diff --git "a/dryrun.log" "b/dryrun.log"
new file mode 100644--- /dev/null
+++ "b/dryrun.log"
@@ -0,0 +1,5807 @@
+
Loading weights:   0%|          | 0/311 [00:00<?, ?it/s]
Loading weights:   0%|          | 1/311 [00:00<00:00, 20164.92it/s, Materializing param=lm_head.weight]
Loading weights:   0%|          | 1/311 [00:00<00:00, 2978.91it/s, Materializing param=lm_head.weight] 
Loading weights:   1%|          | 2/311 [00:00<00:00, 3125.41it/s, Materializing param=model.embed_tokens.weight]
Loading weights:   1%|          | 2/311 [00:00<00:00, 2495.87it/s, Materializing param=model.embed_tokens.weight]
Loading weights:   1%|          | 3/311 [00:00<00:00, 2667.00it/s, Materializing param=model.layers.0.input_layernorm.weight]
Loading weights:   1%|          | 3/311 [00:00<00:00, 2499.59it/s, Materializing param=model.layers.0.input_layernorm.weight]
Loading weights:   1%|▏         | 4/311 [00:00<00:00, 2539.69it/s, Materializing param=model.layers.0.mlp.down_proj.weight]  
Loading weights:   1%|▏         | 4/311 [00:00<00:00, 2452.81it/s, Materializing param=model.layers.0.mlp.down_proj.weight]
Loading weights:   2%|▏         | 5/311 [00:00<00:00, 2540.16it/s, Materializing param=model.layers.0.mlp.gate_proj.weight]
Loading weights:   2%|▏         | 5/311 [00:00<00:00, 2361.92it/s, Materializing param=model.layers.0.mlp.gate_proj.weight]
Loading weights:   2%|▏         | 6/311 [00:00<00:00, 2500.83it/s, Materializing param=model.layers.0.mlp.up_proj.weight]  
Loading weights:   2%|▏         | 6/311 [00:00<00:00, 2377.27it/s, Materializing param=model.layers.0.mlp.up_proj.weight]
Loading weights:   2%|▏         | 7/311 [00:00<00:00, 2462.27it/s, Materializing param=model.layers.0.post_attention_layernorm.weight]
Loading weights:   2%|▏         | 7/311 [00:00<00:00, 2420.26it/s, Materializing param=model.layers.0.post_attention_layernorm.weight]
Loading weights:   3%|▎         | 8/311 [00:00<00:00, 2437.66it/s, Materializing param=model.layers.0.self_attn.k_norm.weight]        
Loading weights:   3%|▎         | 8/311 [00:00<00:00, 2199.86it/s, Materializing param=model.layers.0.self_attn.k_norm.weight]
Loading weights:   3%|▎         | 9/311 [00:00<00:00, 2227.59it/s, Materializing param=model.layers.0.self_attn.k_proj.weight]
Loading weights:   3%|▎         | 9/311 [00:00<00:00, 2195.84it/s, Materializing param=model.layers.0.self_attn.k_proj.weight]
Loading weights:   3%|▎         | 10/311 [00:00<00:00, 2292.47it/s, Materializing param=model.layers.0.self_attn.o_proj.weight]
Loading weights:   3%|▎         | 10/311 [00:00<00:00, 2218.50it/s, Materializing param=model.layers.0.self_attn.o_proj.weight]
Loading weights:   4%|▎         | 11/311 [00:00<00:00, 2236.09it/s, Materializing param=model.layers.0.self_attn.q_norm.weight]
Loading weights:   4%|▎         | 11/311 [00:00<00:00, 2211.66it/s, Materializing param=model.layers.0.self_attn.q_norm.weight]
Loading weights:   4%|▍         | 12/311 [00:00<00:00, 2326.18it/s, Materializing param=model.layers.0.self_attn.q_proj.weight]
Loading weights:   4%|▍         | 12/311 [00:00<00:00, 2213.74it/s, Materializing param=model.layers.0.self_attn.q_proj.weight]
Loading weights:   4%|▍         | 13/311 [00:00<00:00, 2257.52it/s, Materializing param=model.layers.0.self_attn.v_proj.weight]
Loading weights:   4%|▍         | 13/311 [00:00<00:00, 2214.61it/s, Materializing param=model.layers.0.self_attn.v_proj.weight]
Loading weights:   5%|▍         | 14/311 [00:00<00:00, 2199.59it/s, Materializing param=model.layers.1.input_layernorm.weight] 
Loading weights:   5%|▍         | 14/311 [00:00<00:00, 2183.07it/s, Materializing param=model.layers.1.input_layernorm.weight]
Loading weights:   5%|▍         | 15/311 [00:00<00:00, 2259.94it/s, Materializing param=model.layers.1.mlp.down_proj.weight]  
Loading weights:   5%|▍         | 15/311 [00:00<00:00, 2242.14it/s, Materializing param=model.layers.1.mlp.down_proj.weight]
Loading weights:   5%|▌         | 16/311 [00:00<00:00, 2315.46it/s, Materializing param=model.layers.1.mlp.gate_proj.weight]
Loading weights:   5%|▌         | 16/311 [00:00<00:00, 2300.38it/s, Materializing param=model.layers.1.mlp.gate_proj.weight]
Loading weights:   5%|▌         | 17/311 [00:00<00:00, 2319.10it/s, Materializing param=model.layers.1.mlp.up_proj.weight]  
Loading weights:   5%|▌         | 17/311 [00:00<00:00, 2298.92it/s, Materializing param=model.layers.1.mlp.up_proj.weight]
Loading weights:   6%|▌         | 18/311 [00:00<00:00, 2313.04it/s, Materializing param=model.layers.1.post_attention_layernorm.weight]
Loading weights:   6%|▌         | 18/311 [00:00<00:00, 2286.55it/s, Materializing param=model.layers.1.post_attention_layernorm.weight]
Loading weights:   6%|▌         | 19/311 [00:00<00:00, 2308.64it/s, Materializing param=model.layers.1.self_attn.k_norm.weight]        
Loading weights:   6%|▌         | 19/311 [00:00<00:00, 2255.45it/s, Materializing param=model.layers.1.self_attn.k_norm.weight]
Loading weights:   6%|▋         | 20/311 [00:00<00:00, 2344.37it/s, Materializing param=model.layers.1.self_attn.k_proj.weight]
Loading weights:   6%|▋         | 20/311 [00:00<00:00, 2298.94it/s, Materializing param=model.layers.1.self_attn.k_proj.weight]
Loading weights:   7%|▋         | 21/311 [00:00<00:00, 2286.79it/s, Materializing param=model.layers.1.self_attn.o_proj.weight]
Loading weights:   7%|▋         | 21/311 [00:00<00:00, 2260.85it/s, Materializing param=model.layers.1.self_attn.o_proj.weight]
Loading weights:   7%|▋         | 22/311 [00:00<00:00, 2339.86it/s, Materializing param=model.layers.1.self_attn.q_norm.weight]
Loading weights:   7%|▋         | 22/311 [00:00<00:00, 2323.01it/s, Materializing param=model.layers.1.self_attn.q_norm.weight]
Loading weights:   7%|▋         | 23/311 [00:00<00:00, 2398.18it/s, Materializing param=model.layers.1.self_attn.q_proj.weight]
Loading weights:   7%|▋         | 23/311 [00:00<00:00, 2331.35it/s, Materializing param=model.layers.1.self_attn.q_proj.weight]
Loading weights:   8%|▊         | 24/311 [00:00<00:00, 2395.60it/s, Materializing param=model.layers.1.self_attn.v_proj.weight]
Loading weights:   8%|▊         | 24/311 [00:00<00:00, 2320.87it/s, Materializing param=model.layers.1.self_attn.v_proj.weight]
Loading weights:   8%|▊         | 25/311 [00:00<00:00, 2353.07it/s, Materializing param=model.layers.2.input_layernorm.weight] 
Loading weights:   8%|▊         | 25/311 [00:00<00:00, 2310.15it/s, Materializing param=model.layers.2.input_layernorm.weight]
Loading weights:   8%|▊         | 26/311 [00:00<00:00, 2355.28it/s, Materializing param=model.layers.2.mlp.down_proj.weight]  
Loading weights:   8%|▊         | 26/311 [00:00<00:00, 2341.88it/s, Materializing param=model.layers.2.mlp.down_proj.weight]
Loading weights:   9%|▊         | 27/311 [00:00<00:00, 2410.16it/s, Materializing param=model.layers.2.mlp.gate_proj.weight]
Loading weights:   9%|▊         | 27/311 [00:00<00:00, 2387.10it/s, Materializing param=model.layers.2.mlp.gate_proj.weight]
Loading weights:   9%|▉         | 28/311 [00:00<00:00, 2394.45it/s, Materializing param=model.layers.2.mlp.up_proj.weight]  
Loading weights:   9%|▉         | 28/311 [00:00<00:00, 2382.64it/s, Materializing param=model.layers.2.mlp.up_proj.weight]
Loading weights:   9%|▉         | 29/311 [00:00<00:00, 2399.06it/s, Materializing param=model.layers.2.post_attention_layernorm.weight]
Loading weights:   9%|▉         | 29/311 [00:00<00:00, 2390.43it/s, Materializing param=model.layers.2.post_attention_layernorm.weight]
Loading weights:  10%|▉         | 30/311 [00:00<00:00, 2428.34it/s, Materializing param=model.layers.2.self_attn.k_norm.weight]        
Loading weights:  10%|▉         | 30/311 [00:00<00:00, 2378.85it/s, Materializing param=model.layers.2.self_attn.k_norm.weight]
Loading weights:  10%|▉         | 31/311 [00:00<00:00, 2427.89it/s, Materializing param=model.layers.2.self_attn.k_proj.weight]
Loading weights:  10%|▉         | 31/311 [00:00<00:00, 2411.46it/s, Materializing param=model.layers.2.self_attn.k_proj.weight]
Loading weights:  10%|█         | 32/311 [00:00<00:00, 2441.83it/s, Materializing param=model.layers.2.self_attn.o_proj.weight]
Loading weights:  10%|█         | 32/311 [00:00<00:00, 2418.69it/s, Materializing param=model.layers.2.self_attn.o_proj.weight]
Loading weights:  11%|█         | 33/311 [00:00<00:00, 2429.05it/s, Materializing param=model.layers.2.self_attn.q_norm.weight]
Loading weights:  11%|█         | 33/311 [00:00<00:00, 2400.28it/s, Materializing param=model.layers.2.self_attn.q_norm.weight]
Loading weights:  11%|█         | 34/311 [00:00<00:00, 2452.26it/s, Materializing param=model.layers.2.self_attn.q_proj.weight]
Loading weights:  11%|█         | 34/311 [00:00<00:00, 2394.41it/s, Materializing param=model.layers.2.self_attn.q_proj.weight]
Loading weights:  11%|█▏        | 35/311 [00:00<00:00, 2433.01it/s, Materializing param=model.layers.2.self_attn.v_proj.weight]
Loading weights:  11%|█▏        | 35/311 [00:00<00:00, 2422.89it/s, Materializing param=model.layers.2.self_attn.v_proj.weight]
Loading weights:  12%|█▏        | 36/311 [00:00<00:00, 2465.10it/s, Materializing param=model.layers.3.input_layernorm.weight] 
Loading weights:  12%|█▏        | 36/311 [00:00<00:00, 2431.09it/s, Materializing param=model.layers.3.input_layernorm.weight]
Loading weights:  12%|█▏        | 37/311 [00:00<00:00, 2463.63it/s, Materializing param=model.layers.3.mlp.down_proj.weight]  
Loading weights:  12%|█▏        | 37/311 [00:00<00:00, 2448.94it/s, Materializing param=model.layers.3.mlp.down_proj.weight]
Loading weights:  12%|█▏        | 38/311 [00:00<00:00, 2467.73it/s, Materializing param=model.layers.3.mlp.gate_proj.weight]
Loading weights:  12%|█▏        | 38/311 [00:00<00:00, 2453.72it/s, Materializing param=model.layers.3.mlp.gate_proj.weight]
Loading weights:  13%|█▎        | 39/311 [00:00<00:00, 2483.76it/s, Materializing param=model.layers.3.mlp.up_proj.weight]  
Loading weights:  13%|█▎        | 39/311 [00:00<00:00, 2474.74it/s, Materializing param=model.layers.3.mlp.up_proj.weight]
Loading weights:  13%|█▎        | 40/311 [00:00<00:00, 2516.08it/s, Materializing param=model.layers.3.post_attention_layernorm.weight]
Loading weights:  13%|█▎        | 40/311 [00:00<00:00, 2482.79it/s, Materializing param=model.layers.3.post_attention_layernorm.weight]
Loading weights:  13%|█▎        | 41/311 [00:00<00:00, 2490.68it/s, Materializing param=model.layers.3.self_attn.k_norm.weight]        
Loading weights:  13%|█▎        | 41/311 [00:00<00:00, 2461.31it/s, Materializing param=model.layers.3.self_attn.k_norm.weight]
Loading weights:  14%|█▎        | 42/311 [00:00<00:00, 2493.57it/s, Materializing param=model.layers.3.self_attn.k_proj.weight]
Loading weights:  14%|█▎        | 42/311 [00:00<00:00, 2476.71it/s, Materializing param=model.layers.3.self_attn.k_proj.weight]
Loading weights:  14%|█▍        | 43/311 [00:00<00:00, 2504.62it/s, Materializing param=model.layers.3.self_attn.o_proj.weight]
Loading weights:  14%|█▍        | 43/311 [00:00<00:00, 2492.40it/s, Materializing param=model.layers.3.self_attn.o_proj.weight]
Loading weights:  14%|█▍        | 44/311 [00:00<00:00, 2508.04it/s, Materializing param=model.layers.3.self_attn.q_norm.weight]
Loading weights:  14%|█▍        | 44/311 [00:00<00:00, 2498.00it/s, Materializing param=model.layers.3.self_attn.q_norm.weight]
Loading weights:  14%|█▍        | 45/311 [00:00<00:00, 2505.99it/s, Materializing param=model.layers.3.self_attn.q_proj.weight]
Loading weights:  14%|█▍        | 45/311 [00:00<00:00, 2473.25it/s, Materializing param=model.layers.3.self_attn.q_proj.weight]
Loading weights:  15%|█▍        | 46/311 [00:00<00:00, 2504.13it/s, Materializing param=model.layers.3.self_attn.v_proj.weight]
Loading weights:  15%|█▍        | 46/311 [00:00<00:00, 2479.83it/s, Materializing param=model.layers.3.self_attn.v_proj.weight]
Loading weights:  15%|█▌        | 47/311 [00:00<00:00, 2517.56it/s, Materializing param=model.layers.4.input_layernorm.weight] 
Loading weights:  15%|█▌        | 47/311 [00:00<00:00, 2474.76it/s, Materializing param=model.layers.4.input_layernorm.weight]
Loading weights:  15%|█▌        | 48/311 [00:00<00:00, 2466.63it/s, Materializing param=model.layers.4.mlp.down_proj.weight]  
Loading weights:  15%|█▌        | 48/311 [00:00<00:00, 2457.99it/s, Materializing param=model.layers.4.mlp.down_proj.weight]
Loading weights:  16%|█▌        | 49/311 [00:00<00:00, 2460.92it/s, Materializing param=model.layers.4.mlp.gate_proj.weight]
Loading weights:  16%|█▌        | 49/311 [00:00<00:00, 2455.71it/s, Materializing param=model.layers.4.mlp.gate_proj.weight]
Loading weights:  16%|█▌        | 50/311 [00:00<00:00, 2445.00it/s, Materializing param=model.layers.4.mlp.up_proj.weight]  
Loading weights:  16%|█▌        | 50/311 [00:00<00:00, 2427.23it/s, Materializing param=model.layers.4.mlp.up_proj.weight]
Loading weights:  16%|█▋        | 51/311 [00:00<00:00, 2444.85it/s, Materializing param=model.layers.4.post_attention_layernorm.weight]
Loading weights:  16%|█▋        | 51/311 [00:00<00:00, 2436.35it/s, Materializing param=model.layers.4.post_attention_layernorm.weight]
Loading weights:  17%|█▋        | 52/311 [00:00<00:00, 2442.45it/s, Materializing param=model.layers.4.self_attn.k_norm.weight]        
Loading weights:  17%|█▋        | 52/311 [00:00<00:00, 2426.94it/s, Materializing param=model.layers.4.self_attn.k_norm.weight]
Loading weights:  17%|█▋        | 53/311 [00:00<00:00, 2448.54it/s, Materializing param=model.layers.4.self_attn.k_proj.weight]
Loading weights:  17%|█▋        | 53/311 [00:00<00:00, 2430.15it/s, Materializing param=model.layers.4.self_attn.k_proj.weight]
Loading weights:  17%|█▋        | 54/311 [00:00<00:00, 2453.15it/s, Materializing param=model.layers.4.self_attn.o_proj.weight]
Loading weights:  17%|█▋        | 54/311 [00:00<00:00, 2438.10it/s, Materializing param=model.layers.4.self_attn.o_proj.weight]
Loading weights:  18%|█▊        | 55/311 [00:00<00:00, 2430.40it/s, Materializing param=model.layers.4.self_attn.q_norm.weight]
Loading weights:  18%|█▊        | 55/311 [00:00<00:00, 2423.89it/s, Materializing param=model.layers.4.self_attn.q_norm.weight]
Loading weights:  18%|█▊        | 56/311 [00:00<00:00, 2433.87it/s, Materializing param=model.layers.4.self_attn.q_proj.weight]
Loading weights:  18%|█▊        | 56/311 [00:00<00:00, 2424.15it/s, Materializing param=model.layers.4.self_attn.q_proj.weight]
Loading weights:  18%|█▊        | 57/311 [00:00<00:00, 2422.51it/s, Materializing param=model.layers.4.self_attn.v_proj.weight]
Loading weights:  18%|█▊        | 57/311 [00:00<00:00, 2391.38it/s, Materializing param=model.layers.4.self_attn.v_proj.weight]
Loading weights:  19%|█▊        | 58/311 [00:00<00:00, 2423.49it/s, Materializing param=model.layers.5.input_layernorm.weight] 
Loading weights:  19%|█▊        | 58/311 [00:00<00:00, 2405.28it/s, Materializing param=model.layers.5.input_layernorm.weight]
Loading weights:  19%|█▉        | 59/311 [00:00<00:00, 2411.65it/s, Materializing param=model.layers.5.mlp.down_proj.weight]  
Loading weights:  19%|█▉        | 59/311 [00:00<00:00, 2405.67it/s, Materializing param=model.layers.5.mlp.down_proj.weight]
Loading weights:  19%|█▉        | 60/311 [00:00<00:00, 2419.93it/s, Materializing param=model.layers.5.mlp.gate_proj.weight]
Loading weights:  19%|█▉        | 60/311 [00:00<00:00, 2393.87it/s, Materializing param=model.layers.5.mlp.gate_proj.weight]
Loading weights:  20%|█▉        | 61/311 [00:00<00:00, 2391.79it/s, Materializing param=model.layers.5.mlp.up_proj.weight]  
Loading weights:  20%|█▉        | 61/311 [00:00<00:00, 2380.69it/s, Materializing param=model.layers.5.mlp.up_proj.weight]
Loading weights:  20%|█▉        | 62/311 [00:00<00:00, 2401.17it/s, Materializing param=model.layers.5.post_attention_layernorm.weight]
Loading weights:  20%|█▉        | 62/311 [00:00<00:00, 2376.03it/s, Materializing param=model.layers.5.post_attention_layernorm.weight]
Loading weights:  20%|██        | 63/311 [00:00<00:00, 2385.41it/s, Materializing param=model.layers.5.self_attn.k_norm.weight]        
Loading weights:  20%|██        | 63/311 [00:00<00:00, 2380.36it/s, Materializing param=model.layers.5.self_attn.k_norm.weight]
Loading weights:  21%|██        | 64/311 [00:00<00:00, 2394.82it/s, Materializing param=model.layers.5.self_attn.k_proj.weight]
Loading weights:  21%|██        | 64/311 [00:00<00:00, 2387.56it/s, Materializing param=model.layers.5.self_attn.k_proj.weight]
Loading weights:  21%|██        | 65/311 [00:00<00:00, 2380.19it/s, Materializing param=model.layers.5.self_attn.o_proj.weight]
Loading weights:  21%|██        | 65/311 [00:00<00:00, 2368.08it/s, Materializing param=model.layers.5.self_attn.o_proj.weight]
Loading weights:  21%|██        | 66/311 [00:00<00:00, 2396.06it/s, Materializing param=model.layers.5.self_attn.q_norm.weight]
Loading weights:  21%|██        | 66/311 [00:00<00:00, 2389.92it/s, Materializing param=model.layers.5.self_attn.q_norm.weight]
Loading weights:  22%|██▏       | 67/311 [00:00<00:00, 2400.20it/s, Materializing param=model.layers.5.self_attn.q_proj.weight]
Loading weights:  22%|██▏       | 67/311 [00:00<00:00, 2394.05it/s, Materializing param=model.layers.5.self_attn.q_proj.weight]
Loading weights:  22%|██▏       | 68/311 [00:00<00:00, 2400.84it/s, Materializing param=model.layers.5.self_attn.v_proj.weight]
Loading weights:  22%|██▏       | 68/311 [00:00<00:00, 2382.07it/s, Materializing param=model.layers.5.self_attn.v_proj.weight]
Loading weights:  22%|██▏       | 69/311 [00:00<00:00, 2386.68it/s, Materializing param=model.layers.6.input_layernorm.weight] 
Loading weights:  22%|██▏       | 69/311 [00:00<00:00, 2382.93it/s, Materializing param=model.layers.6.input_layernorm.weight]
Loading weights:  23%|██▎       | 70/311 [00:00<00:00, 2410.06it/s, Materializing param=model.layers.6.mlp.down_proj.weight]  
Loading weights:  23%|██▎       | 70/311 [00:00<00:00, 2394.65it/s, Materializing param=model.layers.6.mlp.down_proj.weight]
Loading weights:  23%|██▎       | 71/311 [00:00<00:00, 2391.26it/s, Materializing param=model.layers.6.mlp.gate_proj.weight]
Loading weights:  23%|██▎       | 71/311 [00:00<00:00, 2380.96it/s, Materializing param=model.layers.6.mlp.gate_proj.weight]
Loading weights:  23%|██▎       | 72/311 [00:00<00:00, 2393.61it/s, Materializing param=model.layers.6.mlp.up_proj.weight]  
Loading weights:  23%|██▎       | 72/311 [00:00<00:00, 2381.23it/s, Materializing param=model.layers.6.mlp.up_proj.weight]
Loading weights:  23%|██▎       | 73/311 [00:00<00:00, 2390.31it/s, Materializing param=model.layers.6.post_attention_layernorm.weight]
Loading weights:  23%|██▎       | 73/311 [00:00<00:00, 2370.40it/s, Materializing param=model.layers.6.post_attention_layernorm.weight]
Loading weights:  24%|██▍       | 74/311 [00:00<00:00, 2390.73it/s, Materializing param=model.layers.6.self_attn.k_norm.weight]        
Loading weights:  24%|██▍       | 74/311 [00:00<00:00, 2383.16it/s, Materializing param=model.layers.6.self_attn.k_norm.weight]
Loading weights:  24%|██▍       | 75/311 [00:00<00:00, 2394.25it/s, Materializing param=model.layers.6.self_attn.k_proj.weight]
Loading weights:  24%|██▍       | 75/311 [00:00<00:00, 2386.02it/s, Materializing param=model.layers.6.self_attn.k_proj.weight]
Loading weights:  24%|██▍       | 76/311 [00:00<00:00, 2406.35it/s, Materializing param=model.layers.6.self_attn.o_proj.weight]
Loading weights:  24%|██▍       | 76/311 [00:00<00:00, 2403.12it/s, Materializing param=model.layers.6.self_attn.o_proj.weight]
Loading weights:  25%|██▍       | 77/311 [00:00<00:00, 2415.59it/s, Materializing param=model.layers.6.self_attn.q_norm.weight]
Loading weights:  25%|██▍       | 77/311 [00:00<00:00, 2402.97it/s, Materializing param=model.layers.6.self_attn.q_norm.weight]
Loading weights:  25%|██▌       | 78/311 [00:00<00:00, 2413.83it/s, Materializing param=model.layers.6.self_attn.q_proj.weight]
Loading weights:  25%|██▌       | 78/311 [00:00<00:00, 2404.94it/s, Materializing param=model.layers.6.self_attn.q_proj.weight]
Loading weights:  25%|██▌       | 79/311 [00:00<00:00, 2403.23it/s, Materializing param=model.layers.6.self_attn.v_proj.weight]
Loading weights:  25%|██▌       | 79/311 [00:00<00:00, 2399.28it/s, Materializing param=model.layers.6.self_attn.v_proj.weight]
Loading weights:  26%|██▌       | 80/311 [00:00<00:00, 2394.06it/s, Materializing param=model.layers.7.input_layernorm.weight] 
Loading weights:  26%|██▌       | 80/311 [00:00<00:00, 2389.51it/s, Materializing param=model.layers.7.input_layernorm.weight]
Loading weights:  26%|██▌       | 81/311 [00:00<00:00, 2380.34it/s, Materializing param=model.layers.7.mlp.down_proj.weight]  
Loading weights:  26%|██▌       | 81/311 [00:00<00:00, 2374.83it/s, Materializing param=model.layers.7.mlp.down_proj.weight]
Loading weights:  26%|██▋       | 82/311 [00:00<00:00, 2384.71it/s, Materializing param=model.layers.7.mlp.gate_proj.weight]
Loading weights:  26%|██▋       | 82/311 [00:00<00:00, 2376.43it/s, Materializing param=model.layers.7.mlp.gate_proj.weight]
Loading weights:  27%|██▋       | 83/311 [00:00<00:00, 2392.20it/s, Materializing param=model.layers.7.mlp.up_proj.weight]  
Loading weights:  27%|██▋       | 83/311 [00:00<00:00, 2388.11it/s, Materializing param=model.layers.7.mlp.up_proj.weight]
Loading weights:  27%|██▋       | 84/311 [00:00<00:00, 2374.58it/s, Materializing param=model.layers.7.post_attention_layernorm.weight]
Loading weights:  27%|██▋       | 84/311 [00:00<00:00, 2371.32it/s, Materializing param=model.layers.7.post_attention_layernorm.weight]
Loading weights:  27%|██▋       | 85/311 [00:00<00:00, 2382.09it/s, Materializing param=model.layers.7.self_attn.k_norm.weight]        
Loading weights:  27%|██▋       | 85/311 [00:00<00:00, 2372.39it/s, Materializing param=model.layers.7.self_attn.k_norm.weight]
Loading weights:  28%|██▊       | 86/311 [00:00<00:00, 2392.31it/s, Materializing param=model.layers.7.self_attn.k_proj.weight]
Loading weights:  28%|██▊       | 86/311 [00:00<00:00, 2389.63it/s, Materializing param=model.layers.7.self_attn.k_proj.weight]
Loading weights:  28%|██▊       | 87/311 [00:00<00:00, 2395.39it/s, Materializing param=model.layers.7.self_attn.o_proj.weight]
Loading weights:  28%|██▊       | 87/311 [00:00<00:00, 2390.67it/s, Materializing param=model.layers.7.self_attn.o_proj.weight]
Loading weights:  28%|██▊       | 88/311 [00:00<00:00, 2396.48it/s, Materializing param=model.layers.7.self_attn.q_norm.weight]
Loading weights:  28%|██▊       | 88/311 [00:00<00:00, 2391.28it/s, Materializing param=model.layers.7.self_attn.q_norm.weight]
Loading weights:  29%|██▊       | 89/311 [00:00<00:00, 2393.58it/s, Materializing param=model.layers.7.self_attn.q_proj.weight]
Loading weights:  29%|██▊       | 89/311 [00:00<00:00, 2387.04it/s, Materializing param=model.layers.7.self_attn.q_proj.weight]
Loading weights:  29%|██▉       | 90/311 [00:00<00:00, 2398.42it/s, Materializing param=model.layers.7.self_attn.v_proj.weight]
Loading weights:  29%|██▉       | 90/311 [00:00<00:00, 2395.71it/s, Materializing param=model.layers.7.self_attn.v_proj.weight]
Loading weights:  29%|██▉       | 91/311 [00:00<00:00, 2405.31it/s, Materializing param=model.layers.8.input_layernorm.weight] 
Loading weights:  29%|██▉       | 91/311 [00:00<00:00, 2396.10it/s, Materializing param=model.layers.8.input_layernorm.weight]
Loading weights:  30%|██▉       | 92/311 [00:00<00:00, 2411.30it/s, Materializing param=model.layers.8.mlp.down_proj.weight]  
Loading weights:  30%|██▉       | 92/311 [00:00<00:00, 2408.61it/s, Materializing param=model.layers.8.mlp.down_proj.weight]
Loading weights:  30%|██▉       | 93/311 [00:00<00:00, 2410.58it/s, Materializing param=model.layers.8.mlp.gate_proj.weight]
Loading weights:  30%|██▉       | 93/311 [00:00<00:00, 2407.02it/s, Materializing param=model.layers.8.mlp.gate_proj.weight]
Loading weights:  30%|███       | 94/311 [00:00<00:00, 2426.51it/s, Materializing param=model.layers.8.mlp.up_proj.weight]  
Loading weights:  30%|███       | 94/311 [00:00<00:00, 2421.34it/s, Materializing param=model.layers.8.mlp.up_proj.weight]
Loading weights:  31%|███       | 95/311 [00:00<00:00, 2426.18it/s, Materializing param=model.layers.8.post_attention_layernorm.weight]
Loading weights:  31%|███       | 95/311 [00:00<00:00, 2423.57it/s, Materializing param=model.layers.8.post_attention_layernorm.weight]
Loading weights:  31%|███       | 96/311 [00:00<00:00, 2428.72it/s, Materializing param=model.layers.8.self_attn.k_norm.weight]        
Loading weights:  31%|███       | 96/311 [00:00<00:00, 2422.51it/s, Materializing param=model.layers.8.self_attn.k_norm.weight]
Loading weights:  31%|███       | 97/311 [00:00<00:00, 2437.18it/s, Materializing param=model.layers.8.self_attn.k_proj.weight]
Loading weights:  31%|███       | 97/311 [00:00<00:00, 2434.59it/s, Materializing param=model.layers.8.self_attn.k_proj.weight]
Loading weights:  32%|███▏      | 98/311 [00:00<00:00, 2424.01it/s, Materializing param=model.layers.8.self_attn.o_proj.weight]
Loading weights:  32%|███▏      | 98/311 [00:00<00:00, 2410.84it/s, Materializing param=model.layers.8.self_attn.o_proj.weight]
Loading weights:  32%|███▏      | 99/311 [00:00<00:00, 2420.86it/s, Materializing param=model.layers.8.self_attn.q_norm.weight]
Loading weights:  32%|███▏      | 99/311 [00:00<00:00, 2413.08it/s, Materializing param=model.layers.8.self_attn.q_norm.weight]
Loading weights:  32%|███▏      | 100/311 [00:00<00:00, 2412.53it/s, Materializing param=model.layers.8.self_attn.q_proj.weight]
Loading weights:  32%|███▏      | 100/311 [00:00<00:00, 2407.77it/s, Materializing param=model.layers.8.self_attn.q_proj.weight]
Loading weights:  32%|███▏      | 101/311 [00:00<00:00, 2425.79it/s, Materializing param=model.layers.8.self_attn.v_proj.weight]
Loading weights:  32%|███▏      | 101/311 [00:00<00:00, 2423.54it/s, Materializing param=model.layers.8.self_attn.v_proj.weight]
Loading weights:  33%|███▎      | 102/311 [00:00<00:00, 2443.87it/s, Materializing param=model.layers.9.input_layernorm.weight] 
Loading weights:  33%|███▎      | 102/311 [00:00<00:00, 2441.81it/s, Materializing param=model.layers.9.input_layernorm.weight]
Loading weights:  33%|███▎      | 103/311 [00:00<00:00, 2461.43it/s, Materializing param=model.layers.9.mlp.down_proj.weight]  
Loading weights:  33%|███▎      | 103/311 [00:00<00:00, 2459.40it/s, Materializing param=model.layers.9.mlp.down_proj.weight]
Loading weights:  33%|███▎      | 104/311 [00:00<00:00, 2479.57it/s, Materializing param=model.layers.9.mlp.gate_proj.weight]
Loading weights:  33%|███▎      | 104/311 [00:00<00:00, 2477.52it/s, Materializing param=model.layers.9.mlp.gate_proj.weight]
Loading weights:  34%|███▍      | 105/311 [00:00<00:00, 2497.66it/s, Materializing param=model.layers.9.mlp.up_proj.weight]  
Loading weights:  34%|███▍      | 105/311 [00:00<00:00, 2495.66it/s, Materializing param=model.layers.9.mlp.up_proj.weight]
Loading weights:  34%|███▍      | 106/311 [00:00<00:00, 2515.77it/s, Materializing param=model.layers.9.post_attention_layernorm.weight]
Loading weights:  34%|███▍      | 106/311 [00:00<00:00, 2513.66it/s, Materializing param=model.layers.9.post_attention_layernorm.weight]
Loading weights:  34%|███▍      | 107/311 [00:00<00:00, 2533.71it/s, Materializing param=model.layers.9.self_attn.k_norm.weight]        
Loading weights:  34%|███▍      | 107/311 [00:00<00:00, 2530.61it/s, Materializing param=model.layers.9.self_attn.k_norm.weight]
Loading weights:  35%|███▍      | 108/311 [00:00<00:00, 2550.10it/s, Materializing param=model.layers.9.self_attn.k_proj.weight]
Loading weights:  35%|███▍      | 108/311 [00:00<00:00, 2548.02it/s, Materializing param=model.layers.9.self_attn.k_proj.weight]
Loading weights:  35%|███▌      | 109/311 [00:00<00:00, 2567.97it/s, Materializing param=model.layers.9.self_attn.o_proj.weight]
Loading weights:  35%|███▌      | 109/311 [00:00<00:00, 2565.83it/s, Materializing param=model.layers.9.self_attn.o_proj.weight]
Loading weights:  35%|███▌      | 110/311 [00:00<00:00, 2585.74it/s, Materializing param=model.layers.9.self_attn.q_norm.weight]
Loading weights:  35%|███▌      | 110/311 [00:00<00:00, 2583.68it/s, Materializing param=model.layers.9.self_attn.q_norm.weight]
Loading weights:  36%|███▌      | 111/311 [00:00<00:00, 2603.44it/s, Materializing param=model.layers.9.self_attn.q_proj.weight]
Loading weights:  36%|███▌      | 111/311 [00:00<00:00, 2601.40it/s, Materializing param=model.layers.9.self_attn.q_proj.weight]
Loading weights:  36%|███▌      | 112/311 [00:00<00:00, 2621.10it/s, Materializing param=model.layers.9.self_attn.v_proj.weight]
Loading weights:  36%|███▌      | 112/311 [00:00<00:00, 2619.07it/s, Materializing param=model.layers.9.self_attn.v_proj.weight]
Loading weights:  36%|███▋      | 113/311 [00:00<00:00, 2637.85it/s, Materializing param=model.layers.10.input_layernorm.weight]
Loading weights:  36%|███▋      | 113/311 [00:00<00:00, 2635.67it/s, Materializing param=model.layers.10.input_layernorm.weight]
Loading weights:  37%|███▋      | 114/311 [00:00<00:00, 2655.30it/s, Materializing param=model.layers.10.mlp.down_proj.weight]  
Loading weights:  37%|███▋      | 114/311 [00:00<00:00, 2653.21it/s, Materializing param=model.layers.10.mlp.down_proj.weight]
Loading weights:  37%|███▋      | 115/311 [00:00<00:00, 2672.66it/s, Materializing param=model.layers.10.mlp.gate_proj.weight]
Loading weights:  37%|███▋      | 115/311 [00:00<00:00, 2670.55it/s, Materializing param=model.layers.10.mlp.gate_proj.weight]
Loading weights:  37%|███▋      | 116/311 [00:00<00:00, 2689.93it/s, Materializing param=model.layers.10.mlp.up_proj.weight]  
Loading weights:  37%|███▋      | 116/311 [00:00<00:00, 2687.84it/s, Materializing param=model.layers.10.mlp.up_proj.weight]
Loading weights:  38%|███▊      | 117/311 [00:00<00:00, 2707.21it/s, Materializing param=model.layers.10.post_attention_layernorm.weight]
Loading weights:  38%|███▊      | 117/311 [00:00<00:00, 2704.35it/s, Materializing param=model.layers.10.post_attention_layernorm.weight]
Loading weights:  38%|███▊      | 118/311 [00:00<00:00, 2722.99it/s, Materializing param=model.layers.10.self_attn.k_norm.weight]        
Loading weights:  38%|███▊      | 118/311 [00:00<00:00, 2720.79it/s, Materializing param=model.layers.10.self_attn.k_norm.weight]
Loading weights:  38%|███▊      | 119/311 [00:00<00:00, 2739.86it/s, Materializing param=model.layers.10.self_attn.k_proj.weight]
Loading weights:  38%|███▊      | 119/311 [00:00<00:00, 2737.71it/s, Materializing param=model.layers.10.self_attn.k_proj.weight]
Loading weights:  39%|███▊      | 120/311 [00:00<00:00, 2756.81it/s, Materializing param=model.layers.10.self_attn.o_proj.weight]
Loading weights:  39%|███▊      | 120/311 [00:00<00:00, 2754.67it/s, Materializing param=model.layers.10.self_attn.o_proj.weight]
Loading weights:  39%|███▉      | 121/311 [00:00<00:00, 2773.95it/s, Materializing param=model.layers.10.self_attn.q_norm.weight]
Loading weights:  39%|███▉      | 121/311 [00:00<00:00, 2771.86it/s, Materializing param=model.layers.10.self_attn.q_norm.weight]
Loading weights:  39%|███▉      | 122/311 [00:00<00:00, 2791.09it/s, Materializing param=model.layers.10.self_attn.q_proj.weight]
Loading weights:  39%|███▉      | 122/311 [00:00<00:00, 2788.96it/s, Materializing param=model.layers.10.self_attn.q_proj.weight]
Loading weights:  40%|███▉      | 123/311 [00:00<00:00, 2807.31it/s, Materializing param=model.layers.10.self_attn.v_proj.weight]
Loading weights:  40%|███▉      | 123/311 [00:00<00:00, 2804.91it/s, Materializing param=model.layers.10.self_attn.v_proj.weight]
Loading weights:  40%|███▉      | 124/311 [00:00<00:00, 2823.82it/s, Materializing param=model.layers.11.input_layernorm.weight] 
Loading weights:  40%|███▉      | 124/311 [00:00<00:00, 2821.52it/s, Materializing param=model.layers.11.input_layernorm.weight]
Loading weights:  40%|████      | 125/311 [00:00<00:00, 2840.58it/s, Materializing param=model.layers.11.mlp.down_proj.weight]  
Loading weights:  40%|████      | 125/311 [00:00<00:00, 2838.52it/s, Materializing param=model.layers.11.mlp.down_proj.weight]
Loading weights:  41%|████      | 126/311 [00:00<00:00, 2857.42it/s, Materializing param=model.layers.11.mlp.gate_proj.weight]
Loading weights:  41%|████      | 126/311 [00:00<00:00, 2855.35it/s, Materializing param=model.layers.11.mlp.gate_proj.weight]
Loading weights:  41%|████      | 127/311 [00:00<00:00, 2874.30it/s, Materializing param=model.layers.11.mlp.up_proj.weight]  
Loading weights:  41%|████      | 127/311 [00:00<00:00, 2872.24it/s, Materializing param=model.layers.11.mlp.up_proj.weight]
Loading weights:  41%|████      | 128/311 [00:00<00:00, 2890.19it/s, Materializing param=model.layers.11.post_attention_layernorm.weight]
Loading weights:  41%|████      | 128/311 [00:00<00:00, 2887.57it/s, Materializing param=model.layers.11.post_attention_layernorm.weight]
Loading weights:  41%|████▏     | 129/311 [00:00<00:00, 2905.98it/s, Materializing param=model.layers.11.self_attn.k_norm.weight]        
Loading weights:  41%|████▏     | 129/311 [00:00<00:00, 2903.75it/s, Materializing param=model.layers.11.self_attn.k_norm.weight]
Loading weights:  42%|████▏     | 130/311 [00:00<00:00, 2922.28it/s, Materializing param=model.layers.11.self_attn.k_proj.weight]
Loading weights:  42%|████▏     | 130/311 [00:00<00:00, 2920.06it/s, Materializing param=model.layers.11.self_attn.k_proj.weight]
Loading weights:  42%|████▏     | 131/311 [00:00<00:00, 2938.57it/s, Materializing param=model.layers.11.self_attn.o_proj.weight]
Loading weights:  42%|████▏     | 131/311 [00:00<00:00, 2936.39it/s, Materializing param=model.layers.11.self_attn.o_proj.weight]
Loading weights:  42%|████▏     | 132/311 [00:00<00:00, 2954.87it/s, Materializing param=model.layers.11.self_attn.q_norm.weight]
Loading weights:  42%|████▏     | 132/311 [00:00<00:00, 2952.70it/s, Materializing param=model.layers.11.self_attn.q_norm.weight]
Loading weights:  43%|████▎     | 133/311 [00:00<00:00, 2971.12it/s, Materializing param=model.layers.11.self_attn.q_proj.weight]
Loading weights:  43%|████▎     | 133/311 [00:00<00:00, 2968.15it/s, Materializing param=model.layers.11.self_attn.q_proj.weight]
Loading weights:  43%|████▎     | 134/311 [00:00<00:00, 2986.17it/s, Materializing param=model.layers.11.self_attn.v_proj.weight]
Loading weights:  43%|████▎     | 134/311 [00:00<00:00, 2983.94it/s, Materializing param=model.layers.11.self_attn.v_proj.weight]
Loading weights:  43%|████▎     | 135/311 [00:00<00:00, 3002.13it/s, Materializing param=model.layers.12.input_layernorm.weight] 
Loading weights:  43%|████▎     | 135/311 [00:00<00:00, 2999.90it/s, Materializing param=model.layers.12.input_layernorm.weight]
Loading weights:  44%|████▎     | 136/311 [00:00<00:00, 3018.24it/s, Materializing param=model.layers.12.mlp.down_proj.weight]  
Loading weights:  44%|████▎     | 136/311 [00:00<00:00, 3016.00it/s, Materializing param=model.layers.12.mlp.down_proj.weight]
Loading weights:  44%|████▍     | 137/311 [00:00<00:00, 3034.04it/s, Materializing param=model.layers.12.mlp.gate_proj.weight]
Loading weights:  44%|████▍     | 137/311 [00:00<00:00, 3031.77it/s, Materializing param=model.layers.12.mlp.gate_proj.weight]
Loading weights:  44%|████▍     | 138/311 [00:00<00:00, 3048.96it/s, Materializing param=model.layers.12.mlp.up_proj.weight]  
Loading weights:  44%|████▍     | 138/311 [00:00<00:00, 3046.13it/s, Materializing param=model.layers.12.mlp.up_proj.weight]
Loading weights:  45%|████▍     | 139/311 [00:00<00:00, 3063.55it/s, Materializing param=model.layers.12.post_attention_layernorm.weight]
Loading weights:  45%|████▍     | 139/311 [00:00<00:00, 3061.12it/s, Materializing param=model.layers.12.post_attention_layernorm.weight]
Loading weights:  45%|████▌     | 140/311 [00:00<00:00, 3079.03it/s, Materializing param=model.layers.12.self_attn.k_norm.weight]        
Loading weights:  45%|████▌     | 140/311 [00:00<00:00, 3076.73it/s, Materializing param=model.layers.12.self_attn.k_norm.weight]
Loading weights:  45%|████▌     | 141/311 [00:00<00:00, 3094.67it/s, Materializing param=model.layers.12.self_attn.k_proj.weight]
Loading weights:  45%|████▌     | 141/311 [00:00<00:00, 3092.38it/s, Materializing param=model.layers.12.self_attn.k_proj.weight]
Loading weights:  46%|████▌     | 142/311 [00:00<00:00, 3110.17it/s, Materializing param=model.layers.12.self_attn.o_proj.weight]
Loading weights:  46%|████▌     | 142/311 [00:00<00:00, 3107.82it/s, Materializing param=model.layers.12.self_attn.o_proj.weight]
Loading weights:  46%|████▌     | 143/311 [00:00<00:00, 3125.64it/s, Materializing param=model.layers.12.self_attn.q_norm.weight]
Loading weights:  46%|████▌     | 143/311 [00:00<00:00, 3122.61it/s, Materializing param=model.layers.12.self_attn.q_norm.weight]
Loading weights:  46%|████▋     | 144/311 [00:00<00:00, 3139.91it/s, Materializing param=model.layers.12.self_attn.q_proj.weight]
Loading weights:  46%|████▋     | 144/311 [00:00<00:00, 3137.59it/s, Materializing param=model.layers.12.self_attn.q_proj.weight]
Loading weights:  47%|████▋     | 145/311 [00:00<00:00, 3155.22it/s, Materializing param=model.layers.12.self_attn.v_proj.weight]
Loading weights:  47%|████▋     | 145/311 [00:00<00:00, 3152.89it/s, Materializing param=model.layers.12.self_attn.v_proj.weight]
Loading weights:  47%|████▋     | 146/311 [00:00<00:00, 3170.54it/s, Materializing param=model.layers.13.input_layernorm.weight] 
Loading weights:  47%|████▋     | 146/311 [00:00<00:00, 3168.28it/s, Materializing param=model.layers.13.input_layernorm.weight]
Loading weights:  47%|████▋     | 147/311 [00:00<00:00, 3186.02it/s, Materializing param=model.layers.13.mlp.down_proj.weight]  
Loading weights:  47%|████▋     | 147/311 [00:00<00:00, 3183.77it/s, Materializing param=model.layers.13.mlp.down_proj.weight]
Loading weights:  48%|████▊     | 148/311 [00:00<00:00, 3201.38it/s, Materializing param=model.layers.13.mlp.gate_proj.weight]
Loading weights:  48%|████▊     | 148/311 [00:00<00:00, 3197.88it/s, Materializing param=model.layers.13.mlp.gate_proj.weight]
Loading weights:  48%|████▊     | 149/311 [00:00<00:00, 3214.97it/s, Materializing param=model.layers.13.mlp.up_proj.weight]  
Loading weights:  48%|████▊     | 149/311 [00:00<00:00, 3212.61it/s, Materializing param=model.layers.13.mlp.up_proj.weight]
Loading weights:  48%|████▊     | 150/311 [00:00<00:00, 3230.02it/s, Materializing param=model.layers.13.post_attention_layernorm.weight]
Loading weights:  48%|████▊     | 150/311 [00:00<00:00, 3227.66it/s, Materializing param=model.layers.13.post_attention_layernorm.weight]
Loading weights:  49%|████▊     | 151/311 [00:00<00:00, 3245.02it/s, Materializing param=model.layers.13.self_attn.k_norm.weight]        
Loading weights:  49%|████▊     | 151/311 [00:00<00:00, 3242.66it/s, Materializing param=model.layers.13.self_attn.k_norm.weight]
Loading weights:  49%|████▉     | 152/311 [00:00<00:00, 3259.91it/s, Materializing param=model.layers.13.self_attn.k_proj.weight]
Loading weights:  49%|████▉     | 152/311 [00:00<00:00, 3257.55it/s, Materializing param=model.layers.13.self_attn.k_proj.weight]
Loading weights:  49%|████▉     | 153/311 [00:00<00:00, 3274.83it/s, Materializing param=model.layers.13.self_attn.o_proj.weight]
Loading weights:  49%|████▉     | 153/311 [00:00<00:00, 3272.36it/s, Materializing param=model.layers.13.self_attn.o_proj.weight]
Loading weights:  50%|████▉     | 154/311 [00:00<00:00, 3288.39it/s, Materializing param=model.layers.13.self_attn.q_norm.weight]
Loading weights:  50%|████▉     | 154/311 [00:00<00:00, 3285.87it/s, Materializing param=model.layers.13.self_attn.q_norm.weight]
Loading weights:  50%|████▉     | 155/311 [00:00<00:00, 3303.02it/s, Materializing param=model.layers.13.self_attn.q_proj.weight]
Loading weights:  50%|████▉     | 155/311 [00:00<00:00, 3300.62it/s, Materializing param=model.layers.13.self_attn.q_proj.weight]
Loading weights:  50%|█████     | 156/311 [00:00<00:00, 3317.69it/s, Materializing param=model.layers.13.self_attn.v_proj.weight]
Loading weights:  50%|█████     | 156/311 [00:00<00:00, 3315.35it/s, Materializing param=model.layers.13.self_attn.v_proj.weight]
Loading weights:  50%|█████     | 157/311 [00:00<00:00, 3332.52it/s, Materializing param=model.layers.14.input_layernorm.weight] 
Loading weights:  50%|█████     | 157/311 [00:00<00:00, 3330.21it/s, Materializing param=model.layers.14.input_layernorm.weight]
Loading weights:  51%|█████     | 158/311 [00:00<00:00, 3347.32it/s, Materializing param=model.layers.14.mlp.down_proj.weight]  
Loading weights:  51%|█████     | 158/311 [00:00<00:00, 3344.99it/s, Materializing param=model.layers.14.mlp.down_proj.weight]
Loading weights:  51%|█████     | 159/311 [00:00<00:00, 3360.67it/s, Materializing param=model.layers.14.mlp.gate_proj.weight]
Loading weights:  51%|█████     | 159/311 [00:00<00:00, 3357.93it/s, Materializing param=model.layers.14.mlp.gate_proj.weight]
Loading weights:  51%|█████▏    | 160/311 [00:00<00:00, 3374.58it/s, Materializing param=model.layers.14.mlp.up_proj.weight]  
Loading weights:  51%|█████▏    | 160/311 [00:00<00:00, 3372.22it/s, Materializing param=model.layers.14.mlp.up_proj.weight]
Loading weights:  52%|█████▏    | 161/311 [00:00<00:00, 3389.02it/s, Materializing param=model.layers.14.post_attention_layernorm.weight]
Loading weights:  52%|█████▏    | 161/311 [00:00<00:00, 3386.57it/s, Materializing param=model.layers.14.post_attention_layernorm.weight]
Loading weights:  52%|█████▏    | 162/311 [00:00<00:00, 3403.50it/s, Materializing param=model.layers.14.self_attn.k_norm.weight]        
Loading weights:  52%|█████▏    | 162/311 [00:00<00:00, 3401.11it/s, Materializing param=model.layers.14.self_attn.k_norm.weight]
Loading weights:  52%|█████▏    | 163/311 [00:00<00:00, 3417.78it/s, Materializing param=model.layers.14.self_attn.k_proj.weight]
Loading weights:  52%|█████▏    | 163/311 [00:00<00:00, 3415.37it/s, Materializing param=model.layers.14.self_attn.k_proj.weight]
Loading weights:  53%|█████▎    | 164/311 [00:00<00:00, 3431.28it/s, Materializing param=model.layers.14.self_attn.o_proj.weight]
Loading weights:  53%|█████▎    | 164/311 [00:00<00:00, 3428.58it/s, Materializing param=model.layers.14.self_attn.o_proj.weight]
Loading weights:  53%|█████▎    | 165/311 [00:00<00:00, 3444.99it/s, Materializing param=model.layers.14.self_attn.q_norm.weight]
Loading weights:  53%|█████▎    | 165/311 [00:00<00:00, 3442.52it/s, Materializing param=model.layers.14.self_attn.q_norm.weight]
Loading weights:  53%|█████▎    | 166/311 [00:00<00:00, 3459.07it/s, Materializing param=model.layers.14.self_attn.q_proj.weight]
Loading weights:  53%|█████▎    | 166/311 [00:00<00:00, 3456.70it/s, Materializing param=model.layers.14.self_attn.q_proj.weight]
Loading weights:  54%|█████▎    | 167/311 [00:00<00:00, 3473.07it/s, Materializing param=model.layers.14.self_attn.v_proj.weight]
Loading weights:  54%|█████▎    | 167/311 [00:00<00:00, 3470.66it/s, Materializing param=model.layers.14.self_attn.v_proj.weight]
Loading weights:  54%|█████▍    | 168/311 [00:00<00:00, 3487.12it/s, Materializing param=model.layers.15.input_layernorm.weight] 
Loading weights:  54%|█████▍    | 168/311 [00:00<00:00, 3484.69it/s, Materializing param=model.layers.15.input_layernorm.weight]
Loading weights:  54%|█████▍    | 169/311 [00:00<00:00, 3499.95it/s, Materializing param=model.layers.15.mlp.down_proj.weight]  
Loading weights:  54%|█████▍    | 169/311 [00:00<00:00, 3496.89it/s, Materializing param=model.layers.15.mlp.down_proj.weight]
Loading weights:  55%|█████▍    | 170/311 [00:00<00:00, 3513.04it/s, Materializing param=model.layers.15.mlp.gate_proj.weight]
Loading weights:  55%|█████▍    | 170/311 [00:00<00:00, 3510.67it/s, Materializing param=model.layers.15.mlp.gate_proj.weight]
Loading weights:  55%|█████▍    | 171/311 [00:00<00:00, 3527.12it/s, Materializing param=model.layers.15.mlp.up_proj.weight]  
Loading weights:  55%|█████▍    | 171/311 [00:00<00:00, 3524.85it/s, Materializing param=model.layers.15.mlp.up_proj.weight]
Loading weights:  55%|█████▌    | 172/311 [00:00<00:00, 3541.22it/s, Materializing param=model.layers.15.post_attention_layernorm.weight]
Loading weights:  55%|█████▌    | 172/311 [00:00<00:00, 3538.77it/s, Materializing param=model.layers.15.post_attention_layernorm.weight]
Loading weights:  56%|█████▌    | 173/311 [00:00<00:00, 3555.03it/s, Materializing param=model.layers.15.self_attn.k_norm.weight]        
Loading weights:  56%|█████▌    | 173/311 [00:00<00:00, 3552.65it/s, Materializing param=model.layers.15.self_attn.k_norm.weight]
Loading weights:  56%|█████▌    | 174/311 [00:00<00:00, 3568.75it/s, Materializing param=model.layers.15.self_attn.k_proj.weight]
Loading weights:  56%|█████▌    | 174/311 [00:00<00:00, 3565.59it/s, Materializing param=model.layers.15.self_attn.k_proj.weight]
Loading weights:  56%|█████▋    | 175/311 [00:00<00:00, 3581.41it/s, Materializing param=model.layers.15.self_attn.o_proj.weight]
Loading weights:  56%|█████▋    | 175/311 [00:00<00:00, 3579.02it/s, Materializing param=model.layers.15.self_attn.o_proj.weight]
Loading weights:  57%|█████▋    | 176/311 [00:00<00:00, 3595.16it/s, Materializing param=model.layers.15.self_attn.q_norm.weight]
Loading weights:  57%|█████▋    | 176/311 [00:00<00:00, 3592.80it/s, Materializing param=model.layers.15.self_attn.q_norm.weight]
Loading weights:  57%|█████▋    | 177/311 [00:00<00:00, 3608.85it/s, Materializing param=model.layers.15.self_attn.q_proj.weight]
Loading weights:  57%|█████▋    | 177/311 [00:00<00:00, 3606.43it/s, Materializing param=model.layers.15.self_attn.q_proj.weight]
Loading weights:  57%|█████▋    | 178/311 [00:00<00:00, 3622.55it/s, Materializing param=model.layers.15.self_attn.v_proj.weight]
Loading weights:  57%|█████▋    | 178/311 [00:00<00:00, 3620.11it/s, Materializing param=model.layers.15.self_attn.v_proj.weight]
Loading weights:  58%|█████▊    | 179/311 [00:00<00:00, 3636.13it/s, Materializing param=model.layers.16.input_layernorm.weight] 
Loading weights:  58%|█████▊    | 179/311 [00:00<00:00, 3632.50it/s, Materializing param=model.layers.16.input_layernorm.weight]
Loading weights:  58%|█████▊    | 180/311 [00:00<00:00, 3648.07it/s, Materializing param=model.layers.16.mlp.down_proj.weight]  
Loading weights:  58%|█████▊    | 180/311 [00:00<00:00, 3645.60it/s, Materializing param=model.layers.16.mlp.down_proj.weight]
Loading weights:  58%|█████▊    | 181/311 [00:00<00:00, 3661.40it/s, Materializing param=model.layers.16.mlp.gate_proj.weight]
Loading weights:  58%|█████▊    | 181/311 [00:00<00:00, 3659.07it/s, Materializing param=model.layers.16.mlp.gate_proj.weight]
Loading weights:  59%|█████▊    | 182/311 [00:00<00:00, 3675.14it/s, Materializing param=model.layers.16.mlp.up_proj.weight]  
Loading weights:  59%|█████▊    | 182/311 [00:00<00:00, 3672.72it/s, Materializing param=model.layers.16.mlp.up_proj.weight]
Loading weights:  59%|█████▉    | 183/311 [00:00<00:00, 3688.53it/s, Materializing param=model.layers.16.post_attention_layernorm.weight]
Loading weights:  59%|█████▉    | 183/311 [00:00<00:00, 3686.12it/s, Materializing param=model.layers.16.post_attention_layernorm.weight]
Loading weights:  59%|█████▉    | 184/311 [00:00<00:00, 3702.07it/s, Materializing param=model.layers.16.self_attn.k_norm.weight]        
Loading weights:  59%|█████▉    | 184/311 [00:00<00:00, 3699.71it/s, Materializing param=model.layers.16.self_attn.k_norm.weight]
Loading weights:  59%|█████▉    | 185/311 [00:00<00:00, 3714.78it/s, Materializing param=model.layers.16.self_attn.k_proj.weight]
Loading weights:  59%|█████▉    | 185/311 [00:00<00:00, 3712.08it/s, Materializing param=model.layers.16.self_attn.k_proj.weight]
Loading weights:  60%|█████▉    | 186/311 [00:00<00:00, 3727.43it/s, Materializing param=model.layers.16.self_attn.o_proj.weight]
Loading weights:  60%|█████▉    | 186/311 [00:00<00:00, 3724.80it/s, Materializing param=model.layers.16.self_attn.o_proj.weight]
Loading weights:  60%|██████    | 187/311 [00:00<00:00, 3740.16it/s, Materializing param=model.layers.16.self_attn.q_norm.weight]
Loading weights:  60%|██████    | 187/311 [00:00<00:00, 3737.70it/s, Materializing param=model.layers.16.self_attn.q_norm.weight]
Loading weights:  60%|██████    | 188/311 [00:00<00:00, 3753.27it/s, Materializing param=model.layers.16.self_attn.q_proj.weight]
Loading weights:  60%|██████    | 188/311 [00:00<00:00, 3750.88it/s, Materializing param=model.layers.16.self_attn.q_proj.weight]
Loading weights:  61%|██████    | 189/311 [00:00<00:00, 3766.52it/s, Materializing param=model.layers.16.self_attn.v_proj.weight]
Loading weights:  61%|██████    | 189/311 [00:00<00:00, 3764.15it/s, Materializing param=model.layers.16.self_attn.v_proj.weight]
Loading weights:  61%|██████    | 190/311 [00:00<00:00, 3778.71it/s, Materializing param=model.layers.17.input_layernorm.weight] 
Loading weights:  61%|██████    | 190/311 [00:00<00:00, 3775.66it/s, Materializing param=model.layers.17.input_layernorm.weight]
Loading weights:  61%|██████▏   | 191/311 [00:00<00:00, 3790.97it/s, Materializing param=model.layers.17.mlp.down_proj.weight]  
Loading weights:  61%|██████▏   | 191/311 [00:00<00:00, 3788.48it/s, Materializing param=model.layers.17.mlp.down_proj.weight]
Loading weights:  62%|██████▏   | 192/311 [00:00<00:00, 3803.82it/s, Materializing param=model.layers.17.mlp.gate_proj.weight]
Loading weights:  62%|██████▏   | 192/311 [00:00<00:00, 3801.32it/s, Materializing param=model.layers.17.mlp.gate_proj.weight]
Loading weights:  62%|██████▏   | 193/311 [00:00<00:00, 3816.58it/s, Materializing param=model.layers.17.mlp.up_proj.weight]  
Loading weights:  62%|██████▏   | 193/311 [00:00<00:00, 3814.19it/s, Materializing param=model.layers.17.mlp.up_proj.weight]
Loading weights:  62%|██████▏   | 194/311 [00:00<00:00, 3829.50it/s, Materializing param=model.layers.17.post_attention_layernorm.weight]
Loading weights:  62%|██████▏   | 194/311 [00:00<00:00, 3826.97it/s, Materializing param=model.layers.17.post_attention_layernorm.weight]
Loading weights:  63%|██████▎   | 195/311 [00:00<00:00, 3842.29it/s, Materializing param=model.layers.17.self_attn.k_norm.weight]        
Loading weights:  63%|██████▎   | 195/311 [00:00<00:00, 3838.95it/s, Materializing param=model.layers.17.self_attn.k_norm.weight]
Loading weights:  63%|██████▎   | 196/311 [00:00<00:00, 3853.65it/s, Materializing param=model.layers.17.self_attn.k_proj.weight]
Loading weights:  63%|██████▎   | 196/311 [00:00<00:00, 3851.19it/s, Materializing param=model.layers.17.self_attn.k_proj.weight]
Loading weights:  63%|██████▎   | 197/311 [00:00<00:00, 3866.24it/s, Materializing param=model.layers.17.self_attn.o_proj.weight]
Loading weights:  63%|██████▎   | 197/311 [00:00<00:00, 3863.78it/s, Materializing param=model.layers.17.self_attn.o_proj.weight]
Loading weights:  64%|██████▎   | 198/311 [00:00<00:00, 3878.99it/s, Materializing param=model.layers.17.self_attn.q_norm.weight]
Loading weights:  64%|██████▎   | 198/311 [00:00<00:00, 3876.56it/s, Materializing param=model.layers.17.self_attn.q_norm.weight]
Loading weights:  64%|██████▍   | 199/311 [00:00<00:00, 3891.65it/s, Materializing param=model.layers.17.self_attn.q_proj.weight]
Loading weights:  64%|██████▍   | 199/311 [00:00<00:00, 3889.21it/s, Materializing param=model.layers.17.self_attn.q_proj.weight]
Loading weights:  64%|██████▍   | 200/311 [00:00<00:00, 3904.20it/s, Materializing param=model.layers.17.self_attn.v_proj.weight]
Loading weights:  64%|██████▍   | 200/311 [00:00<00:00, 3900.66it/s, Materializing param=model.layers.17.self_attn.v_proj.weight]
Loading weights:  65%|██████▍   | 201/311 [00:00<00:00, 3914.74it/s, Materializing param=model.layers.18.input_layernorm.weight] 
Loading weights:  65%|██████▍   | 201/311 [00:00<00:00, 3912.11it/s, Materializing param=model.layers.18.input_layernorm.weight]
Loading weights:  65%|██████▍   | 202/311 [00:00<00:00, 3926.87it/s, Materializing param=model.layers.18.mlp.down_proj.weight]  
Loading weights:  65%|██████▍   | 202/311 [00:00<00:00, 3924.43it/s, Materializing param=model.layers.18.mlp.down_proj.weight]
Loading weights:  65%|██████▌   | 203/311 [00:00<00:00, 3939.33it/s, Materializing param=model.layers.18.mlp.gate_proj.weight]
Loading weights:  65%|██████▌   | 203/311 [00:00<00:00, 3936.89it/s, Materializing param=model.layers.18.mlp.gate_proj.weight]
Loading weights:  66%|██████▌   | 204/311 [00:00<00:00, 3951.85it/s, Materializing param=model.layers.18.mlp.up_proj.weight]  
Loading weights:  66%|██████▌   | 204/311 [00:00<00:00, 3949.46it/s, Materializing param=model.layers.18.mlp.up_proj.weight]
Loading weights:  66%|██████▌   | 205/311 [00:00<00:00, 3964.46it/s, Materializing param=model.layers.18.post_attention_layernorm.weight]
Loading weights:  66%|██████▌   | 205/311 [00:00<00:00, 3961.94it/s, Materializing param=model.layers.18.post_attention_layernorm.weight]
Loading weights:  66%|██████▌   | 206/311 [00:00<00:00, 3975.92it/s, Materializing param=model.layers.18.self_attn.k_norm.weight]        
Loading weights:  66%|██████▌   | 206/311 [00:00<00:00, 3973.43it/s, Materializing param=model.layers.18.self_attn.k_norm.weight]
Loading weights:  67%|██████���   | 207/311 [00:00<00:00, 3988.24it/s, Materializing param=model.layers.18.self_attn.k_proj.weight]
Loading weights:  67%|██████▋   | 207/311 [00:00<00:00, 3985.75it/s, Materializing param=model.layers.18.self_attn.k_proj.weight]
Loading weights:  67%|██████▋   | 208/311 [00:00<00:00, 4000.36it/s, Materializing param=model.layers.18.self_attn.o_proj.weight]
Loading weights:  67%|██████▋   | 208/311 [00:00<00:00, 3997.93it/s, Materializing param=model.layers.18.self_attn.o_proj.weight]
Loading weights:  67%|██████▋   | 209/311 [00:00<00:00, 4012.62it/s, Materializing param=model.layers.18.self_attn.q_norm.weight]
Loading weights:  67%|██████▋   | 209/311 [00:00<00:00, 4010.22it/s, Materializing param=model.layers.18.self_attn.q_norm.weight]
Loading weights:  68%|██████▊   | 210/311 [00:00<00:00, 4024.93it/s, Materializing param=model.layers.18.self_attn.q_proj.weight]
Loading weights:  68%|██████▊   | 210/311 [00:00<00:00, 4022.56it/s, Materializing param=model.layers.18.self_attn.q_proj.weight]
Loading weights:  68%|██████▊   | 211/311 [00:00<00:00, 4035.98it/s, Materializing param=model.layers.18.self_attn.v_proj.weight]
Loading weights:  68%|██████▊   | 211/311 [00:00<00:00, 4032.76it/s, Materializing param=model.layers.18.self_attn.v_proj.weight]
Loading weights:  68%|██████▊   | 212/311 [00:00<00:00, 4047.03it/s, Materializing param=model.layers.19.input_layernorm.weight] 
Loading weights:  68%|██████▊   | 212/311 [00:00<00:00, 4044.50it/s, Materializing param=model.layers.19.input_layernorm.weight]
Loading weights:  68%|██████▊   | 213/311 [00:00<00:00, 4059.06it/s, Materializing param=model.layers.19.mlp.down_proj.weight]  
Loading weights:  68%|██████▊   | 213/311 [00:00<00:00, 4056.48it/s, Materializing param=model.layers.19.mlp.down_proj.weight]
Loading weights:  69%|██████▉   | 214/311 [00:00<00:00, 4070.83it/s, Materializing param=model.layers.19.mlp.gate_proj.weight]
Loading weights:  69%|██████▉   | 214/311 [00:00<00:00, 4068.39it/s, Materializing param=model.layers.19.mlp.gate_proj.weight]
Loading weights:  69%|██████▉   | 215/311 [00:00<00:00, 4082.91it/s, Materializing param=model.layers.19.mlp.up_proj.weight]  
Loading weights:  69%|██████▉   | 215/311 [00:00<00:00, 4080.41it/s, Materializing param=model.layers.19.mlp.up_proj.weight]
Loading weights:  69%|██████▉   | 216/311 [00:00<00:00, 4095.02it/s, Materializing param=model.layers.19.post_attention_layernorm.weight]
Loading weights:  69%|██████▉   | 216/311 [00:00<00:00, 4091.75it/s, Materializing param=model.layers.19.post_attention_layernorm.weight]
Loading weights:  70%|██████▉   | 217/311 [00:00<00:00, 4105.72it/s, Materializing param=model.layers.19.self_attn.k_norm.weight]        
Loading weights:  70%|██████▉   | 217/311 [00:00<00:00, 4103.18it/s, Materializing param=model.layers.19.self_attn.k_norm.weight]
Loading weights:  70%|███████   | 218/311 [00:00<00:00, 4117.51it/s, Materializing param=model.layers.19.self_attn.k_proj.weight]
Loading weights:  70%|███████   | 218/311 [00:00<00:00, 4115.02it/s, Materializing param=model.layers.19.self_attn.k_proj.weight]
Loading weights:  70%|███████   | 219/311 [00:00<00:00, 4129.44it/s, Materializing param=model.layers.19.self_attn.o_proj.weight]
Loading weights:  70%|███████   | 219/311 [00:00<00:00, 4126.92it/s, Materializing param=model.layers.19.self_attn.o_proj.weight]
Loading weights:  71%|███████   | 220/311 [00:00<00:00, 4141.07it/s, Materializing param=model.layers.19.self_attn.q_norm.weight]
Loading weights:  71%|███████   | 220/311 [00:00<00:00, 4138.57it/s, Materializing param=model.layers.19.self_attn.q_norm.weight]
Loading weights:  71%|███████   | 221/311 [00:00<00:00, 4152.66it/s, Materializing param=model.layers.19.self_attn.q_proj.weight]
Loading weights:  71%|███████   | 221/311 [00:00<00:00, 4149.17it/s, Materializing param=model.layers.19.self_attn.q_proj.weight]
Loading weights:  71%|███████▏  | 222/311 [00:00<00:00, 4162.37it/s, Materializing param=model.layers.19.self_attn.v_proj.weight]
Loading weights:  71%|███████▏  | 222/311 [00:00<00:00, 4159.81it/s, Materializing param=model.layers.19.self_attn.v_proj.weight]
Loading weights:  72%|███████▏  | 223/311 [00:00<00:00, 4174.00it/s, Materializing param=model.layers.20.input_layernorm.weight] 
Loading weights:  72%|███████▏  | 223/311 [00:00<00:00, 4171.56it/s, Materializing param=model.layers.20.input_layernorm.weight]
Loading weights:  72%|███████▏  | 224/311 [00:00<00:00, 4185.63it/s, Materializing param=model.layers.20.mlp.down_proj.weight]  
Loading weights:  72%|███████▏  | 224/311 [00:00<00:00, 4183.17it/s, Materializing param=model.layers.20.mlp.down_proj.weight]
Loading weights:  72%|███████▏  | 225/311 [00:00<00:00, 4196.92it/s, Materializing param=model.layers.20.mlp.gate_proj.weight]
Loading weights:  72%|███████▏  | 225/311 [00:00<00:00, 4194.43it/s, Materializing param=model.layers.20.mlp.gate_proj.weight]
Loading weights:  73%|███████▎  | 226/311 [00:00<00:00, 4208.42it/s, Materializing param=model.layers.20.mlp.up_proj.weight]  
Loading weights:  73%|███████▎  | 226/311 [00:00<00:00, 4205.97it/s, Materializing param=model.layers.20.mlp.up_proj.weight]
Loading weights:  73%|███████▎  | 227/311 [00:00<00:00, 4219.00it/s, Materializing param=model.layers.20.post_attention_layernorm.weight]
Loading weights:  73%|███████▎  | 227/311 [00:00<00:00, 4216.15it/s, Materializing param=model.layers.20.post_attention_layernorm.weight]
Loading weights:  73%|███████▎  | 228/311 [00:00<00:00, 4229.91it/s, Materializing param=model.layers.20.self_attn.k_norm.weight]        
Loading weights:  73%|███████▎  | 228/311 [00:00<00:00, 4227.40it/s, Materializing param=model.layers.20.self_attn.k_norm.weight]
Loading weights:  74%|███████▎  | 229/311 [00:00<00:00, 4241.27it/s, Materializing param=model.layers.20.self_attn.k_proj.weight]
Loading weights:  74%|███████▎  | 229/311 [00:00<00:00, 4238.80it/s, Materializing param=model.layers.20.self_attn.k_proj.weight]
Loading weights:  74%|███████▍  | 230/311 [00:00<00:00, 4252.58it/s, Materializing param=model.layers.20.self_attn.o_proj.weight]
Loading weights:  74%|███████▍  | 230/311 [00:00<00:00, 4250.11it/s, Materializing param=model.layers.20.self_attn.o_proj.weight]
Loading weights:  74%|███████▍  | 231/311 [00:00<00:00, 4264.15it/s, Materializing param=model.layers.20.self_attn.q_norm.weight]
Loading weights:  74%|███████▍  | 231/311 [00:00<00:00, 4261.66it/s, Materializing param=model.layers.20.self_attn.q_norm.weight]
Loading weights:  75%|███████▍  | 232/311 [00:00<00:00, 4274.58it/s, Materializing param=model.layers.20.self_attn.q_proj.weight]
Loading weights:  75%|███████▍  | 232/311 [00:00<00:00, 4271.35it/s, Materializing param=model.layers.20.self_attn.q_proj.weight]
Loading weights:  75%|███████▍  | 233/311 [00:00<00:00, 4284.78it/s, Materializing param=model.layers.20.self_attn.v_proj.weight]
Loading weights:  75%|███████▍  | 233/311 [00:00<00:00, 4282.15it/s, Materializing param=model.layers.20.self_attn.v_proj.weight]
Loading weights:  75%|███████▌  | 234/311 [00:00<00:00, 4295.66it/s, Materializing param=model.layers.21.input_layernorm.weight] 
Loading weights:  75%|███████▌  | 234/311 [00:00<00:00, 4293.08it/s, Materializing param=model.layers.21.input_layernorm.weight]
Loading weights:  76%|███████▌  | 235/311 [00:00<00:00, 4306.93it/s, Materializing param=model.layers.21.mlp.down_proj.weight]  
Loading weights:  76%|███████▌  | 235/311 [00:00<00:00, 4304.42it/s, Materializing param=model.layers.21.mlp.down_proj.weight]
Loading weights:  76%|███████▌  | 236/311 [00:00<00:00, 4318.25it/s, Materializing param=model.layers.21.mlp.gate_proj.weight]
Loading weights:  76%|███████▌  | 236/311 [00:00<00:00, 4315.77it/s, Materializing param=model.layers.21.mlp.gate_proj.weight]
Loading weights:  76%|███████▌  | 237/311 [00:00<00:00, 4329.32it/s, Materializing param=model.layers.21.mlp.up_proj.weight]  
Loading weights:  76%|███████▌  | 237/311 [00:00<00:00, 4326.25it/s, Materializing param=model.layers.21.mlp.up_proj.weight]
Loading weights:  77%|███████▋  | 238/311 [00:00<00:00, 4339.29it/s, Materializing param=model.layers.21.post_attention_layernorm.weight]
Loading weights:  77%|███████▋  | 238/311 [00:00<00:00, 4336.65it/s, Materializing param=model.layers.21.post_attention_layernorm.weight]
Loading weights:  77%|███████▋  | 239/311 [00:00<00:00, 4348.60it/s, Materializing param=model.layers.21.self_attn.k_norm.weight]        
Loading weights:  77%|███████▋  | 239/311 [00:00<00:00, 4345.77it/s, Materializing param=model.layers.21.self_attn.k_norm.weight]
Loading weights:  77%|███████▋  | 240/311 [00:00<00:00, 4359.15it/s, Materializing param=model.layers.21.self_attn.k_proj.weight]
Loading weights:  77%|███████▋  | 240/311 [00:00<00:00, 4356.55it/s, Materializing param=model.layers.21.self_attn.k_proj.weight]
Loading weights:  77%|███████▋  | 241/311 [00:00<00:00, 4370.05it/s, Materializing param=model.layers.21.self_attn.o_proj.weight]
Loading weights:  77%|███████▋  | 241/311 [00:00<00:00, 4367.54it/s, Materializing param=model.layers.21.self_attn.o_proj.weight]
Loading weights:  78%|███████▊  | 242/311 [00:00<00:00, 4381.10it/s, Materializing param=model.layers.21.self_attn.q_norm.weight]
Loading weights:  78%|███████▊  | 242/311 [00:00<00:00, 4377.58it/s, Materializing param=model.layers.21.self_attn.q_norm.weight]
Loading weights:  78%|███████▊  | 243/311 [00:00<00:00, 4390.20it/s, Materializing param=model.layers.21.self_attn.q_proj.weight]
Loading weights:  78%|███████▊  | 243/311 [00:00<00:00, 4387.56it/s, Materializing param=model.layers.21.self_attn.q_proj.weight]
Loading weights:  78%|███████▊  | 244/311 [00:00<00:00, 4400.99it/s, Materializing param=model.layers.21.self_attn.v_proj.weight]
Loading weights:  78%|███████▊  | 244/311 [00:00<00:00, 4398.47it/s, Materializing param=model.layers.21.self_attn.v_proj.weight]
Loading weights:  79%|███████▉  | 245/311 [00:00<00:00, 4411.74it/s, Materializing param=model.layers.22.input_layernorm.weight] 
Loading weights:  79%|███████▉  | 245/311 [00:00<00:00, 4409.18it/s, Materializing param=model.layers.22.input_layernorm.weight]
Loading weights:  79%|███████▉  | 246/311 [00:00<00:00, 4422.57it/s, Materializing param=model.layers.22.mlp.down_proj.weight]  
Loading weights:  79%|███████▉  | 246/311 [00:00<00:00, 4420.09it/s, Materializing param=model.layers.22.mlp.down_proj.weight]
Loading weights:  79%|███████▉  | 247/311 [00:00<00:00, 4433.33it/s, Materializing param=model.layers.22.mlp.gate_proj.weight]
Loading weights:  79%|███████▉  | 247/311 [00:00<00:00, 4430.79it/s, Materializing param=model.layers.22.mlp.gate_proj.weight]
Loading weights:  80%|███████▉  | 248/311 [00:00<00:00, 4442.99it/s, Materializing param=model.layers.22.mlp.up_proj.weight]  
Loading weights:  80%|███████▉  | 248/311 [00:00<00:00, 4440.20it/s, Materializing param=model.layers.22.mlp.up_proj.weight]
Loading weights:  80%|████████  | 249/311 [00:00<00:00, 4453.33it/s, Materializing param=model.layers.22.post_attention_layernorm.weight]
Loading weights:  80%|████████  | 249/311 [00:00<00:00, 4450.77it/s, Materializing param=model.layers.22.post_attention_layernorm.weight]
Loading weights:  80%|████████  | 250/311 [00:00<00:00, 4464.08it/s, Materializing param=model.layers.22.self_attn.k_norm.weight]        
Loading weights:  80%|████████  | 250/311 [00:00<00:00, 4461.44it/s, Materializing param=model.layers.22.self_attn.k_norm.weight]
Loading weights:  81%|████████  | 251/311 [00:00<00:00, 4474.71it/s, Materializing param=model.layers.22.self_attn.k_proj.weight]
Loading weights:  81%|████████  | 251/311 [00:00<00:00, 4472.15it/s, Materializing param=model.layers.22.self_attn.k_proj.weight]
Loading weights:  81%|████████  | 252/311 [00:00<00:00, 4485.33it/s, Materializing param=model.layers.22.self_attn.o_proj.weight]
Loading weights:  81%|████████  | 252/311 [00:00<00:00, 4482.80it/s, Materializing param=model.layers.22.self_attn.o_proj.weight]
Loading weights:  81%|████████▏ | 253/311 [00:00<00:00, 4494.65it/s, Materializing param=model.layers.22.self_attn.q_norm.weight]
Loading weights:  81%|████████▏ | 253/311 [00:00<00:00, 4491.51it/s, Materializing param=model.layers.22.self_attn.q_norm.weight]
Loading weights:  82%|████████▏ | 254/311 [00:00<00:00, 4504.15it/s, Materializing param=model.layers.22.self_attn.q_proj.weight]
Loading weights:  82%|████████▏ | 254/311 [00:00<00:00, 4501.56it/s, Materializing param=model.layers.22.self_attn.q_proj.weight]
Loading weights:  82%|████████▏ | 255/311 [00:00<00:00, 4514.71it/s, Materializing param=model.layers.22.self_attn.v_proj.weight]
Loading weights:  82%|████████▏ | 255/311 [00:00<00:00, 4512.19it/s, Materializing param=model.layers.22.self_attn.v_proj.weight]
Loading weights:  82%|████████▏ | 256/311 [00:00<00:00, 4525.27it/s, Materializing param=model.layers.23.input_layernorm.weight] 
Loading weights:  82%|████████▏ | 256/311 [00:00<00:00, 4522.58it/s, Materializing param=model.layers.23.input_layernorm.weight]
Loading weights:  83%|████████▎ | 257/311 [00:00<00:00, 4535.57it/s, Materializing param=model.layers.23.mlp.down_proj.weight]  
Loading weights:  83%|████████▎ | 257/311 [00:00<00:00, 4533.01it/s, Materializing param=model.layers.23.mlp.down_proj.weight]
Loading weights:  83%|████████▎ | 258/311 [00:00<00:00, 4545.77it/s, Materializing param=model.layers.23.mlp.gate_proj.weight]
Loading weights:  83%|████████▎ | 258/311 [00:00<00:00, 4542.49it/s, Materializing param=model.layers.23.mlp.gate_proj.weight]
Loading weights:  83%|████████▎ | 259/311 [00:00<00:00, 4554.82it/s, Materializing param=model.layers.23.mlp.up_proj.weight]  
Loading weights:  83%|████████▎ | 259/311 [00:00<00:00, 4552.28it/s, Materializing param=model.layers.23.mlp.up_proj.weight]
Loading weights:  84%|████████▎ | 260/311 [00:00<00:00, 4565.13it/s, Materializing param=model.layers.23.post_attention_layernorm.weight]
Loading weights:  84%|████████▎ | 260/311 [00:00<00:00, 4562.57it/s, Materializing param=model.layers.23.post_attention_layernorm.weight]
Loading weights:  84%|████████▍ | 261/311 [00:00<00:00, 4575.53it/s, Materializing param=model.layers.23.self_attn.k_norm.weight]        
Loading weights:  84%|████████▍ | 261/311 [00:00<00:00, 4572.97it/s, Materializing param=model.layers.23.self_attn.k_norm.weight]
Loading weights:  84%|████████▍ | 262/311 [00:00<00:00, 4585.79it/s, Materializing param=model.layers.23.self_attn.k_proj.weight]
Loading weights:  84%|████████▍ | 262/311 [00:00<00:00, 4583.29it/s, Materializing param=model.layers.23.self_attn.k_proj.weight]
Loading weights:  85%|████████▍ | 263/311 [00:00<00:00, 4596.14it/s, Materializing param=model.layers.23.self_attn.o_proj.weight]
Loading weights:  85%|████████▍ | 263/311 [00:00<00:00, 4592.39it/s, Materializing param=model.layers.23.self_attn.o_proj.weight]
Loading weights:  85%|████████▍ | 264/311 [00:00<00:00, 4604.41it/s, Materializing param=model.layers.23.self_attn.q_norm.weight]
Loading weights:  85%|████████▍ | 264/311 [00:00<00:00, 4601.85it/s, Materializing param=model.layers.23.self_attn.q_norm.weight]
Loading weights:  85%|████████▌ | 265/311 [00:00<00:00, 4614.46it/s, Materializing param=model.layers.23.self_attn.q_proj.weight]
Loading weights:  85%|████████▌ | 265/311 [00:00<00:00, 4611.88it/s, Materializing param=model.layers.23.self_attn.q_proj.weight]
Loading weights:  86%|████████▌ | 266/311 [00:00<00:00, 4624.41it/s, Materializing param=model.layers.23.self_attn.v_proj.weight]
Loading weights:  86%|████████▌ | 266/311 [00:00<00:00, 4621.80it/s, Materializing param=model.layers.23.self_attn.v_proj.weight]
Loading weights:  86%|████████▌ | 267/311 [00:00<00:00, 4634.34it/s, Materializing param=model.layers.24.input_layernorm.weight] 
Loading weights:  86%|████████▌ | 267/311 [00:00<00:00, 4631.79it/s, Materializing param=model.layers.24.input_layernorm.weight]
Loading weights:  86%|████████▌ | 268/311 [00:00<00:00, 4644.34it/s, Materializing param=model.layers.24.mlp.down_proj.weight]  
Loading weights:  86%|████████▌ | 268/311 [00:00<00:00, 4641.86it/s, Materializing param=model.layers.24.mlp.down_proj.weight]
Loading weights:  86%|████████▋ | 269/311 [00:00<00:00, 4653.53it/s, Materializing param=model.layers.24.mlp.gate_proj.weight]
Loading weights:  86%|████████▋ | 269/311 [00:00<00:00, 4650.81it/s, Materializing param=model.layers.24.mlp.gate_proj.weight]
Loading weights:  87%|████████▋ | 270/311 [00:00<00:00, 4663.18it/s, Materializing param=model.layers.24.mlp.up_proj.weight]  
Loading weights:  87%|████████▋ | 270/311 [00:00<00:00, 4660.68it/s, Materializing param=model.layers.24.mlp.up_proj.weight]
Loading weights:  87%|████████▋ | 271/311 [00:00<00:00, 4673.33it/s, Materializing param=model.layers.24.post_attention_layernorm.weight]
Loading weights:  87%|████████▋ | 271/311 [00:00<00:00, 4670.76it/s, Materializing param=model.layers.24.post_attention_layernorm.weight]
Loading weights:  87%|████████▋ | 272/311 [00:00<00:00, 4683.31it/s, Materializing param=model.layers.24.self_attn.k_norm.weight]        
Loading weights:  87%|████████▋ | 272/311 [00:00<00:00, 4680.76it/s, Materializing param=model.layers.24.self_attn.k_norm.weight]
Loading weights:  88%|████████▊ | 273/311 [00:00<00:00, 4693.17it/s, Materializing param=model.layers.24.self_attn.k_proj.weight]
Loading weights:  88%|████████▊ | 273/311 [00:00<00:00, 4690.71it/s, Materializing param=model.layers.24.self_attn.k_proj.weight]
Loading weights:  88%|████████▊ | 274/311 [00:00<00:00, 4702.27it/s, Materializing param=model.layers.24.self_attn.o_proj.weight]
Loading weights:  88%|████████▊ | 274/311 [00:00<00:00, 4699.10it/s, Materializing param=model.layers.24.self_attn.o_proj.weight]
Loading weights:  88%|████████▊ | 275/311 [00:00<00:00, 4711.12it/s, Materializing param=model.layers.24.self_attn.q_norm.weight]
Loading weights:  88%|████████▊ | 275/311 [00:00<00:00, 4708.49it/s, Materializing param=model.layers.24.self_attn.q_norm.weight]
Loading weights:  89%|████████▊ | 276/311 [00:00<00:00, 4720.79it/s, Materializing param=model.layers.24.self_attn.q_proj.weight]
Loading weights:  89%|████████▊ | 276/311 [00:00<00:00, 4718.25it/s, Materializing param=model.layers.24.self_attn.q_proj.weight]
Loading weights:  89%|████████▉ | 277/311 [00:00<00:00, 4730.51it/s, Materializing param=model.layers.24.self_attn.v_proj.weight]
Loading weights:  89%|████████▉ | 277/311 [00:00<00:00, 4727.91it/s, Materializing param=model.layers.24.self_attn.v_proj.weight]
Loading weights:  89%|████████▉ | 278/311 [00:00<00:00, 4740.23it/s, Materializing param=model.layers.25.input_layernorm.weight] 
Loading weights:  89%|████████▉ | 278/311 [00:00<00:00, 4737.69it/s, Materializing param=model.layers.25.input_layernorm.weight]
Loading weights:  90%|████████▉ | 279/311 [00:00<00:00, 4750.10it/s, Materializing param=model.layers.25.mlp.down_proj.weight]  
Loading weights:  90%|████████▉ | 279/311 [00:00<00:00, 4747.58it/s, Materializing param=model.layers.25.mlp.down_proj.weight]
Loading weights:  90%|█████████ | 280/311 [00:00<00:00, 4758.63it/s, Materializing param=model.layers.25.mlp.gate_proj.weight]
Loading weights:  90%|█████████ | 280/311 [00:00<00:00, 4756.01it/s, Materializing param=model.layers.25.mlp.gate_proj.weight]
Loading weights:  90%|█████████ | 281/311 [00:00<00:00, 4768.11it/s, Materializing param=model.layers.25.mlp.up_proj.weight]  
Loading weights:  90%|█████████ | 281/311 [00:00<00:00, 4765.54it/s, Materializing param=model.layers.25.mlp.up_proj.weight]
Loading weights:  91%|█████████ | 282/311 [00:00<00:00, 4777.79it/s, Materializing param=model.layers.25.post_attention_layernorm.weight]
Loading weights:  91%|█████████ | 282/311 [00:00<00:00, 4775.22it/s, Materializing param=model.layers.25.post_attention_layernorm.weight]
Loading weights:  91%|█████████ | 283/311 [00:00<00:00, 4787.67it/s, Materializing param=model.layers.25.self_attn.k_norm.weight]        
Loading weights:  91%|█████████ | 283/311 [00:00<00:00, 4785.18it/s, Materializing param=model.layers.25.self_attn.k_norm.weight]
Loading weights:  91%|█████████▏| 284/311 [00:00<00:00, 4797.33it/s, Materializing param=model.layers.25.self_attn.k_proj.weight]
Loading weights:  91%|█████████▏| 284/311 [00:00<00:00, 4794.74it/s, Materializing param=model.layers.25.self_attn.k_proj.weight]
Loading weights:  92%|█████████▏| 285/311 [00:00<00:00, 4805.69it/s, Materializing param=model.layers.25.self_attn.o_proj.weight]
Loading weights:  92%|█████████▏| 285/311 [00:00<00:00, 4802.66it/s, Materializing param=model.layers.25.self_attn.o_proj.weight]
Loading weights:  92%|█████████▏| 286/311 [00:00<00:00, 4814.52it/s, Materializing param=model.layers.25.self_attn.q_norm.weight]
Loading weights:  92%|█████████▏| 286/311 [00:00<00:00, 4811.95it/s, Materializing param=model.layers.25.self_attn.q_norm.weight]
Loading weights:  92%|█████████▏| 287/311 [00:00<00:00, 4824.07it/s, Materializing param=model.layers.25.self_attn.q_proj.weight]
Loading weights:  92%|█████████▏| 287/311 [00:00<00:00, 4821.58it/s, Materializing param=model.layers.25.self_attn.q_proj.weight]
Loading weights:  93%|█████████▎| 288/311 [00:00<00:00, 4833.75it/s, Materializing param=model.layers.25.self_attn.v_proj.weight]
Loading weights:  93%|█████████▎| 288/311 [00:00<00:00, 4831.30it/s, Materializing param=model.layers.25.self_attn.v_proj.weight]
Loading weights:  93%|█████████▎| 289/311 [00:00<00:00, 4843.60it/s, Materializing param=model.layers.26.input_layernorm.weight] 
Loading weights:  93%|█████████▎| 289/311 [00:00<00:00, 4841.16it/s, Materializing param=model.layers.26.input_layernorm.weight]
Loading weights:  93%|█████████▎| 290/311 [00:00<00:00, 4853.32it/s, Materializing param=model.layers.26.mlp.down_proj.weight]  
Loading weights:  93%|█████████▎| 290/311 [00:00<00:00, 4850.01it/s, Materializing param=model.layers.26.mlp.down_proj.weight]
Loading weights:  94%|█████████▎| 291/311 [00:00<00:00, 4861.58it/s, Materializing param=model.layers.26.mlp.gate_proj.weight]
Loading weights:  94%|█████████▎| 291/311 [00:00<00:00, 4858.92it/s, Materializing param=model.layers.26.mlp.gate_proj.weight]
Loading weights:  94%|█████████▍| 292/311 [00:00<00:00, 4870.76it/s, Materializing param=model.layers.26.mlp.up_proj.weight]  
Loading weights:  94%|█████████▍| 292/311 [00:00<00:00, 4868.20it/s, Materializing param=model.layers.26.mlp.up_proj.weight]
Loading weights:  94%|█████████▍| 293/311 [00:00<00:00, 4880.16it/s, Materializing param=model.layers.26.post_attention_layernorm.weight]
Loading weights:  94%|█████████▍| 293/311 [00:00<00:00, 4877.60it/s, Materializing param=model.layers.26.post_attention_layernorm.weight]
Loading weights:  95%|█████████▍| 294/311 [00:00<00:00, 4889.65it/s, Materializing param=model.layers.26.self_attn.k_norm.weight]        
Loading weights:  95%|█████████▍| 294/311 [00:00<00:00, 4886.94it/s, Materializing param=model.layers.26.self_attn.k_norm.weight]
Loading weights:  95%|█████████▍| 295/311 [00:00<00:00, 4898.74it/s, Materializing param=model.layers.26.self_attn.k_proj.weight]
Loading weights:  95%|█████████▍| 295/311 [00:00<00:00, 4895.24it/s, Materializing param=model.layers.26.self_attn.k_proj.weight]
Loading weights:  95%|█████████▌| 296/311 [00:00<00:00, 4906.05it/s, Materializing param=model.layers.26.self_attn.o_proj.weight]
Loading weights:  95%|█████████▌| 296/311 [00:00<00:00, 4903.39it/s, Materializing param=model.layers.26.self_attn.o_proj.weight]
Loading weights:  95%|█████████▌| 297/311 [00:00<00:00, 4915.18it/s, Materializing param=model.layers.26.self_attn.q_norm.weight]
Loading weights:  95%|█████████▌| 297/311 [00:00<00:00, 4912.54it/s, Materializing param=model.layers.26.self_attn.q_norm.weight]
Loading weights:  96%|█████████▌| 298/311 [00:00<00:00, 4924.25it/s, Materializing param=model.layers.26.self_attn.q_proj.weight]
Loading weights:  96%|█████████▌| 298/311 [00:00<00:00, 4921.67it/s, Materializing param=model.layers.26.self_attn.q_proj.weight]
Loading weights:  96%|█████████▌| 299/311 [00:00<00:00, 4933.50it/s, Materializing param=model.layers.26.self_attn.v_proj.weight]
Loading weights:  96%|█████████▌| 299/311 [00:00<00:00, 4930.92it/s, Materializing param=model.layers.26.self_attn.v_proj.weight]
Loading weights:  96%|█████████▋| 300/311 [00:00<00:00, 4942.56it/s, Materializing param=model.layers.27.input_layernorm.weight] 
Loading weights:  96%|█████████▋| 300/311 [00:00<00:00, 4940.02it/s, Materializing param=model.layers.27.input_layernorm.weight]
Loading weights:  97%|█████████▋| 301/311 [00:00<00:00, 4950.94it/s, Materializing param=model.layers.27.mlp.down_proj.weight]  
Loading weights:  97%|█████████▋| 301/311 [00:00<00:00, 4948.19it/s, Materializing param=model.layers.27.mlp.down_proj.weight]
Loading weights:  97%|█████████▋| 302/311 [00:00<00:00, 4959.73it/s, Materializing param=model.layers.27.mlp.gate_proj.weight]
Loading weights:  97%|█████████▋| 302/311 [00:00<00:00, 4957.15it/s, Materializing param=model.layers.27.mlp.gate_proj.weight]
Loading weights:  97%|█████████▋| 303/311 [00:00<00:00, 4968.84it/s, Materializing param=model.layers.27.mlp.up_proj.weight]  
Loading weights:  97%|█████████▋| 303/311 [00:00<00:00, 4966.29it/s, Materializing param=model.layers.27.mlp.up_proj.weight]
Loading weights:  98%|█████████▊| 304/311 [00:00<00:00, 4978.05it/s, Materializing param=model.layers.27.post_attention_layernorm.weight]
Loading weights:  98%|█████████▊| 304/311 [00:00<00:00, 4975.45it/s, Materializing param=model.layers.27.post_attention_layernorm.weight]
Loading weights:  98%|█████████▊| 305/311 [00:00<00:00, 4987.18it/s, Materializing param=model.layers.27.self_attn.k_norm.weight]        
Loading weights:  98%|█████████▊| 305/311 [00:00<00:00, 4984.62it/s, Materializing param=model.layers.27.self_attn.k_norm.weight]
Loading weights:  98%|█████████▊| 306/311 [00:00<00:00, 4994.89it/s, Materializing param=model.layers.27.self_attn.k_proj.weight]
Loading weights:  98%|█████████▊| 306/311 [00:00<00:00, 4991.63it/s, Materializing param=model.layers.27.self_attn.k_proj.weight]
Loading weights:  99%|█████████▊| 307/311 [00:00<00:00, 5002.61it/s, Materializing param=model.layers.27.self_attn.o_proj.weight]
Loading weights:  99%|█████████▊| 307/311 [00:00<00:00, 5000.04it/s, Materializing param=model.layers.27.self_attn.o_proj.weight]
Loading weights:  99%|█████████▉| 308/311 [00:00<00:00, 5011.33it/s, Materializing param=model.layers.27.self_attn.q_norm.weight]
Loading weights:  99%|█████████▉| 308/311 [00:00<00:00, 5008.78it/s, Materializing param=model.layers.27.self_attn.q_norm.weight]
Loading weights:  99%|█████████▉| 309/311 [00:00<00:00, 5020.30it/s, Materializing param=model.layers.27.self_attn.q_proj.weight]
Loading weights:  99%|█████████▉| 309/311 [00:00<00:00, 5017.77it/s, Materializing param=model.layers.27.self_attn.q_proj.weight]
Loading weights: 100%|█████████▉| 310/311 [00:00<00:00, 5029.10it/s, Materializing param=model.layers.27.self_attn.v_proj.weight]
Loading weights: 100%|█████████▉| 310/311 [00:00<00:00, 5026.54it/s, Materializing param=model.layers.27.self_attn.v_proj.weight]
Loading weights: 100%|██████████| 311/311 [00:00<00:00, 5037.90it/s, Materializing param=model.norm.weight]                      
Loading weights: 100%|██████████| 311/311 [00:00<00:00, 5035.35it/s, Materializing param=model.norm.weight]
Loading weights: 100%|██████████| 311/311 [00:00<00:00, 5028.02it/s, Materializing param=model.norm.weight]
+The tied weights mapping and config for this model specifies to tie model.embed_tokens.weight to lm_head.weight, but both are present in the checkpoints, so we will NOT tie them. You should update the config with `tie_word_embeddings=False` to silence this warning
+
Map:   0%|          | 0/5758 [00:00<?, ? examples/s]
Map:   6%|▌         | 319/5758 [00:00<00:01, 3169.08 examples/s]
Map:  11%|█▏        | 649/5758 [00:00<00:01, 3245.62 examples/s]
Map:  17%|█▋        | 998/5758 [00:00<00:01, 3344.70 examples/s]
Map:  26%|██▌       | 1480/5758 [00:00<00:01, 3270.73 examples/s]
Map:  32%|███▏      | 1842/5758 [00:00<00:01, 3378.29 examples/s]
Map:  38%|███▊      | 2188/5758 [00:00<00:01, 3399.28 examples/s]
Map:  44%|████▍     | 2532/5758 [00:00<00:00, 3407.42 examples/s]
Map:  50%|█████     | 2888/5758 [00:00<00:00, 3452.14 examples/s]
Map:  59%|█████▉    | 3392/5758 [00:01<00:00, 3412.00 examples/s]
Map:  65%|██████▍   | 3741/5758 [00:01<00:00, 3429.19 examples/s]
Map:  74%|███████▍  | 4248/5758 [00:01<00:00, 3404.45 examples/s]
Map:  83%|████████▎ | 4759/5758 [00:01<00:00, 3401.45 examples/s]
Map:  92%|█████████▏| 5271/5758 [00:01<00:00, 3400.56 examples/s]
Map:  98%|█████████▊| 5627/5758 [00:01<00:00, 3436.53 examples/s]
Map: 100%|██████████| 5758/5758 [00:01<00:00, 3386.88 examples/s]
+Number of training samples: 5758
+
Adding EOS to train dataset:   0%|          | 0/5758 [00:00<?, ? examples/s]
Adding EOS to train dataset: 100%|██████████| 5758/5758 [00:00<00:00, 58296.41 examples/s]
+
Tokenizing train dataset:   0%|          | 0/5758 [00:00<?, ? examples/s]
Tokenizing train dataset:   2%|▏         | 108/5758 [00:00<00:05, 1059.23 examples/s]
Tokenizing train dataset:   4%|▍         | 219/5758 [00:00<00:05, 1083.34 examples/s]
Tokenizing train dataset:   6%|▌         | 337/5758 [00:00<00:04, 1123.12 examples/s]
Tokenizing train dataset:   9%|▊         | 498/5758 [00:00<00:04, 1084.67 examples/s]
Tokenizing train dataset:  11%|█▏        | 650/5758 [00:00<00:04, 1054.07 examples/s]
Tokenizing train dataset:  13%|█▎        | 760/5758 [00:00<00:04, 1064.72 examples/s]
Tokenizing train dataset:  16%|█▌        | 923/5758 [00:00<00:04, 1069.50 examples/s]
Tokenizing train dataset:  18%|█▊        | 1048/5758 [00:01<00:04, 947.70 examples/s]
Tokenizing train dataset:  20%|██        | 1180/5758 [00:01<00:04, 922.41 examples/s]
Tokenizing train dataset:  22%|██▏       | 1277/5758 [00:01<00:04, 931.08 examples/s]
Tokenizing train dataset:  24%|██▍       | 1382/5758 [00:01<00:04, 956.55 examples/s]
Tokenizing train dataset:  27%|██▋       | 1544/5758 [00:01<00:04, 991.26 examples/s]
Tokenizing train dataset:  29%|██▉       | 1662/5758 [00:01<00:03, 1034.81 examples/s]
Tokenizing train dataset:  31%|███       | 1775/5758 [00:01<00:03, 1046.16 examples/s]
Tokenizing train dataset:  33%|███▎      | 1882/5758 [00:01<00:03, 1050.01 examples/s]
Tokenizing train dataset:  35%|███▍      | 1994/5758 [00:01<00:03, 1066.71 examples/s]
Tokenizing train dataset:  37%|███▋      | 2107/5758 [00:02<00:03, 946.83 examples/s] 
Tokenizing train dataset:  38%|███▊      | 2210/5758 [00:02<00:03, 965.94 examples/s]
Tokenizing train dataset:  41%|████      | 2356/5758 [00:02<00:03, 967.15 examples/s]
Tokenizing train dataset:  43%|████▎     | 2460/5758 [00:02<00:03, 982.53 examples/s]
Tokenizing train dataset:  45%|████▍     | 2566/5758 [00:02<00:03, 1000.47 examples/s]
Tokenizing train dataset:  46%|████▋     | 2670/5758 [00:02<00:03, 1007.49 examples/s]
Tokenizing train dataset:  48%|████▊     | 2774/5758 [00:02<00:02, 1012.54 examples/s]
Tokenizing train dataset:  50%|█████     | 2899/5758 [00:02<00:02, 1077.12 examples/s]
Tokenizing train dataset:  53%|█████▎    | 3048/5758 [00:03<00:02, 961.74 examples/s] 
Tokenizing train dataset:  55%|█████▌    | 3183/5758 [00:03<00:02, 937.15 examples/s]
Tokenizing train dataset:  57%|█████▋    | 3287/5758 [00:03<00:02, 958.48 examples/s]
Tokenizing train dataset:  59%|█████▉    | 3392/5758 [00:03<00:02, 981.10 examples/s]
Tokenizing train dataset:  61%|██████    | 3502/5758 [00:03<00:02, 1007.26 examples/s]
Tokenizing train dataset:  63%|██████▎   | 3648/5758 [00:03<00:02, 990.49 examples/s] 
Tokenizing train dataset:  65%|██████▌   | 3757/5758 [00:03<00:01, 1015.47 examples/s]
Tokenizing train dataset:  67%|██████▋   | 3865/5758 [00:03<00:01, 1029.23 examples/s]
Tokenizing train dataset:  69%|██████▉   | 3974/5758 [00:03<00:01, 1044.28 examples/s]
Tokenizing train dataset:  71%|███████   | 4099/5758 [00:04<00:02, 821.99 examples/s] 
Tokenizing train dataset:  74%|███████▎  | 4235/5758 [00:04<00:01, 845.17 examples/s]
Tokenizing train dataset:  76%|███████▌  | 4366/5758 [00:04<00:01, 850.74 examples/s]
Tokenizing train dataset:  77%|███████▋  | 4462/5758 [00:04<00:01, 872.76 examples/s]
Tokenizing train dataset:  79%|███████▉  | 4565/5758 [00:04<00:01, 904.98 examples/s]
Tokenizing train dataset:  81%|████████  | 4669/5758 [00:04<00:01, 934.76 examples/s]
Tokenizing train dataset:  83%|████████▎ | 4777/5758 [00:04<00:01, 971.68 examples/s]
Tokenizing train dataset:  85%|████████▌ | 4899/5758 [00:04<00:00, 1034.32 examples/s]
Tokenizing train dataset:  88%|████████▊ | 5044/5758 [00:05<00:00, 912.25 examples/s] 
Tokenizing train dataset:  90%|█████████ | 5183/5758 [00:05<00:00, 915.48 examples/s]
Tokenizing train dataset:  92%|█████████▏| 5281/5758 [00:05<00:00, 927.79 examples/s]
Tokenizing train dataset:  93%|█████████▎| 5381/5758 [00:05<00:00, 943.56 examples/s]
Tokenizing train dataset:  95%|█████████▌| 5478/5758 [00:05<00:00, 947.55 examples/s]
Tokenizing train dataset:  97%|█████████▋| 5582/5758 [00:05<00:00, 971.00 examples/s]
Tokenizing train dataset:  99%|█████████▉| 5689/5758 [00:05<00:00, 996.69 examples/s]
Tokenizing train dataset: 100%|██████████| 5758/5758 [00:05<00:00, 971.03 examples/s]
+
Truncating train dataset:   0%|          | 0/5758 [00:00<?, ? examples/s]
Truncating train dataset: 100%|██████████| 5758/5758 [00:00<00:00, 688194.30 examples/s]
+The tokenizer has new PAD/BOS/EOS tokens that differ from the model config and generation config. The model config and generation config were aligned accordingly, being updated with the tokenizer's values. Updated tokens: {'bos_token_id': None, 'pad_token_id': 151645}.
+Qwen3-0.6B FULL FT | Single GPU
+Dry run...
+
  0%|          | 0/5758 [00:00<?, ?it/s]
  0%|          | 1/5758 [00:01<2:05:11,  1.30s/it]
                                                  

  0%|          | 1/5758 [00:01<2:05:11,  1.30s/it]
  0%|          | 2/5758 [00:01<1:11:35,  1.34it/s]
                                                  

  0%|          | 2/5758 [00:01<1:11:35,  1.34it/s]
  0%|          | 3/5758 [00:01<52:51,  1.81it/s]  
                                                

  0%|          | 3/5758 [00:01<52:51,  1.81it/s]
  0%|          | 4/5758 [00:02<44:48,  2.14it/s]
                                                

  0%|          | 4/5758 [00:02<44:48,  2.14it/s]
  0%|          | 5/5758 [00:02<39:02,  2.46it/s]
                                                

  0%|          | 5/5758 [00:02<39:02,  2.46it/s]
  0%|          | 6/5758 [00:02<35:31,  2.70it/s]
                                                

  0%|          | 6/5758 [00:02<35:31,  2.70it/s]
  0%|          | 7/5758 [00:03<35:08,  2.73it/s]
                                                

  0%|          | 7/5758 [00:03<35:08,  2.73it/s]
  0%|          | 8/5758 [00:03<34:04,  2.81it/s]
                                                

  0%|          | 8/5758 [00:03<34:04,  2.81it/s]
  0%|          | 9/5758 [00:03<33:16,  2.88it/s]
                                                

  0%|          | 9/5758 [00:03<33:16,  2.88it/s]
  0%|          | 10/5758 [00:04<31:53,  3.00it/s]
                                                 

  0%|          | 10/5758 [00:04<31:53,  3.00it/s]
  0%|          | 11/5758 [00:04<30:31,  3.14it/s]
                                                 

  0%|          | 11/5758 [00:04<30:31,  3.14it/s]
  0%|          | 12/5758 [00:04<31:42,  3.02it/s]
                                                 

  0%|          | 12/5758 [00:04<31:42,  3.02it/s]
  0%|          | 13/5758 [00:05<33:15,  2.88it/s]
                                                 

  0%|          | 13/5758 [00:05<33:15,  2.88it/s]
  0%|          | 14/5758 [00:05<31:25,  3.05it/s]
                                                 

  0%|          | 14/5758 [00:05<31:25,  3.05it/s]
  0%|          | 15/5758 [00:05<30:36,  3.13it/s]
                                                 

  0%|          | 15/5758 [00:05<30:36,  3.13it/s]
  0%|          | 16/5758 [00:06<31:31,  3.04it/s]
                                                 

  0%|          | 16/5758 [00:06<31:31,  3.04it/s]
  0%|          | 17/5758 [00:06<33:26,  2.86it/s]
                                                 

  0%|          | 17/5758 [00:06<33:26,  2.86it/s]
  0%|          | 18/5758 [00:06<34:12,  2.80it/s]
                                                 

  0%|          | 18/5758 [00:06<34:12,  2.80it/s]
  0%|          | 19/5758 [00:07<34:44,  2.75it/s]
                                                 

  0%|          | 19/5758 [00:07<34:44,  2.75it/s]
  0%|          | 20/5758 [00:07<34:59,  2.73it/s]
                                                 

  0%|          | 20/5758 [00:07<34:59,  2.73it/s]
  0%|          | 21/5758 [00:08<34:53,  2.74it/s]
                                                 

  0%|          | 21/5758 [00:08<34:53,  2.74it/s]
  0%|          | 22/5758 [00:08<36:57,  2.59it/s]
                                                 

  0%|          | 22/5758 [00:08<36:57,  2.59it/s]
  0%|          | 23/5758 [00:08<35:51,  2.67it/s]
                                                 

  0%|          | 23/5758 [00:08<35:51,  2.67it/s]
  0%|          | 24/5758 [00:09<33:53,  2.82it/s]
                                                 

  0%|          | 24/5758 [00:09<33:53,  2.82it/s]
  0%|          | 25/5758 [00:09<32:07,  2.97it/s]
                                                 

  0%|          | 25/5758 [00:09<32:07,  2.97it/s]
  0%|          | 26/5758 [00:09<31:42,  3.01it/s]
                                                 

  0%|          | 26/5758 [00:09<31:42,  3.01it/s]
  0%|          | 27/5758 [00:10<31:57,  2.99it/s]
                                                 

  0%|          | 27/5758 [00:10<31:57,  2.99it/s]
  0%|          | 28/5758 [00:10<30:57,  3.09it/s]
                                                 

  0%|          | 28/5758 [00:10<30:57,  3.09it/s]
  1%|          | 29/5758 [00:10<31:37,  3.02it/s]
                                                 

  1%|          | 29/5758 [00:10<31:37,  3.02it/s]
  1%|          | 30/5758 [00:11<31:37,  3.02it/s]
                                                 

  1%|          | 30/5758 [00:11<31:37,  3.02it/s]
  1%|          | 31/5758 [00:11<31:38,  3.02it/s]
                                                 

  1%|          | 31/5758 [00:11<31:38,  3.02it/s]
  1%|          | 32/5758 [00:11<30:38,  3.11it/s]
                                                 

  1%|          | 32/5758 [00:11<30:38,  3.11it/s]
  1%|          | 33/5758 [00:12<30:46,  3.10it/s]
                                                 

  1%|          | 33/5758 [00:12<30:46,  3.10it/s]
  1%|          | 34/5758 [00:12<29:44,  3.21it/s]
                                                 

  1%|          | 34/5758 [00:12<29:44,  3.21it/s]
  1%|          | 35/5758 [00:12<29:04,  3.28it/s]
                                                 

  1%|          | 35/5758 [00:12<29:04,  3.28it/s]
  1%|          | 36/5758 [00:12<28:39,  3.33it/s]
                                                 

  1%|          | 36/5758 [00:12<28:39,  3.33it/s]
  1%|          | 37/5758 [00:13<29:00,  3.29it/s]
                                                 

  1%|          | 37/5758 [00:13<29:00,  3.29it/s]
  1%|          | 38/5758 [00:13<30:03,  3.17it/s]
                                                 

  1%|          | 38/5758 [00:13<30:03,  3.17it/s]
  1%|          | 39/5758 [00:13<30:01,  3.18it/s]
                                                 

  1%|          | 39/5758 [00:13<30:01,  3.18it/s]
  1%|          | 40/5758 [00:14<29:19,  3.25it/s]
                                                 

  1%|          | 40/5758 [00:14<29:19,  3.25it/s]
  1%|          | 41/5758 [00:14<30:25,  3.13it/s]
                                                 

  1%|          | 41/5758 [00:14<30:25,  3.13it/s]
  1%|          | 42/5758 [00:14<30:17,  3.14it/s]
                                                 

  1%|          | 42/5758 [00:14<30:17,  3.14it/s]
  1%|          | 43/5758 [00:15<31:30,  3.02it/s]
                                                 

  1%|          | 43/5758 [00:15<31:30,  3.02it/s]
  1%|          | 44/5758 [00:15<32:27,  2.93it/s]
                                                 

  1%|          | 44/5758 [00:15<32:27,  2.93it/s]
  1%|          | 45/5758 [00:15<32:16,  2.95it/s]
                                                 

  1%|          | 45/5758 [00:15<32:16,  2.95it/s]
  1%|          | 46/5758 [00:16<30:55,  3.08it/s]
                                                 

  1%|          | 46/5758 [00:16<30:55,  3.08it/s]
  1%|          | 47/5758 [00:16<30:23,  3.13it/s]
                                                 

  1%|          | 47/5758 [00:16<30:23,  3.13it/s]
  1%|          | 48/5758 [00:17<35:38,  2.67it/s]
                                                 

  1%|          | 48/5758 [00:17<35:38,  2.67it/s]
  1%|          | 49/5758 [00:17<34:08,  2.79it/s]
                                                 

  1%|          | 49/5758 [00:17<34:08,  2.79it/s]
  1%|          | 50/5758 [00:17<32:42,  2.91it/s]
                                                 

  1%|          | 50/5758 [00:17<32:42,  2.91it/s]
  1%|          | 51/5758 [00:17<31:17,  3.04it/s]
                                                 

  1%|          | 51/5758 [00:17<31:17,  3.04it/s]
  1%|          | 52/5758 [00:18<31:42,  3.00it/s]
                                                 

  1%|          | 52/5758 [00:18<31:42,  3.00it/s]
  1%|          | 53/5758 [00:18<32:22,  2.94it/s]
                                                 

  1%|          | 53/5758 [00:18<32:22,  2.94it/s]
  1%|          | 54/5758 [00:18<31:34,  3.01it/s]
                                                 

  1%|          | 54/5758 [00:18<31:34,  3.01it/s]
  1%|          | 55/5758 [00:19<32:59,  2.88it/s]
                                                 

  1%|          | 55/5758 [00:19<32:59,  2.88it/s]
  1%|          | 56/5758 [00:19<31:22,  3.03it/s]
                                                 
{'loss': '3.121', 'grad_norm': '54', 'learning_rate': '1e-05', 'num_tokens': '514', 'mean_token_accuracy': '0.5029', 'epoch': '0.0001737'}
+{'loss': '2.244', 'grad_norm': '46.5', 'learning_rate': '9.998e-06', 'num_tokens': '1197', 'mean_token_accuracy': '0.566', 'epoch': '0.0003473'}
+{'loss': '4.943', 'grad_norm': '106', 'learning_rate': '9.997e-06', 'num_tokens': '1278', 'mean_token_accuracy': '0.3', 'epoch': '0.000521'}
+{'loss': '1.945', 'grad_norm': '22.38', 'learning_rate': '9.995e-06', 'num_tokens': '2126', 'mean_token_accuracy': '0.6328', 'epoch': '0.0006947'}
+{'loss': '2.9', 'grad_norm': '42', 'learning_rate': '9.993e-06', 'num_tokens': '2281', 'mean_token_accuracy': '0.513', 'epoch': '0.0008684'}
+{'loss': '1.037', 'grad_norm': '12.75', 'learning_rate': '9.991e-06', 'num_tokens': '3549', 'mean_token_accuracy': '0.7395', 'epoch': '0.001042'}
+{'loss': '3.215', 'grad_norm': '23.38', 'learning_rate': '9.99e-06', 'num_tokens': '4004', 'mean_token_accuracy': '0.4604', 'epoch': '0.001216'}
+{'loss': '3.812', 'grad_norm': '36.25', 'learning_rate': '9.988e-06', 'num_tokens': '4216', 'mean_token_accuracy': '0.3744', 'epoch': '0.001389'}
+{'loss': '2.052', 'grad_norm': '17.12', 'learning_rate': '9.986e-06', 'num_tokens': '5075', 'mean_token_accuracy': '0.5186', 'epoch': '0.001563'}
+{'loss': '3.176', 'grad_norm': '36.25', 'learning_rate': '9.984e-06', 'num_tokens': '5242', 'mean_token_accuracy': '0.4217', 'epoch': '0.001737'}
+{'loss': '1.098', 'grad_norm': '12.94', 'learning_rate': '9.983e-06', 'num_tokens': '6460', 'mean_token_accuracy': '0.7075', 'epoch': '0.00191'}
+{'loss': '1.613', 'grad_norm': '11.19', 'learning_rate': '9.981e-06', 'num_tokens': '8415', 'mean_token_accuracy': '0.6668', 'epoch': '0.002084'}
+{'loss': '3.08', 'grad_norm': '34.25', 'learning_rate': '9.979e-06', 'num_tokens': '8549', 'mean_token_accuracy': '0.4511', 'epoch': '0.002258'}
+{'loss': '2.748', 'grad_norm': '32.5', 'learning_rate': '9.977e-06', 'num_tokens': '8699', 'mean_token_accuracy': '0.4295', 'epoch': '0.002431'}
+{'loss': '1.63', 'grad_norm': '14.81', 'learning_rate': '9.976e-06', 'num_tokens': '9338', 'mean_token_accuracy': '0.5721', 'epoch': '0.002605'}
+{'loss': '1.575', 'grad_norm': '10.38', 'learning_rate': '9.974e-06', 'num_tokens': '1.115e+04', 'mean_token_accuracy': '0.6431', 'epoch': '0.002779'}
+{'loss': '1.656', 'grad_norm': '20.75', 'learning_rate': '9.972e-06', 'num_tokens': '1.144e+04', 'mean_token_accuracy': '0.619', 'epoch': '0.002952'}
+{'loss': '2.093', 'grad_norm': '23.25', 'learning_rate': '9.97e-06', 'num_tokens': '1.171e+04', 'mean_token_accuracy': '0.588', 'epoch': '0.003126'}
+{'loss': '1.838', 'grad_norm': '17', 'learning_rate': '9.969e-06', 'num_tokens': '1.226e+04', 'mean_token_accuracy': '0.5509', 'epoch': '0.0033'}
+{'loss': '2.525', 'grad_norm': '20.5', 'learning_rate': '9.967e-06', 'num_tokens': '1.279e+04', 'mean_token_accuracy': '0.5369', 'epoch': '0.003473'}
+{'loss': '2.361', 'grad_norm': '20.38', 'learning_rate': '9.965e-06', 'num_tokens': '1.331e+04', 'mean_token_accuracy': '0.5742', 'epoch': '0.003647'}
+{'loss': '1.525', 'grad_norm': '8.812', 'learning_rate': '9.964e-06', 'num_tokens': '1.476e+04', 'mean_token_accuracy': '0.6862', 'epoch': '0.003821'}
+{'loss': '2.947', 'grad_norm': '18.62', 'learning_rate': '9.962e-06', 'num_tokens': '1.521e+04', 'mean_token_accuracy': '0.4742', 'epoch': '0.003994'}
+{'loss': '1.687', 'grad_norm': '13.75', 'learning_rate': '9.96e-06', 'num_tokens': '1.586e+04', 'mean_token_accuracy': '0.5864', 'epoch': '0.004168'}
+{'loss': '1.466', 'grad_norm': '13.62', 'learning_rate': '9.958e-06', 'num_tokens': '1.649e+04', 'mean_token_accuracy': '0.6248', 'epoch': '0.004342'}
+{'loss': '1.783', 'grad_norm': '13.56', 'learning_rate': '9.957e-06', 'num_tokens': '1.727e+04', 'mean_token_accuracy': '0.5638', 'epoch': '0.004515'}
+{'loss': '3.403', 'grad_norm': '27.25', 'learning_rate': '9.955e-06', 'num_tokens': '1.75e+04', 'mean_token_accuracy': '0.4063', 'epoch': '0.004689'}
+{'loss': '1.357', 'grad_norm': '11.94', 'learning_rate': '9.953e-06', 'num_tokens': '1.818e+04', 'mean_token_accuracy': '0.6164', 'epoch': '0.004863'}
+{'loss': '2.692', 'grad_norm': '23.38', 'learning_rate': '9.951e-06', 'num_tokens': '1.842e+04', 'mean_token_accuracy': '0.5021', 'epoch': '0.005036'}
+{'loss': '1.698', 'grad_norm': '11.56', 'learning_rate': '9.95e-06', 'num_tokens': '1.923e+04', 'mean_token_accuracy': '0.6414', 'epoch': '0.00521'}
+{'loss': '3.115', 'grad_norm': '24.12', 'learning_rate': '9.948e-06', 'num_tokens': '1.945e+04', 'mean_token_accuracy': '0.4612', 'epoch': '0.005384'}
+{'loss': '2.13', 'grad_norm': '13.25', 'learning_rate': '9.946e-06', 'num_tokens': '2.01e+04', 'mean_token_accuracy': '0.595', 'epoch': '0.005557'}
+{'loss': '2.579', 'grad_norm': '26.5', 'learning_rate': '9.944e-06', 'num_tokens': '2.031e+04', 'mean_token_accuracy': '0.5337', 'epoch': '0.005731'}
+{'loss': '0.9654', 'grad_norm': '9.062', 'learning_rate': '9.943e-06', 'num_tokens': '2.156e+04', 'mean_token_accuracy': '0.7159', 'epoch': '0.005905'}
+{'loss': '1.789', 'grad_norm': '21.88', 'learning_rate': '9.941e-06', 'num_tokens': '2.172e+04', 'mean_token_accuracy': '0.6581', 'epoch': '0.006078'}
+{'loss': '1.62', 'grad_norm': '13', 'learning_rate': '9.939e-06', 'num_tokens': '2.235e+04', 'mean_token_accuracy': '0.5965', 'epoch': '0.006252'}
+{'loss': '2.643', 'grad_norm': '21.62', 'learning_rate': '9.937e-06', 'num_tokens': '2.254e+04', 'mean_token_accuracy': '0.5355', 'epoch': '0.006426'}
+{'loss': '1.901', 'grad_norm': '9.438', 'learning_rate': '9.936e-06', 'num_tokens': '2.438e+04', 'mean_token_accuracy': '0.6042', 'epoch': '0.0066'}
+{'loss': '1.737', 'grad_norm': '12.62', 'learning_rate': '9.934e-06', 'num_tokens': '2.513e+04', 'mean_token_accuracy': '0.5667', 'epoch': '0.006773'}
+{'loss': '1.76', 'grad_norm': '14', 'learning_rate': '9.932e-06', 'num_tokens': '2.578e+04', 'mean_token_accuracy': '0.5672', 'epoch': '0.006947'}
+{'loss': '1.681', 'grad_norm': '11.19', 'learning_rate': '9.931e-06', 'num_tokens': '2.669e+04', 'mean_token_accuracy': '0.6284', 'epoch': '0.007121'}
+{'loss': '2.317', 'grad_norm': '15.81', 'learning_rate': '9.929e-06', 'num_tokens': '2.706e+04', 'mean_token_accuracy': '0.5815', 'epoch': '0.007294'}
+{'loss': '2.36', 'grad_norm': '16.25', 'learning_rate': '9.927e-06', 'num_tokens': '2.758e+04', 'mean_token_accuracy': '0.5624', 'epoch': '0.007468'}
+{'loss': '1.414', 'grad_norm': '10', 'learning_rate': '9.925e-06', 'num_tokens': '2.857e+04', 'mean_token_accuracy': '0.6721', 'epoch': '0.007642'}
+{'loss': '1.83', 'grad_norm': '13.94', 'learning_rate': '9.924e-06', 'num_tokens': '2.902e+04', 'mean_token_accuracy': '0.644', 'epoch': '0.007815'}
+{'loss': '3.106', 'grad_norm': '42.5', 'learning_rate': '9.922e-06', 'num_tokens': '2.909e+04', 'mean_token_accuracy': '0.5362', 'epoch': '0.007989'}
+{'loss': '2.637', 'grad_norm': '14.12', 'learning_rate': '9.92e-06', 'num_tokens': '2.982e+04', 'mean_token_accuracy': '0.4917', 'epoch': '0.008163'}
+{'loss': '1.642', 'grad_norm': '7.406', 'learning_rate': '9.918e-06', 'num_tokens': '3.202e+04', 'mean_token_accuracy': '0.651', 'epoch': '0.008336'}
+{'loss': '1.751', 'grad_norm': '11.19', 'learning_rate': '9.917e-06', 'num_tokens': '3.284e+04', 'mean_token_accuracy': '0.5721', 'epoch': '0.00851'}
+{'loss': '2.928', 'grad_norm': '32.75', 'learning_rate': '9.915e-06', 'num_tokens': '3.292e+04', 'mean_token_accuracy': '0.5122', 'epoch': '0.008684'}
+{'loss': '1.528', 'grad_norm': '12', 'learning_rate': '9.913e-06', 'num_tokens': '3.358e+04', 'mean_token_accuracy': '0.5982', 'epoch': '0.008857'}
+{'loss': '2.584', 'grad_norm': '21', 'learning_rate': '9.911e-06', 'num_tokens': '3.382e+04', 'mean_token_accuracy': '0.5322', 'epoch': '0.009031'}
+{'loss': '2.314', 'grad_norm': '34.75', 'learning_rate': '9.91e-06', 'num_tokens': '3.393e+04', 'mean_token_accuracy': '0.5545', 'epoch': '0.009205'}
+{'loss': '2.651', 'grad_norm': '16.62', 'learning_rate': '9.908e-06', 'num_tokens': '3.431e+04', 'mean_token_accuracy': '0.4908', 'epoch': '0.009378'}
+{'loss': '2.302', 'grad_norm': '26.5', 'learning_rate': '9.906e-06', 'num_tokens': '3.444e+04', 'mean_token_accuracy': '0.5985', 'epoch': '0.009552'}
+
  1%|          | 56/5758 [00:19<31:22,  3.03it/s]
  1%|          | 57/5758 [00:20<32:51,  2.89it/s]
                                                 

  1%|          | 57/5758 [00:20<32:51,  2.89it/s]
  1%|          | 58/5758 [00:20<32:25,  2.93it/s]
                                                 

  1%|          | 58/5758 [00:20<32:25,  2.93it/s]
  1%|          | 59/5758 [00:20<31:50,  2.98it/s]
                                                 

  1%|          | 59/5758 [00:20<31:50,  2.98it/s]
  1%|          | 60/5758 [00:20<30:58,  3.07it/s]
                                                 

  1%|          | 60/5758 [00:20<30:58,  3.07it/s]
  1%|          | 61/5758 [00:21<30:08,  3.15it/s]
                                                 

  1%|          | 61/5758 [00:21<30:08,  3.15it/s]
  1%|          | 62/5758 [00:21<31:01,  3.06it/s]
                                                 

  1%|          | 62/5758 [00:21<31:01,  3.06it/s]
  1%|          | 63/5758 [00:21<29:50,  3.18it/s]
                                                 

  1%|          | 63/5758 [00:21<29:50,  3.18it/s]
  1%|          | 64/5758 [00:22<29:59,  3.16it/s]
                                                 

  1%|          | 64/5758 [00:22<29:59,  3.16it/s]
  1%|          | 65/5758 [00:22<30:12,  3.14it/s]
                                                 

  1%|          | 65/5758 [00:22<30:12,  3.14it/s]
  1%|          | 66/5758 [00:22<30:20,  3.13it/s]
                                                 

  1%|          | 66/5758 [00:22<30:20,  3.13it/s]
  1%|          | 67/5758 [00:23<30:59,  3.06it/s]
                                                 

  1%|          | 67/5758 [00:23<30:59,  3.06it/s]
  1%|          | 68/5758 [00:23<30:13,  3.14it/s]
                                                 

  1%|          | 68/5758 [00:23<30:13,  3.14it/s]
  1%|          | 69/5758 [00:23<32:09,  2.95it/s]
                                                 

  1%|          | 69/5758 [00:23<32:09,  2.95it/s]
  1%|          | 70/5758 [00:24<33:55,  2.79it/s]
                                                 

  1%|          | 70/5758 [00:24<33:55,  2.79it/s]
  1%|          | 71/5758 [00:24<34:40,  2.73it/s]
                                                 

  1%|          | 71/5758 [00:24<34:40,  2.73it/s]
  1%|▏         | 72/5758 [00:24<32:31,  2.91it/s]
                                                 

  1%|▏         | 72/5758 [00:24<32:31,  2.91it/s]
  1%|▏         | 73/5758 [00:25<32:47,  2.89it/s]
                                                 

  1%|▏         | 73/5758 [00:25<32:47,  2.89it/s]
  1%|▏         | 74/5758 [00:25<35:39,  2.66it/s]
                                                 

  1%|▏         | 74/5758 [00:25<35:39,  2.66it/s]
  1%|▏         | 75/5758 [00:26<33:47,  2.80it/s]
                                                 

  1%|▏         | 75/5758 [00:26<33:47,  2.80it/s]
  1%|▏         | 76/5758 [00:26<32:11,  2.94it/s]
                                                 

  1%|▏         | 76/5758 [00:26<32:11,  2.94it/s]
  1%|▏         | 77/5758 [00:26<33:17,  2.84it/s]
                                                 

  1%|▏         | 77/5758 [00:26<33:17,  2.84it/s]
  1%|▏         | 78/5758 [00:27<32:39,  2.90it/s]
                                                 

  1%|▏         | 78/5758 [00:27<32:39,  2.90it/s]
  1%|▏         | 79/5758 [00:27<33:05,  2.86it/s]
                                                 

  1%|▏         | 79/5758 [00:27<33:05,  2.86it/s]
  1%|▏         | 80/5758 [00:27<32:33,  2.91it/s]
                                                 

  1%|▏         | 80/5758 [00:27<32:33,  2.91it/s]
  1%|▏         | 81/5758 [00:28<31:05,  3.04it/s]
                                                 

  1%|▏         | 81/5758 [00:28<31:05,  3.04it/s]
  1%|▏         | 82/5758 [00:28<32:56,  2.87it/s]
                                                 

  1%|▏         | 82/5758 [00:28<32:56,  2.87it/s]
  1%|▏         | 83/5758 [00:28<32:30,  2.91it/s]
                                                 

  1%|▏         | 83/5758 [00:28<32:30,  2.91it/s]
  1%|▏         | 84/5758 [00:29<31:00,  3.05it/s]
                                                 

  1%|▏         | 84/5758 [00:29<31:00,  3.05it/s]
  1%|▏         | 85/5758 [00:29<30:17,  3.12it/s]
                                                 

  1%|▏         | 85/5758 [00:29<30:17,  3.12it/s]
  1%|▏         | 86/5758 [00:29<30:24,  3.11it/s]
                                                 

  1%|▏         | 86/5758 [00:29<30:24,  3.11it/s]
  2%|▏         | 87/5758 [00:30<30:11,  3.13it/s]
                                                 

  2%|▏         | 87/5758 [00:30<30:11,  3.13it/s]
  2%|▏         | 88/5758 [00:30<31:03,  3.04it/s]
                                                 

  2%|▏         | 88/5758 [00:30<31:03,  3.04it/s]
  2%|▏         | 89/5758 [00:30<31:02,  3.04it/s]
                                                 

  2%|▏         | 89/5758 [00:30<31:02,  3.04it/s]
  2%|▏         | 90/5758 [00:31<30:27,  3.10it/s]
                                                 

  2%|▏         | 90/5758 [00:31<30:27,  3.10it/s]
  2%|▏         | 91/5758 [00:31<29:23,  3.21it/s]
                                                 

  2%|▏         | 91/5758 [00:31<29:23,  3.21it/s]
  2%|▏         | 92/5758 [00:31<29:25,  3.21it/s]
                                                 

  2%|▏         | 92/5758 [00:31<29:25,  3.21it/s]
  2%|▏         | 93/5758 [00:31<28:52,  3.27it/s]
                                                 

  2%|▏         | 93/5758 [00:31<28:52,  3.27it/s]
  2%|▏         | 94/5758 [00:32<28:37,  3.30it/s]
                                                 

  2%|▏         | 94/5758 [00:32<28:37,  3.30it/s]
  2%|▏         | 95/5758 [00:32<29:56,  3.15it/s]
                                                 

  2%|▏         | 95/5758 [00:32<29:56,  3.15it/s]
  2%|▏         | 96/5758 [00:32<31:07,  3.03it/s]
                                                 

  2%|▏         | 96/5758 [00:32<31:07,  3.03it/s]
  2%|▏         | 97/5758 [00:33<31:37,  2.98it/s]
                                                 

  2%|▏         | 97/5758 [00:33<31:37,  2.98it/s]
  2%|▏         | 98/5758 [00:33<31:29,  2.99it/s]
                                                 

  2%|▏         | 98/5758 [00:33<31:29,  2.99it/s]
  2%|▏         | 99/5758 [00:33<31:32,  2.99it/s]
                                                 

  2%|▏         | 99/5758 [00:33<31:32,  2.99it/s]
  2%|▏         | 100/5758 [00:34<30:47,  3.06it/s]
                                                  

  2%|▏         | 100/5758 [00:34<30:47,  3.06it/s]
  2%|▏         | 101/5758 [00:34<33:04,  2.85it/s]
                                                  

  2%|▏         | 101/5758 [00:34<33:04,  2.85it/s]
  2%|▏         | 102/5758 [00:35<33:15,  2.83it/s]
                                                  

  2%|▏         | 102/5758 [00:35<33:15,  2.83it/s]
  2%|▏         | 103/5758 [00:35<31:45,  2.97it/s]
                                                  

  2%|▏         | 103/5758 [00:35<31:45,  2.97it/s]
  2%|▏         | 104/5758 [00:35<31:21,  3.00it/s]
                                                  

  2%|▏         | 104/5758 [00:35<31:21,  3.00it/s]
  2%|▏         | 105/5758 [00:36<32:41,  2.88it/s]
                                                  

  2%|▏         | 105/5758 [00:36<32:41,  2.88it/s]
  2%|▏         | 106/5758 [00:36<31:14,  3.02it/s]
                                                  

  2%|▏         | 106/5758 [00:36<31:14,  3.02it/s]
  2%|▏         | 107/5758 [00:36<30:05,  3.13it/s]
                                                  

  2%|▏         | 107/5758 [00:36<30:05,  3.13it/s]
  2%|▏         | 108/5758 [00:36<29:52,  3.15it/s]
                                                  

  2%|▏         | 108/5758 [00:36<29:52,  3.15it/s]
  2%|▏         | 109/5758 [00:37<30:10,  3.12it/s]
                                                  

  2%|▏         | 109/5758 [00:37<30:10,  3.12it/s]
  2%|▏         | 110/5758 [00:37<31:01,  3.03it/s]
                                                  
{'loss': '1.822', 'grad_norm': '13.56', 'learning_rate': '9.904e-06', 'num_tokens': '3.504e+04', 'mean_token_accuracy': '0.5373', 'epoch': '0.009726'}
+{'loss': '1.723', 'grad_norm': '13.56', 'learning_rate': '9.903e-06', 'num_tokens': '3.562e+04', 'mean_token_accuracy': '0.5801', 'epoch': '0.009899'}
+{'loss': '1.703', 'grad_norm': '12.81', 'learning_rate': '9.901e-06', 'num_tokens': '3.604e+04', 'mean_token_accuracy': '0.6827', 'epoch': '0.01007'}
+{'loss': '1.718', 'grad_norm': '9.125', 'learning_rate': '9.899e-06', 'num_tokens': '3.758e+04', 'mean_token_accuracy': '0.6357', 'epoch': '0.01025'}
+{'loss': '2.407', 'grad_norm': '13.12', 'learning_rate': '9.898e-06', 'num_tokens': '3.825e+04', 'mean_token_accuracy': '0.5574', 'epoch': '0.01042'}
+{'loss': '1.563', 'grad_norm': '11.94', 'learning_rate': '9.896e-06', 'num_tokens': '3.892e+04', 'mean_token_accuracy': '0.5797', 'epoch': '0.01059'}
+{'loss': '2.33', 'grad_norm': '19.38', 'learning_rate': '9.894e-06', 'num_tokens': '3.916e+04', 'mean_token_accuracy': '0.5378', 'epoch': '0.01077'}
+{'loss': '1.671', 'grad_norm': '12.88', 'learning_rate': '9.892e-06', 'num_tokens': '3.976e+04', 'mean_token_accuracy': '0.5646', 'epoch': '0.01094'}
+{'loss': '1.601', 'grad_norm': '14.5', 'learning_rate': '9.891e-06', 'num_tokens': '4.015e+04', 'mean_token_accuracy': '0.6501', 'epoch': '0.01111'}
+{'loss': '1.611', 'grad_norm': '11.44', 'learning_rate': '9.889e-06', 'num_tokens': '4.094e+04', 'mean_token_accuracy': '0.5849', 'epoch': '0.01129'}
+{'loss': '2.313', 'grad_norm': '32.5', 'learning_rate': '9.887e-06', 'num_tokens': '4.104e+04', 'mean_token_accuracy': '0.6022', 'epoch': '0.01146'}
+{'loss': '1.305', 'grad_norm': '11.44', 'learning_rate': '9.885e-06', 'num_tokens': '4.198e+04', 'mean_token_accuracy': '0.6454', 'epoch': '0.01164'}
+{'loss': '2.703', 'grad_norm': '16.25', 'learning_rate': '9.884e-06', 'num_tokens': '4.231e+04', 'mean_token_accuracy': '0.4863', 'epoch': '0.01181'}
+{'loss': '2.101', 'grad_norm': '17.12', 'learning_rate': '9.882e-06', 'num_tokens': '4.26e+04', 'mean_token_accuracy': '0.5574', 'epoch': '0.01198'}
+{'loss': '2.257', 'grad_norm': '8.438', 'learning_rate': '9.88e-06', 'num_tokens': '4.497e+04', 'mean_token_accuracy': '0.5545', 'epoch': '0.01216'}
+{'loss': '1.461', 'grad_norm': '8.625', 'learning_rate': '9.878e-06', 'num_tokens': '4.608e+04', 'mean_token_accuracy': '0.6769', 'epoch': '0.01233'}
+{'loss': '1.358', 'grad_norm': '13.19', 'learning_rate': '9.877e-06', 'num_tokens': '4.669e+04', 'mean_token_accuracy': '0.6073', 'epoch': '0.0125'}
+{'loss': '1.719', 'grad_norm': '10', 'learning_rate': '9.875e-06', 'num_tokens': '4.769e+04', 'mean_token_accuracy': '0.6171', 'epoch': '0.01268'}
+{'loss': '1.869', 'grad_norm': '20.12', 'learning_rate': '9.873e-06', 'num_tokens': '4.789e+04', 'mean_token_accuracy': '0.6305', 'epoch': '0.01285'}
+{'loss': '1.734', 'grad_norm': '9.625', 'learning_rate': '9.871e-06', 'num_tokens': '4.926e+04', 'mean_token_accuracy': '0.6184', 'epoch': '0.01303'}
+{'loss': '3.086', 'grad_norm': '26.62', 'learning_rate': '9.87e-06', 'num_tokens': '4.943e+04', 'mean_token_accuracy': '0.4882', 'epoch': '0.0132'}
+{'loss': '1.331', 'grad_norm': '12.19', 'learning_rate': '9.868e-06', 'num_tokens': '5e+04', 'mean_token_accuracy': '0.6471', 'epoch': '0.01337'}
+{'loss': '1.495', 'grad_norm': '12.19', 'learning_rate': '9.866e-06', 'num_tokens': '5.084e+04', 'mean_token_accuracy': '0.5993', 'epoch': '0.01355'}
+{'loss': '1.56', 'grad_norm': '14.38', 'learning_rate': '9.865e-06', 'num_tokens': '5.136e+04', 'mean_token_accuracy': '0.5827', 'epoch': '0.01372'}
+{'loss': '1.484', 'grad_norm': '13.38', 'learning_rate': '9.863e-06', 'num_tokens': '5.178e+04', 'mean_token_accuracy': '0.7014', 'epoch': '0.01389'}
+{'loss': '2.442', 'grad_norm': '17.38', 'learning_rate': '9.861e-06', 'num_tokens': '5.243e+04', 'mean_token_accuracy': '0.5331', 'epoch': '0.01407'}
+{'loss': '2.476', 'grad_norm': '27', 'learning_rate': '9.859e-06', 'num_tokens': '5.257e+04', 'mean_token_accuracy': '0.5674', 'epoch': '0.01424'}
+{'loss': '1.471', 'grad_norm': '11.56', 'learning_rate': '9.858e-06', 'num_tokens': '5.343e+04', 'mean_token_accuracy': '0.6079', 'epoch': '0.01441'}
+{'loss': '1.6', 'grad_norm': '12.62', 'learning_rate': '9.856e-06', 'num_tokens': '5.408e+04', 'mean_token_accuracy': '0.5849', 'epoch': '0.01459'}
+{'loss': '1.078', 'grad_norm': '13.75', 'learning_rate': '9.854e-06', 'num_tokens': '5.482e+04', 'mean_token_accuracy': '0.75', 'epoch': '0.01476'}
+{'loss': '2.535', 'grad_norm': '19.62', 'learning_rate': '9.852e-06', 'num_tokens': '5.503e+04', 'mean_token_accuracy': '0.5481', 'epoch': '0.01494'}
+{'loss': '1.453', 'grad_norm': '8', 'learning_rate': '9.851e-06', 'num_tokens': '5.655e+04', 'mean_token_accuracy': '0.6748', 'epoch': '0.01511'}
+{'loss': '2.577', 'grad_norm': '22', 'learning_rate': '9.849e-06', 'num_tokens': '5.68e+04', 'mean_token_accuracy': '0.498', 'epoch': '0.01528'}
+{'loss': '1.402', 'grad_norm': '12.12', 'learning_rate': '9.847e-06', 'num_tokens': '5.764e+04', 'mean_token_accuracy': '0.6198', 'epoch': '0.01546'}
+{'loss': '1.312', 'grad_norm': '12.31', 'learning_rate': '9.845e-06', 'num_tokens': '5.838e+04', 'mean_token_accuracy': '0.6213', 'epoch': '0.01563'}
+{'loss': '1.746', 'grad_norm': '31.75', 'learning_rate': '9.844e-06', 'num_tokens': '5.853e+04', 'mean_token_accuracy': '0.649', 'epoch': '0.0158'}
+{'loss': '1.539', 'grad_norm': '12.69', 'learning_rate': '9.842e-06', 'num_tokens': '5.924e+04', 'mean_token_accuracy': '0.6037', 'epoch': '0.01598'}
+{'loss': '1.625', 'grad_norm': '14.56', 'learning_rate': '9.84e-06', 'num_tokens': '5.987e+04', 'mean_token_accuracy': '0.6029', 'epoch': '0.01615'}
+{'loss': '1.423', 'grad_norm': '14.56', 'learning_rate': '9.838e-06', 'num_tokens': '6.02e+04', 'mean_token_accuracy': '0.7341', 'epoch': '0.01633'}
+{'loss': '1.793', 'grad_norm': '12', 'learning_rate': '9.837e-06', 'num_tokens': '6.116e+04', 'mean_token_accuracy': '0.6441', 'epoch': '0.0165'}
+{'loss': '1.867', 'grad_norm': '36.5', 'learning_rate': '9.835e-06', 'num_tokens': '6.128e+04', 'mean_token_accuracy': '0.6126', 'epoch': '0.01667'}
+{'loss': '2.108', 'grad_norm': '19.75', 'learning_rate': '9.833e-06', 'num_tokens': '6.151e+04', 'mean_token_accuracy': '0.6026', 'epoch': '0.01685'}
+{'loss': '1.293', 'grad_norm': '11.06', 'learning_rate': '9.832e-06', 'num_tokens': '6.235e+04', 'mean_token_accuracy': '0.6507', 'epoch': '0.01702'}
+{'loss': '1.629', 'grad_norm': '7.5', 'learning_rate': '9.83e-06', 'num_tokens': '6.403e+04', 'mean_token_accuracy': '0.6702', 'epoch': '0.01719'}
+{'loss': '2.614', 'grad_norm': '18.12', 'learning_rate': '9.828e-06', 'num_tokens': '6.438e+04', 'mean_token_accuracy': '0.52', 'epoch': '0.01737'}
+{'loss': '1.487', 'grad_norm': '13.06', 'learning_rate': '9.826e-06', 'num_tokens': '6.503e+04', 'mean_token_accuracy': '0.6093', 'epoch': '0.01754'}
+{'loss': '1.352', 'grad_norm': '10.44', 'learning_rate': '9.825e-06', 'num_tokens': '6.605e+04', 'mean_token_accuracy': '0.6358', 'epoch': '0.01771'}
+{'loss': '2.564', 'grad_norm': '11.12', 'learning_rate': '9.823e-06', 'num_tokens': '6.742e+04', 'mean_token_accuracy': '0.5212', 'epoch': '0.01789'}
+{'loss': '1.558', 'grad_norm': '11.31', 'learning_rate': '9.821e-06', 'num_tokens': '6.824e+04', 'mean_token_accuracy': '0.623', 'epoch': '0.01806'}
+{'loss': '2.208', 'grad_norm': '14.88', 'learning_rate': '9.819e-06', 'num_tokens': '6.881e+04', 'mean_token_accuracy': '0.5573', 'epoch': '0.01824'}
+{'loss': '1.446', 'grad_norm': '12.88', 'learning_rate': '9.818e-06', 'num_tokens': '6.946e+04', 'mean_token_accuracy': '0.5981', 'epoch': '0.01841'}
+{'loss': '1.596', 'grad_norm': '13.5', 'learning_rate': '9.816e-06', 'num_tokens': '7.009e+04', 'mean_token_accuracy': '0.5934', 'epoch': '0.01858'}
+{'loss': '1.578', 'grad_norm': '12.38', 'learning_rate': '9.814e-06', 'num_tokens': '7.085e+04', 'mean_token_accuracy': '0.5947', 'epoch': '0.01876'}
+{'loss': '1.291', 'grad_norm': '27.5', 'learning_rate': '9.812e-06', 'num_tokens': '7.106e+04', 'mean_token_accuracy': '0.7062', 'epoch': '0.01893'}
+
  2%|▏         | 110/5758 [00:37<31:01,  3.03it/s]
  2%|▏         | 111/5758 [00:37<29:52,  3.15it/s]
                                                  

  2%|▏         | 111/5758 [00:37<29:52,  3.15it/s]
  2%|▏         | 112/5758 [00:38<30:31,  3.08it/s]
                                                  

  2%|▏         | 112/5758 [00:38<30:31,  3.08it/s]
  2%|▏         | 113/5758 [00:38<30:28,  3.09it/s]
                                                  

  2%|▏         | 113/5758 [00:38<30:28,  3.09it/s]
  2%|▏         | 114/5758 [00:38<31:02,  3.03it/s]
                                                  

  2%|▏         | 114/5758 [00:38<31:02,  3.03it/s]
  2%|▏         | 115/5758 [00:39<31:16,  3.01it/s]
                                                  

  2%|▏         | 115/5758 [00:39<31:16,  3.01it/s]
  2%|▏         | 116/5758 [00:39<30:40,  3.07it/s]
                                                  

  2%|▏         | 116/5758 [00:39<30:40,  3.07it/s]
  2%|▏         | 117/5758 [00:39<29:43,  3.16it/s]
                                                  

  2%|▏         | 117/5758 [00:39<29:43,  3.16it/s]
  2%|▏         | 118/5758 [00:40<32:18,  2.91it/s]
                                                  

  2%|▏         | 118/5758 [00:40<32:18,  2.91it/s]
  2%|▏         | 119/5758 [00:40<31:17,  3.00it/s]
                                                  

  2%|▏         | 119/5758 [00:40<31:17,  3.00it/s]
  2%|▏         | 120/5758 [00:40<31:36,  2.97it/s]
                                                  

  2%|▏         | 120/5758 [00:40<31:36,  2.97it/s]
  2%|▏         | 121/5758 [00:41<30:43,  3.06it/s]
                                                  

  2%|▏         | 121/5758 [00:41<30:43,  3.06it/s]
  2%|▏         | 122/5758 [00:41<29:59,  3.13it/s]
                                                  

  2%|▏         | 122/5758 [00:41<29:59,  3.13it/s]
  2%|▏         | 123/5758 [00:41<29:31,  3.18it/s]
                                                  

  2%|▏         | 123/5758 [00:41<29:31,  3.18it/s]
  2%|▏         | 124/5758 [00:42<31:20,  3.00it/s]
                                                  

  2%|▏         | 124/5758 [00:42<31:20,  3.00it/s]
  2%|▏         | 125/5758 [00:42<30:57,  3.03it/s]
                                                  

  2%|▏         | 125/5758 [00:42<30:57,  3.03it/s]
  2%|▏         | 126/5758 [00:42<30:28,  3.08it/s]
                                                  

  2%|▏         | 126/5758 [00:42<30:28,  3.08it/s]
  2%|▏         | 127/5758 [00:43<30:38,  3.06it/s]
                                                  

  2%|▏         | 127/5758 [00:43<30:38,  3.06it/s]
  2%|▏         | 128/5758 [00:43<30:22,  3.09it/s]
                                                  

  2%|▏         | 128/5758 [00:43<30:22,  3.09it/s]
  2%|▏         | 129/5758 [00:43<29:49,  3.15it/s]
                                                  

  2%|▏         | 129/5758 [00:43<29:49,  3.15it/s]
  2%|▏         | 130/5758 [00:44<32:23,  2.90it/s]
                                                  

  2%|▏         | 130/5758 [00:44<32:23,  2.90it/s]
  2%|▏         | 131/5758 [00:44<32:01,  2.93it/s]
                                                  

  2%|▏         | 131/5758 [00:44<32:01,  2.93it/s]
  2%|▏         | 132/5758 [00:44<31:54,  2.94it/s]
                                                  

  2%|▏         | 132/5758 [00:44<31:54,  2.94it/s]
  2%|▏         | 133/5758 [00:45<31:29,  2.98it/s]
                                                  

  2%|▏         | 133/5758 [00:45<31:29,  2.98it/s]
  2%|▏         | 134/5758 [00:45<31:35,  2.97it/s]
                                                  

  2%|▏         | 134/5758 [00:45<31:35,  2.97it/s]
  2%|▏         | 135/5758 [00:45<32:40,  2.87it/s]
                                                  

  2%|▏         | 135/5758 [00:45<32:40,  2.87it/s]
  2%|▏         | 136/5758 [00:46<33:57,  2.76it/s]
                                                  

  2%|▏         | 136/5758 [00:46<33:57,  2.76it/s]
  2%|▏         | 137/5758 [00:46<33:48,  2.77it/s]
                                                  

  2%|▏         | 137/5758 [00:46<33:48,  2.77it/s]
  2%|▏         | 138/5758 [00:46<32:02,  2.92it/s]
                                                  

  2%|▏         | 138/5758 [00:46<32:02,  2.92it/s]
  2%|▏         | 139/5758 [00:47<31:47,  2.95it/s]
                                                  

  2%|▏         | 139/5758 [00:47<31:47,  2.95it/s]
  2%|▏         | 140/5758 [00:47<30:48,  3.04it/s]
                                                  

  2%|▏         | 140/5758 [00:47<30:48,  3.04it/s]
  2%|▏         | 141/5758 [00:47<30:08,  3.11it/s]
                                                  

  2%|▏         | 141/5758 [00:47<30:08,  3.11it/s]
  2%|▏         | 142/5758 [00:48<31:31,  2.97it/s]
                                                  

  2%|▏         | 142/5758 [00:48<31:31,  2.97it/s]
  2%|▏         | 143/5758 [00:48<31:08,  3.01it/s]
                                                  

  2%|▏         | 143/5758 [00:48<31:08,  3.01it/s]
  3%|▎         | 144/5758 [00:48<31:28,  2.97it/s]
                                                  

  3%|▎         | 144/5758 [00:48<31:28,  2.97it/s]
  3%|▎         | 145/5758 [00:49<31:35,  2.96it/s]
                                                  

  3%|▎         | 145/5758 [00:49<31:35,  2.96it/s]
  3%|▎         | 146/5758 [00:49<32:33,  2.87it/s]
                                                  

  3%|▎         | 146/5758 [00:49<32:33,  2.87it/s]
  3%|▎         | 147/5758 [00:49<32:09,  2.91it/s]
                                                  

  3%|▎         | 147/5758 [00:49<32:09,  2.91it/s]
  3%|▎         | 148/5758 [00:50<30:41,  3.05it/s]
                                                  

  3%|▎         | 148/5758 [00:50<30:41,  3.05it/s]
  3%|▎         | 149/5758 [00:50<30:04,  3.11it/s]
                                                  

  3%|▎         | 149/5758 [00:50<30:04,  3.11it/s]
  3%|▎         | 150/5758 [00:50<30:06,  3.10it/s]
                                                  

  3%|▎         | 150/5758 [00:50<30:06,  3.10it/s]
  3%|▎         | 151/5758 [00:51<30:42,  3.04it/s]
                                                  

  3%|▎         | 151/5758 [00:51<30:42,  3.04it/s]
  3%|▎         | 152/5758 [00:51<29:51,  3.13it/s]
                                                  

  3%|▎         | 152/5758 [00:51<29:51,  3.13it/s]
  3%|▎         | 153/5758 [00:51<29:59,  3.11it/s]
                                                  

  3%|▎         | 153/5758 [00:51<29:59,  3.11it/s]
  3%|▎         | 154/5758 [00:52<31:54,  2.93it/s]
                                                  

  3%|▎         | 154/5758 [00:52<31:54,  2.93it/s]
  3%|▎         | 155/5758 [00:52<35:16,  2.65it/s]
                                                  

  3%|▎         | 155/5758 [00:52<35:16,  2.65it/s]
  3%|▎         | 156/5758 [00:53<33:18,  2.80it/s]
                                                  

  3%|▎         | 156/5758 [00:53<33:18,  2.80it/s]
  3%|▎         | 157/5758 [00:53<33:04,  2.82it/s]
                                                  

  3%|▎         | 157/5758 [00:53<33:04,  2.82it/s]
  3%|▎         | 158/5758 [00:53<31:29,  2.96it/s]
                                                  

  3%|▎         | 158/5758 [00:53<31:29,  2.96it/s]
  3%|▎         | 159/5758 [00:53<30:12,  3.09it/s]
                                                  

  3%|▎         | 159/5758 [00:53<30:12,  3.09it/s]
  3%|▎         | 160/5758 [00:54<31:28,  2.96it/s]
                                                  

  3%|▎         | 160/5758 [00:54<31:28,  2.96it/s]
  3%|▎         | 161/5758 [00:54<30:27,  3.06it/s]
                                                  

  3%|▎         | 161/5758 [00:54<30:27,  3.06it/s]
  3%|▎         | 162/5758 [00:55<34:10,  2.73it/s]
                                                  

  3%|▎         | 162/5758 [00:55<34:10,  2.73it/s]
  3%|▎         | 163/5758 [00:55<35:10,  2.65it/s]
                                                  

  3%|▎         | 163/5758 [00:55<35:10,  2.65it/s]
  3%|▎         | 164/5758 [00:55<33:26,  2.79it/s]
                                                  
{'loss': '1.701', 'grad_norm': '10.75', 'learning_rate': '9.811e-06', 'num_tokens': '7.205e+04', 'mean_token_accuracy': '0.5628', 'epoch': '0.0191'}
+{'loss': '1.423', 'grad_norm': '13', 'learning_rate': '9.809e-06', 'num_tokens': '7.267e+04', 'mean_token_accuracy': '0.627', 'epoch': '0.01928'}
+{'loss': '0.8301', 'grad_norm': '7.688', 'learning_rate': '9.807e-06', 'num_tokens': '7.451e+04', 'mean_token_accuracy': '0.7606', 'epoch': '0.01945'}
+{'loss': '1.128', 'grad_norm': '11.5', 'learning_rate': '9.805e-06', 'num_tokens': '7.534e+04', 'mean_token_accuracy': '0.6872', 'epoch': '0.01962'}
+{'loss': '1.381', 'grad_norm': '18.62', 'learning_rate': '9.804e-06', 'num_tokens': '7.557e+04', 'mean_token_accuracy': '0.6826', 'epoch': '0.0198'}
+{'loss': '0.8219', 'grad_norm': '7.625', 'learning_rate': '9.802e-06', 'num_tokens': '7.734e+04', 'mean_token_accuracy': '0.7688', 'epoch': '0.01997'}
+{'loss': '1.372', 'grad_norm': '11.12', 'learning_rate': '9.8e-06', 'num_tokens': '7.81e+04', 'mean_token_accuracy': '0.6301', 'epoch': '0.02015'}
+{'loss': '2.586', 'grad_norm': '17.88', 'learning_rate': '9.799e-06', 'num_tokens': '7.874e+04', 'mean_token_accuracy': '0.5158', 'epoch': '0.02032'}
+{'loss': '2.712', 'grad_norm': '27.88', 'learning_rate': '9.797e-06', 'num_tokens': '7.888e+04', 'mean_token_accuracy': '0.5646', 'epoch': '0.02049'}
+{'loss': '1.526', 'grad_norm': '11.62', 'learning_rate': '9.795e-06', 'num_tokens': '7.963e+04', 'mean_token_accuracy': '0.6051', 'epoch': '0.02067'}
+{'loss': '2.164', 'grad_norm': '19.88', 'learning_rate': '9.793e-06', 'num_tokens': '8.01e+04', 'mean_token_accuracy': '0.586', 'epoch': '0.02084'}
+{'loss': '1.554', 'grad_norm': '13', 'learning_rate': '9.792e-06', 'num_tokens': '8.081e+04', 'mean_token_accuracy': '0.5847', 'epoch': '0.02101'}
+{'loss': '1.088', 'grad_norm': '11.31', 'learning_rate': '9.79e-06', 'num_tokens': '8.149e+04', 'mean_token_accuracy': '0.6912', 'epoch': '0.02119'}
+{'loss': '1.531', 'grad_norm': '11.5', 'learning_rate': '9.788e-06', 'num_tokens': '8.219e+04', 'mean_token_accuracy': '0.6165', 'epoch': '0.02136'}
+{'loss': '1.552', 'grad_norm': '14.5', 'learning_rate': '9.786e-06', 'num_tokens': '8.276e+04', 'mean_token_accuracy': '0.6124', 'epoch': '0.02154'}
+{'loss': '1.313', 'grad_norm': '10.38', 'learning_rate': '9.785e-06', 'num_tokens': '8.354e+04', 'mean_token_accuracy': '0.6923', 'epoch': '0.02171'}
+{'loss': '3.041', 'grad_norm': '13.69', 'learning_rate': '9.783e-06', 'num_tokens': '8.429e+04', 'mean_token_accuracy': '0.4265', 'epoch': '0.02188'}
+{'loss': '1.575', 'grad_norm': '11.75', 'learning_rate': '9.781e-06', 'num_tokens': '8.514e+04', 'mean_token_accuracy': '0.5778', 'epoch': '0.02206'}
+{'loss': '1.637', 'grad_norm': '14.62', 'learning_rate': '9.779e-06', 'num_tokens': '8.552e+04', 'mean_token_accuracy': '0.6464', 'epoch': '0.02223'}
+{'loss': '1.665', 'grad_norm': '13.19', 'learning_rate': '9.778e-06', 'num_tokens': '8.621e+04', 'mean_token_accuracy': '0.5669', 'epoch': '0.0224'}
+{'loss': '2.261', 'grad_norm': '43', 'learning_rate': '9.776e-06', 'num_tokens': '8.628e+04', 'mean_token_accuracy': '0.6471', 'epoch': '0.02258'}
+{'loss': '1.551', 'grad_norm': '14.12', 'learning_rate': '9.774e-06', 'num_tokens': '8.67e+04', 'mean_token_accuracy': '0.6556', 'epoch': '0.02275'}
+{'loss': '0.8183', 'grad_norm': '8.125', 'learning_rate': '9.772e-06', 'num_tokens': '8.846e+04', 'mean_token_accuracy': '0.7544', 'epoch': '0.02292'}
+{'loss': '1.646', 'grad_norm': '12.25', 'learning_rate': '9.771e-06', 'num_tokens': '8.928e+04', 'mean_token_accuracy': '0.5832', 'epoch': '0.0231'}
+{'loss': '1.306', 'grad_norm': '14.56', 'learning_rate': '9.769e-06', 'num_tokens': '8.973e+04', 'mean_token_accuracy': '0.651', 'epoch': '0.02327'}
+{'loss': '1.529', 'grad_norm': '15.25', 'learning_rate': '9.767e-06', 'num_tokens': '9.001e+04', 'mean_token_accuracy': '0.6996', 'epoch': '0.02345'}
+{'loss': '2.113', 'grad_norm': '42', 'learning_rate': '9.766e-06', 'num_tokens': '9.015e+04', 'mean_token_accuracy': '0.6241', 'epoch': '0.02362'}
+{'loss': '1.42', 'grad_norm': '14.88', 'learning_rate': '9.764e-06', 'num_tokens': '9.066e+04', 'mean_token_accuracy': '0.6088', 'epoch': '0.02379'}
+{'loss': '1.451', 'grad_norm': '13.19', 'learning_rate': '9.762e-06', 'num_tokens': '9.133e+04', 'mean_token_accuracy': '0.622', 'epoch': '0.02397'}
+{'loss': '2.29', 'grad_norm': '14.75', 'learning_rate': '9.76e-06', 'num_tokens': '9.219e+04', 'mean_token_accuracy': '0.5405', 'epoch': '0.02414'}
+{'loss': '1.502', 'grad_norm': '13.25', 'learning_rate': '9.759e-06', 'num_tokens': '9.288e+04', 'mean_token_accuracy': '0.5986', 'epoch': '0.02431'}
+{'loss': '1.361', 'grad_norm': '12.5', 'learning_rate': '9.757e-06', 'num_tokens': '9.359e+04', 'mean_token_accuracy': '0.6362', 'epoch': '0.02449'}
+{'loss': '2.262', 'grad_norm': '17.25', 'learning_rate': '9.755e-06', 'num_tokens': '9.386e+04', 'mean_token_accuracy': '0.5956', 'epoch': '0.02466'}
+{'loss': '1.48', 'grad_norm': '10.75', 'learning_rate': '9.753e-06', 'num_tokens': '9.467e+04', 'mean_token_accuracy': '0.5854', 'epoch': '0.02484'}
+{'loss': '1.953', 'grad_norm': '14.62', 'learning_rate': '9.752e-06', 'num_tokens': '9.513e+04', 'mean_token_accuracy': '0.6266', 'epoch': '0.02501'}
+{'loss': '1.282', 'grad_norm': '12.69', 'learning_rate': '9.75e-06', 'num_tokens': '9.558e+04', 'mean_token_accuracy': '0.7058', 'epoch': '0.02518'}
+{'loss': '1.831', 'grad_norm': '9.938', 'learning_rate': '9.748e-06', 'num_tokens': '9.664e+04', 'mean_token_accuracy': '0.6107', 'epoch': '0.02536'}
+{'loss': '2.557', 'grad_norm': '22.25', 'learning_rate': '9.746e-06', 'num_tokens': '9.708e+04', 'mean_token_accuracy': '0.4907', 'epoch': '0.02553'}
+{'loss': '1.399', 'grad_norm': '11.81', 'learning_rate': '9.745e-06', 'num_tokens': '9.769e+04', 'mean_token_accuracy': '0.615', 'epoch': '0.0257'}
+{'loss': '2.277', 'grad_norm': '20.5', 'learning_rate': '9.743e-06', 'num_tokens': '9.787e+04', 'mean_token_accuracy': '0.6158', 'epoch': '0.02588'}
+{'loss': '1.287', 'grad_norm': '10.12', 'learning_rate': '9.741e-06', 'num_tokens': '9.867e+04', 'mean_token_accuracy': '0.6617', 'epoch': '0.02605'}
+{'loss': '1.403', 'grad_norm': '10.75', 'learning_rate': '9.739e-06', 'num_tokens': '9.963e+04', 'mean_token_accuracy': '0.673', 'epoch': '0.02622'}
+{'loss': '1.46', 'grad_norm': '10.56', 'learning_rate': '9.738e-06', 'num_tokens': '1.003e+05', 'mean_token_accuracy': '0.6686', 'epoch': '0.0264'}
+{'loss': '2.324', 'grad_norm': '17.38', 'learning_rate': '9.736e-06', 'num_tokens': '1.007e+05', 'mean_token_accuracy': '0.565', 'epoch': '0.02657'}
+{'loss': '1.373', 'grad_norm': '12.06', 'learning_rate': '9.734e-06', 'num_tokens': '1.013e+05', 'mean_token_accuracy': '0.6985', 'epoch': '0.02675'}
+{'loss': '1.778', 'grad_norm': '17', 'learning_rate': '9.733e-06', 'num_tokens': '1.018e+05', 'mean_token_accuracy': '0.5723', 'epoch': '0.02692'}
+{'loss': '1.604', 'grad_norm': '12.62', 'learning_rate': '9.731e-06', 'num_tokens': '1.025e+05', 'mean_token_accuracy': '0.5793', 'epoch': '0.02709'}
+{'loss': '2.334', 'grad_norm': '25.62', 'learning_rate': '9.729e-06', 'num_tokens': '1.027e+05', 'mean_token_accuracy': '0.4458', 'epoch': '0.02727'}
+{'loss': '1.416', 'grad_norm': '12.19', 'learning_rate': '9.727e-06', 'num_tokens': '1.034e+05', 'mean_token_accuracy': '0.6375', 'epoch': '0.02744'}
+{'loss': '1.746', 'grad_norm': '10.69', 'learning_rate': '9.726e-06', 'num_tokens': '1.046e+05', 'mean_token_accuracy': '0.623', 'epoch': '0.02761'}
+{'loss': '2.008', 'grad_norm': '18.38', 'learning_rate': '9.724e-06', 'num_tokens': '1.049e+05', 'mean_token_accuracy': '0.597', 'epoch': '0.02779'}
+{'loss': '1.305', 'grad_norm': '13.12', 'learning_rate': '9.722e-06', 'num_tokens': '1.055e+05', 'mean_token_accuracy': '0.6461', 'epoch': '0.02796'}
+{'loss': '1.25', 'grad_norm': '5.469', 'learning_rate': '9.72e-06', 'num_tokens': '1.096e+05', 'mean_token_accuracy': '0.6955', 'epoch': '0.02813'}
+{'loss': '2.295', 'grad_norm': '26.38', 'learning_rate': '9.719e-06', 'num_tokens': '1.098e+05', 'mean_token_accuracy': '0.5507', 'epoch': '0.02831'}
+
  3%|▎         | 164/5758 [00:55<33:26,  2.79it/s]
  3%|▎         | 165/5758 [00:56<31:37,  2.95it/s]
                                                  

  3%|▎         | 165/5758 [00:56<31:37,  2.95it/s]
  3%|▎         | 166/5758 [00:56<30:56,  3.01it/s]
                                                  

  3%|▎         | 166/5758 [00:56<30:56,  3.01it/s]
  3%|▎         | 167/5758 [00:56<30:08,  3.09it/s]
                                                  

  3%|▎         | 167/5758 [00:56<30:08,  3.09it/s]
  3%|▎         | 168/5758 [00:57<30:12,  3.08it/s]
                                                  

  3%|▎         | 168/5758 [00:57<30:12,  3.08it/s]
  3%|▎         | 169/5758 [00:57<30:40,  3.04it/s]
                                                  

  3%|▎         | 169/5758 [00:57<30:40,  3.04it/s]
  3%|▎         | 170/5758 [00:57<32:18,  2.88it/s]
                                                  

  3%|▎         | 170/5758 [00:57<32:18,  2.88it/s]
  3%|▎         | 171/5758 [00:58<32:13,  2.89it/s]
                                                  

  3%|▎         | 171/5758 [00:58<32:13,  2.89it/s]
  3%|▎         | 172/5758 [00:58<30:57,  3.01it/s]
                                                  

  3%|▎         | 172/5758 [00:58<30:57,  3.01it/s]
  3%|▎         | 173/5758 [00:58<32:20,  2.88it/s]
                                                  

  3%|▎         | 173/5758 [00:58<32:20,  2.88it/s]
  3%|▎         | 174/5758 [00:59<30:51,  3.02it/s]
                                                  

  3%|▎         | 174/5758 [00:59<30:51,  3.02it/s]
  3%|▎         | 175/5758 [00:59<31:45,  2.93it/s]
                                                  

  3%|▎         | 175/5758 [00:59<31:45,  2.93it/s]
  3%|▎         | 176/5758 [00:59<32:40,  2.85it/s]
                                                  

  3%|▎         | 176/5758 [00:59<32:40,  2.85it/s]
  3%|▎         | 177/5758 [01:00<30:54,  3.01it/s]
                                                  

  3%|▎         | 177/5758 [01:00<30:54,  3.01it/s]
  3%|▎         | 178/5758 [01:00<34:06,  2.73it/s]
                                                  

  3%|▎         | 178/5758 [01:00<34:06,  2.73it/s]
  3%|▎         | 179/5758 [01:00<32:33,  2.86it/s]
                                                  

  3%|▎         | 179/5758 [01:00<32:33,  2.86it/s]
  3%|▎         | 180/5758 [01:01<32:09,  2.89it/s]
                                                  

  3%|▎         | 180/5758 [01:01<32:09,  2.89it/s]
  3%|▎         | 181/5758 [01:01<32:12,  2.89it/s]
                                                  

  3%|▎         | 181/5758 [01:01<32:12,  2.89it/s]
  3%|▎         | 182/5758 [01:01<30:55,  3.01it/s]
                                                  

  3%|▎         | 182/5758 [01:01<30:55,  3.01it/s]
  3%|▎         | 183/5758 [01:02<30:05,  3.09it/s]
                                                  

  3%|▎         | 183/5758 [01:02<30:05,  3.09it/s]
  3%|▎         | 184/5758 [01:02<30:13,  3.07it/s]
                                                  

  3%|▎         | 184/5758 [01:02<30:13,  3.07it/s]
  3%|▎         | 185/5758 [01:02<31:26,  2.95it/s]
                                                  

  3%|▎         | 185/5758 [01:02<31:26,  2.95it/s]
  3%|▎         | 186/5758 [01:03<31:51,  2.91it/s]
                                                  

  3%|▎         | 186/5758 [01:03<31:51,  2.91it/s]
  3%|▎         | 187/5758 [01:03<32:42,  2.84it/s]
                                                  

  3%|▎         | 187/5758 [01:03<32:42,  2.84it/s]
  3%|▎         | 188/5758 [01:03<33:51,  2.74it/s]
                                                  

  3%|▎         | 188/5758 [01:03<33:51,  2.74it/s]
  3%|▎         | 189/5758 [01:04<31:51,  2.91it/s]
                                                  

  3%|▎         | 189/5758 [01:04<31:51,  2.91it/s]
  3%|▎         | 190/5758 [01:04<31:49,  2.92it/s]
                                                  

  3%|▎         | 190/5758 [01:04<31:49,  2.92it/s]
  3%|▎         | 191/5758 [01:04<31:24,  2.95it/s]
                                                  

  3%|▎         | 191/5758 [01:04<31:24,  2.95it/s]
  3%|▎         | 192/5758 [01:05<30:25,  3.05it/s]
                                                  

  3%|▎         | 192/5758 [01:05<30:25,  3.05it/s]
  3%|▎         | 193/5758 [01:05<29:30,  3.14it/s]
                                                  

  3%|▎         | 193/5758 [01:05<29:30,  3.14it/s]
  3%|▎         | 194/5758 [01:05<29:03,  3.19it/s]
                                                  

  3%|▎         | 194/5758 [01:05<29:03,  3.19it/s]
  3%|▎         | 195/5758 [01:06<28:31,  3.25it/s]
                                                  

  3%|▎         | 195/5758 [01:06<28:31,  3.25it/s]
  3%|▎         | 196/5758 [01:06<28:08,  3.29it/s]
                                                  

  3%|▎         | 196/5758 [01:06<28:08,  3.29it/s]
  3%|▎         | 197/5758 [01:06<29:25,  3.15it/s]
                                                  

  3%|▎         | 197/5758 [01:06<29:25,  3.15it/s]
  3%|▎         | 198/5758 [01:07<31:00,  2.99it/s]
                                                  

  3%|▎         | 198/5758 [01:07<31:00,  2.99it/s]
  3%|▎         | 199/5758 [01:07<30:19,  3.05it/s]
                                                  

  3%|▎         | 199/5758 [01:07<30:19,  3.05it/s]
  3%|▎         | 200/5758 [01:07<29:51,  3.10it/s]
                                                  

  3%|▎         | 200/5758 [01:07<29:51,  3.10it/s]
  3%|▎         | 201/5758 [01:08<29:55,  3.10it/s]
                                                  

  3%|▎         | 201/5758 [01:08<29:55,  3.10it/s]
  4%|▎         | 202/5758 [01:08<29:33,  3.13it/s]
                                                  

  4%|▎         | 202/5758 [01:08<29:33,  3.13it/s]
  4%|▎         | 203/5758 [01:08<30:48,  3.00it/s]
                                                  

  4%|▎         | 203/5758 [01:08<30:48,  3.00it/s]
  4%|▎         | 204/5758 [01:09<31:15,  2.96it/s]
                                                  

  4%|▎         | 204/5758 [01:09<31:15,  2.96it/s]
  4%|▎         | 205/5758 [01:09<34:47,  2.66it/s]
                                                  

  4%|▎         | 205/5758 [01:09<34:47,  2.66it/s]
  4%|▎         | 206/5758 [01:09<33:49,  2.74it/s]
                                                  

  4%|▎         | 206/5758 [01:09<33:49,  2.74it/s]
  4%|▎         | 207/5758 [01:10<31:38,  2.92it/s]
                                                  

  4%|▎         | 207/5758 [01:10<31:38,  2.92it/s]
  4%|▎         | 208/5758 [01:10<32:37,  2.84it/s]
                                                  

  4%|▎         | 208/5758 [01:10<32:37,  2.84it/s]
  4%|▎         | 209/5758 [01:10<31:17,  2.96it/s]
                                                  

  4%|▎         | 209/5758 [01:10<31:17,  2.96it/s]
  4%|▎         | 210/5758 [01:11<32:10,  2.87it/s]
                                                  

  4%|▎         | 210/5758 [01:11<32:10,  2.87it/s]
  4%|▎         | 211/5758 [01:11<30:50,  3.00it/s]
                                                  

  4%|▎         | 211/5758 [01:11<30:50,  3.00it/s]
  4%|▎         | 212/5758 [01:11<30:12,  3.06it/s]
                                                  

  4%|▎         | 212/5758 [01:11<30:12,  3.06it/s]
  4%|▎         | 213/5758 [01:12<32:02,  2.88it/s]
                                                  

  4%|▎         | 213/5758 [01:12<32:02,  2.88it/s]
  4%|▎         | 214/5758 [01:12<30:20,  3.05it/s]
                                                  

  4%|▎         | 214/5758 [01:12<30:20,  3.05it/s]
  4%|▎         | 215/5758 [01:12<29:34,  3.12it/s]
                                                  

  4%|▎         | 215/5758 [01:12<29:34,  3.12it/s]
  4%|▍         | 216/5758 [01:13<31:26,  2.94it/s]
                                                  

  4%|▍         | 216/5758 [01:13<31:26,  2.94it/s]
  4%|▍         | 217/5758 [01:13<32:54,  2.81it/s]
                                                  

  4%|▍         | 217/5758 [01:13<32:54,  2.81it/s]
  4%|▍         | 218/5758 [01:13<31:32,  2.93it/s]
                                                  
{'loss': '1.317', 'grad_norm': '12.38', 'learning_rate': '9.717e-06', 'num_tokens': '1.105e+05', 'mean_token_accuracy': '0.6485', 'epoch': '0.02848'}
+{'loss': '1.421', 'grad_norm': '15.94', 'learning_rate': '9.715e-06', 'num_tokens': '1.108e+05', 'mean_token_accuracy': '0.7092', 'epoch': '0.02866'}
+{'loss': '2.155', 'grad_norm': '11.06', 'learning_rate': '9.713e-06', 'num_tokens': '1.123e+05', 'mean_token_accuracy': '0.5559', 'epoch': '0.02883'}
+{'loss': '2.426', 'grad_norm': '17', 'learning_rate': '9.712e-06', 'num_tokens': '1.13e+05', 'mean_token_accuracy': '0.5308', 'epoch': '0.029'}
+{'loss': '1.42', 'grad_norm': '15.19', 'learning_rate': '9.71e-06', 'num_tokens': '1.134e+05', 'mean_token_accuracy': '0.6767', 'epoch': '0.02918'}
+{'loss': '2.279', 'grad_norm': '16.25', 'learning_rate': '9.708e-06', 'num_tokens': '1.139e+05', 'mean_token_accuracy': '0.5646', 'epoch': '0.02935'}
+{'loss': '1.224', 'grad_norm': '13.44', 'learning_rate': '9.706e-06', 'num_tokens': '1.144e+05', 'mean_token_accuracy': '0.6538', 'epoch': '0.02952'}
+{'loss': '2.569', 'grad_norm': '12.69', 'learning_rate': '9.705e-06', 'num_tokens': '1.154e+05', 'mean_token_accuracy': '0.5338', 'epoch': '0.0297'}
+{'loss': '1.269', 'grad_norm': '11.56', 'learning_rate': '9.703e-06', 'num_tokens': '1.16e+05', 'mean_token_accuracy': '0.6429', 'epoch': '0.02987'}
+{'loss': '1.973', 'grad_norm': '19.38', 'learning_rate': '9.701e-06', 'num_tokens': '1.166e+05', 'mean_token_accuracy': '0.5888', 'epoch': '0.03005'}
+{'loss': '1.472', 'grad_norm': '13.5', 'learning_rate': '9.7e-06', 'num_tokens': '1.172e+05', 'mean_token_accuracy': '0.609', 'epoch': '0.03022'}
+{'loss': '1.569', 'grad_norm': '13.69', 'learning_rate': '9.698e-06', 'num_tokens': '1.178e+05', 'mean_token_accuracy': '0.618', 'epoch': '0.03039'}
+{'loss': '1.364', 'grad_norm': '9', 'learning_rate': '9.696e-06', 'num_tokens': '1.189e+05', 'mean_token_accuracy': '0.6613', 'epoch': '0.03057'}
+{'loss': '1.296', 'grad_norm': '12.81', 'learning_rate': '9.694e-06', 'num_tokens': '1.195e+05', 'mean_token_accuracy': '0.6322', 'epoch': '0.03074'}
+{'loss': '2.297', 'grad_norm': '32.5', 'learning_rate': '9.693e-06', 'num_tokens': '1.196e+05', 'mean_token_accuracy': '0.567', 'epoch': '0.03091'}
+{'loss': '1.528', 'grad_norm': '26.25', 'learning_rate': '9.691e-06', 'num_tokens': '1.198e+05', 'mean_token_accuracy': '0.6973', 'epoch': '0.03109'}
+{'loss': '1.838', 'grad_norm': '15.31', 'learning_rate': '9.689e-06', 'num_tokens': '1.202e+05', 'mean_token_accuracy': '0.625', 'epoch': '0.03126'}
+{'loss': '1.279', 'grad_norm': '12', 'learning_rate': '9.687e-06', 'num_tokens': '1.212e+05', 'mean_token_accuracy': '0.6427', 'epoch': '0.03143'}
+{'loss': '2.335', 'grad_norm': '14.19', 'learning_rate': '9.686e-06', 'num_tokens': '1.218e+05', 'mean_token_accuracy': '0.5314', 'epoch': '0.03161'}
+{'loss': '1.835', 'grad_norm': '24.12', 'learning_rate': '9.684e-06', 'num_tokens': '1.22e+05', 'mean_token_accuracy': '0.6337', 'epoch': '0.03178'}
+{'loss': '2.42', 'grad_norm': '20.5', 'learning_rate': '9.682e-06', 'num_tokens': '1.222e+05', 'mean_token_accuracy': '0.5385', 'epoch': '0.03196'}
+{'loss': '2.462', 'grad_norm': '12.75', 'learning_rate': '9.68e-06', 'num_tokens': '1.228e+05', 'mean_token_accuracy': '0.5325', 'epoch': '0.03213'}
+{'loss': '2.686', 'grad_norm': '42.75', 'learning_rate': '9.679e-06', 'num_tokens': '1.229e+05', 'mean_token_accuracy': '0.5217', 'epoch': '0.0323'}
+{'loss': '1.569', 'grad_norm': '9.562', 'learning_rate': '9.677e-06', 'num_tokens': '1.24e+05', 'mean_token_accuracy': '0.6357', 'epoch': '0.03248'}
+{'loss': '1.968', 'grad_norm': '21.5', 'learning_rate': '9.675e-06', 'num_tokens': '1.241e+05', 'mean_token_accuracy': '0.6126', 'epoch': '0.03265'}
+{'loss': '1.603', 'grad_norm': '8.562', 'learning_rate': '9.673e-06', 'num_tokens': '1.254e+05', 'mean_token_accuracy': '0.6534', 'epoch': '0.03282'}
+{'loss': '2.11', 'grad_norm': '23.12', 'learning_rate': '9.672e-06', 'num_tokens': '1.256e+05', 'mean_token_accuracy': '0.625', 'epoch': '0.033'}
+{'loss': '2.027', 'grad_norm': '20.5', 'learning_rate': '9.67e-06', 'num_tokens': '1.26e+05', 'mean_token_accuracy': '0.577', 'epoch': '0.03317'}
+{'loss': '1.306', 'grad_norm': '11.06', 'learning_rate': '9.668e-06', 'num_tokens': '1.267e+05', 'mean_token_accuracy': '0.6327', 'epoch': '0.03334'}
+{'loss': '2.248', 'grad_norm': '13.12', 'learning_rate': '9.667e-06', 'num_tokens': '1.273e+05', 'mean_token_accuracy': '0.5404', 'epoch': '0.03352'}
+{'loss': '1.524', 'grad_norm': '10.81', 'learning_rate': '9.665e-06', 'num_tokens': '1.28e+05', 'mean_token_accuracy': '0.6559', 'epoch': '0.03369'}
+{'loss': '2.408', 'grad_norm': '11.62', 'learning_rate': '9.663e-06', 'num_tokens': '1.293e+05', 'mean_token_accuracy': '0.5491', 'epoch': '0.03387'}
+{'loss': '1.598', 'grad_norm': '14.44', 'learning_rate': '9.661e-06', 'num_tokens': '1.299e+05', 'mean_token_accuracy': '0.5893', 'epoch': '0.03404'}
+{'loss': '2.075', 'grad_norm': '20.5', 'learning_rate': '9.66e-06', 'num_tokens': '1.302e+05', 'mean_token_accuracy': '0.5588', 'epoch': '0.03421'}
+{'loss': '2.056', 'grad_norm': '13.56', 'learning_rate': '9.658e-06', 'num_tokens': '1.307e+05', 'mean_token_accuracy': '0.5775', 'epoch': '0.03439'}
+{'loss': '2.203', 'grad_norm': '15.19', 'learning_rate': '9.656e-06', 'num_tokens': '1.311e+05', 'mean_token_accuracy': '0.5495', 'epoch': '0.03456'}
+{'loss': '2.073', 'grad_norm': '12.31', 'learning_rate': '9.654e-06', 'num_tokens': '1.318e+05', 'mean_token_accuracy': '0.5762', 'epoch': '0.03473'}
+{'loss': '1.302', 'grad_norm': '11', 'learning_rate': '9.653e-06', 'num_tokens': '1.326e+05', 'mean_token_accuracy': '0.6429', 'epoch': '0.03491'}
+{'loss': '1.499', 'grad_norm': '11.88', 'learning_rate': '9.651e-06', 'num_tokens': '1.333e+05', 'mean_token_accuracy': '0.5825', 'epoch': '0.03508'}
+{'loss': '1.442', 'grad_norm': '11.62', 'learning_rate': '9.649e-06', 'num_tokens': '1.339e+05', 'mean_token_accuracy': '0.6755', 'epoch': '0.03526'}
+{'loss': '1.53', 'grad_norm': '38', 'learning_rate': '9.647e-06', 'num_tokens': '1.34e+05', 'mean_token_accuracy': '0.6373', 'epoch': '0.03543'}
+{'loss': '2.663', 'grad_norm': '37.75', 'learning_rate': '9.646e-06', 'num_tokens': '1.341e+05', 'mean_token_accuracy': '0.5047', 'epoch': '0.0356'}
+{'loss': '1.339', 'grad_norm': '12.12', 'learning_rate': '9.644e-06', 'num_tokens': '1.349e+05', 'mean_token_accuracy': '0.6133', 'epoch': '0.03578'}
+{'loss': '1.972', 'grad_norm': '25.12', 'learning_rate': '9.642e-06', 'num_tokens': '1.351e+05', 'mean_token_accuracy': '0.62', 'epoch': '0.03595'}
+{'loss': '2.031', 'grad_norm': '15.56', 'learning_rate': '9.641e-06', 'num_tokens': '1.356e+05', 'mean_token_accuracy': '0.5599', 'epoch': '0.03612'}
+{'loss': '1.396', 'grad_norm': '11.5', 'learning_rate': '9.639e-06', 'num_tokens': '1.363e+05', 'mean_token_accuracy': '0.6913', 'epoch': '0.0363'}
+{'loss': '1.342', 'grad_norm': '13.56', 'learning_rate': '9.637e-06', 'num_tokens': '1.369e+05', 'mean_token_accuracy': '0.6154', 'epoch': '0.03647'}
+{'loss': '2.166', 'grad_norm': '18.88', 'learning_rate': '9.635e-06', 'num_tokens': '1.372e+05', 'mean_token_accuracy': '0.6096', 'epoch': '0.03664'}
+{'loss': '2.321', 'grad_norm': '12.06', 'learning_rate': '9.634e-06', 'num_tokens': '1.379e+05', 'mean_token_accuracy': '0.5634', 'epoch': '0.03682'}
+{'loss': '1.518', 'grad_norm': '27.75', 'learning_rate': '9.632e-06', 'num_tokens': '1.38e+05', 'mean_token_accuracy': '0.6741', 'epoch': '0.03699'}
+{'loss': '1.952', 'grad_norm': '22.25', 'learning_rate': '9.63e-06', 'num_tokens': '1.382e+05', 'mean_token_accuracy': '0.6242', 'epoch': '0.03717'}
+{'loss': '1.777', 'grad_norm': '24.75', 'learning_rate': '9.628e-06', 'num_tokens': '1.384e+05', 'mean_token_accuracy': '0.6012', 'epoch': '0.03734'}
+{'loss': '2.333', 'grad_norm': '19.88', 'learning_rate': '9.627e-06', 'num_tokens': '1.387e+05', 'mean_token_accuracy': '0.5185', 'epoch': '0.03751'}
+{'loss': '1.484', 'grad_norm': '25.25', 'learning_rate': '9.625e-06', 'num_tokens': '1.388e+05', 'mean_token_accuracy': '0.6788', 'epoch': '0.03769'}
+{'loss': '1.278', 'grad_norm': '11.56', 'learning_rate': '9.623e-06', 'num_tokens': '1.395e+05', 'mean_token_accuracy': '0.6449', 'epoch': '0.03786'}
+
  4%|▍         | 218/5758 [01:13<31:32,  2.93it/s]
  4%|▍         | 219/5758 [01:14<31:04,  2.97it/s]
                                                  

  4%|▍         | 219/5758 [01:14<31:04,  2.97it/s]
  4%|▍         | 220/5758 [01:14<29:56,  3.08it/s]
                                                  

  4%|▍         | 220/5758 [01:14<29:56,  3.08it/s]
  4%|▍         | 221/5758 [01:14<31:29,  2.93it/s]
                                                  

  4%|▍         | 221/5758 [01:14<31:29,  2.93it/s]
  4%|▍         | 222/5758 [01:15<31:57,  2.89it/s]
                                                  

  4%|▍         | 222/5758 [01:15<31:57,  2.89it/s]
  4%|▍         | 223/5758 [01:15<30:43,  3.00it/s]
                                                  

  4%|▍         | 223/5758 [01:15<30:43,  3.00it/s]
  4%|▍         | 224/5758 [01:15<29:43,  3.10it/s]
                                                  

  4%|▍         | 224/5758 [01:15<29:43,  3.10it/s]
  4%|▍         | 225/5758 [01:16<30:03,  3.07it/s]
                                                  

  4%|▍         | 225/5758 [01:16<30:03,  3.07it/s]
  4%|▍         | 226/5758 [01:16<29:57,  3.08it/s]
                                                  

  4%|▍         | 226/5758 [01:16<29:57,  3.08it/s]
  4%|▍         | 227/5758 [01:16<30:07,  3.06it/s]
                                                  

  4%|▍         | 227/5758 [01:16<30:07,  3.06it/s]
  4%|▍         | 228/5758 [01:17<29:15,  3.15it/s]
                                                  

  4%|▍         | 228/5758 [01:17<29:15,  3.15it/s]
  4%|▍         | 229/5758 [01:17<28:53,  3.19it/s]
                                                  

  4%|▍         | 229/5758 [01:17<28:53,  3.19it/s]
  4%|▍         | 230/5758 [01:17<28:53,  3.19it/s]
                                                  

  4%|▍         | 230/5758 [01:17<28:53,  3.19it/s]
  4%|▍         | 231/5758 [01:18<32:02,  2.88it/s]
                                                  

  4%|▍         | 231/5758 [01:18<32:02,  2.88it/s]
  4%|▍         | 232/5758 [01:18<31:48,  2.90it/s]
                                                  

  4%|▍         | 232/5758 [01:18<31:48,  2.90it/s]
  4%|▍         | 233/5758 [01:18<30:37,  3.01it/s]
                                                  

  4%|▍         | 233/5758 [01:18<30:37,  3.01it/s]
  4%|▍         | 234/5758 [01:19<29:39,  3.10it/s]
                                                  

  4%|▍         | 234/5758 [01:19<29:39,  3.10it/s]
  4%|▍         | 235/5758 [01:19<29:02,  3.17it/s]
                                                  

  4%|▍         | 235/5758 [01:19<29:02,  3.17it/s]
  4%|▍         | 236/5758 [01:19<32:17,  2.85it/s]
                                                  

  4%|▍         | 236/5758 [01:19<32:17,  2.85it/s]
  4%|▍         | 237/5758 [01:20<33:43,  2.73it/s]
                                                  

  4%|▍         | 237/5758 [01:20<33:43,  2.73it/s]
  4%|▍         | 238/5758 [01:20<32:05,  2.87it/s]
                                                  

  4%|▍         | 238/5758 [01:20<32:05,  2.87it/s]
  4%|▍         | 239/5758 [01:21<32:55,  2.79it/s]
                                                  

  4%|▍         | 239/5758 [01:21<32:55,  2.79it/s]
  4%|▍         | 240/5758 [01:21<31:27,  2.92it/s]
                                                  

  4%|▍         | 240/5758 [01:21<31:27,  2.92it/s]
  4%|▍         | 241/5758 [01:21<31:34,  2.91it/s]
                                                  

  4%|▍         | 241/5758 [01:21<31:34,  2.91it/s]
  4%|▍         | 242/5758 [01:21<30:46,  2.99it/s]
                                                  

  4%|▍         | 242/5758 [01:21<30:46,  2.99it/s]
  4%|▍         | 243/5758 [01:22<29:55,  3.07it/s]
                                                  

  4%|▍         | 243/5758 [01:22<29:55,  3.07it/s]
  4%|▍         | 244/5758 [01:22<29:32,  3.11it/s]
                                                  

  4%|▍         | 244/5758 [01:22<29:32,  3.11it/s]
  4%|▍         | 245/5758 [01:22<28:45,  3.20it/s]
                                                  

  4%|▍         | 245/5758 [01:22<28:45,  3.20it/s]
  4%|▍         | 246/5758 [01:23<28:50,  3.18it/s]
                                                  

  4%|▍         | 246/5758 [01:23<28:50,  3.18it/s]
  4%|▍         | 247/5758 [01:23<29:13,  3.14it/s]
                                                  

  4%|▍         | 247/5758 [01:23<29:13,  3.14it/s]
  4%|▍         | 248/5758 [01:23<28:55,  3.18it/s]
                                                  

  4%|▍         | 248/5758 [01:23<28:55,  3.18it/s]
  4%|▍         | 249/5758 [01:24<28:22,  3.24it/s]
                                                  

  4%|▍         | 249/5758 [01:24<28:22,  3.24it/s]
  4%|▍         | 250/5758 [01:24<30:44,  2.99it/s]
                                                  

  4%|▍         | 250/5758 [01:24<30:44,  2.99it/s]
  4%|▍         | 251/5758 [01:24<29:36,  3.10it/s]
                                                  

  4%|▍         | 251/5758 [01:24<29:36,  3.10it/s]
  4%|▍         | 252/5758 [01:25<28:40,  3.20it/s]
                                                  

  4%|▍         | 252/5758 [01:25<28:40,  3.20it/s]
  4%|▍         | 253/5758 [01:25<30:41,  2.99it/s]
                                                  

  4%|▍         | 253/5758 [01:25<30:41,  2.99it/s]
  4%|▍         | 254/5758 [01:25<29:43,  3.09it/s]
                                                  

  4%|▍         | 254/5758 [01:25<29:43,  3.09it/s]
  4%|▍         | 255/5758 [01:26<33:45,  2.72it/s]
                                                  

  4%|▍         | 255/5758 [01:26<33:45,  2.72it/s]
  4%|▍         | 256/5758 [01:26<33:56,  2.70it/s]
                                                  

  4%|▍         | 256/5758 [01:26<33:56,  2.70it/s]
  4%|▍         | 257/5758 [01:26<32:28,  2.82it/s]
                                                  

  4%|▍         | 257/5758 [01:26<32:28,  2.82it/s]
  4%|▍         | 258/5758 [01:27<31:24,  2.92it/s]
                                                  

  4%|▍         | 258/5758 [01:27<31:24,  2.92it/s]
  4%|▍         | 259/5758 [01:27<29:53,  3.07it/s]
                                                  

  4%|▍         | 259/5758 [01:27<29:53,  3.07it/s]
  5%|▍         | 260/5758 [01:27<28:58,  3.16it/s]
                                                  

  5%|▍         | 260/5758 [01:27<28:58,  3.16it/s]
  5%|▍         | 261/5758 [01:28<28:14,  3.24it/s]
                                                  

  5%|▍         | 261/5758 [01:28<28:14,  3.24it/s]
  5%|▍         | 262/5758 [01:28<28:13,  3.25it/s]
                                                  

  5%|▍         | 262/5758 [01:28<28:13,  3.25it/s]
  5%|▍         | 263/5758 [01:28<29:06,  3.15it/s]
                                                  

  5%|▍         | 263/5758 [01:28<29:06,  3.15it/s]
  5%|▍         | 264/5758 [01:29<28:18,  3.23it/s]
                                                  

  5%|▍         | 264/5758 [01:29<28:18,  3.23it/s]
  5%|▍         | 265/5758 [01:29<28:24,  3.22it/s]
                                                  

  5%|▍         | 265/5758 [01:29<28:24,  3.22it/s]
  5%|▍         | 266/5758 [01:29<27:51,  3.29it/s]
                                                  

  5%|▍         | 266/5758 [01:29<27:51,  3.29it/s]
  5%|▍         | 267/5758 [01:30<28:59,  3.16it/s]
                                                  

  5%|▍         | 267/5758 [01:30<28:59,  3.16it/s]
  5%|▍         | 268/5758 [01:30<28:31,  3.21it/s]
                                                  

  5%|▍         | 268/5758 [01:30<28:31,  3.21it/s]
  5%|▍         | 269/5758 [01:30<29:38,  3.09it/s]
                                                  

  5%|▍         | 269/5758 [01:30<29:38,  3.09it/s]
  5%|▍         | 270/5758 [01:31<31:01,  2.95it/s]
                                                  

  5%|▍         | 270/5758 [01:31<31:01,  2.95it/s]
  5%|▍         | 271/5758 [01:31<32:07,  2.85it/s]
                                                  

  5%|▍         | 271/5758 [01:31<32:07,  2.85it/s]
  5%|▍         | 272/5758 [01:31<31:17,  2.92it/s]
                                                  

  5%|▍         | 272/5758 [01:31<31:17,  2.92it/s]
  5%|▍         | 273/5758 [01:32<30:18,  3.02it/s]
                                                  

  5%|▍         | 273/5758 [01:32<30:18,  3.02it/s]
  5%|▍         | 274/5758 [01:32<29:07,  3.14it/s]
                                                  
{'loss': '1.583', 'grad_norm': '13.12', 'learning_rate': '9.621e-06', 'num_tokens': '1.403e+05', 'mean_token_accuracy': '0.604', 'epoch': '0.03803'}
+{'loss': '1.984', 'grad_norm': '26', 'learning_rate': '9.62e-06', 'num_tokens': '1.405e+05', 'mean_token_accuracy': '0.6503', 'epoch': '0.03821'}
+{'loss': '2.257', 'grad_norm': '18.12', 'learning_rate': '9.618e-06', 'num_tokens': '1.411e+05', 'mean_token_accuracy': '0.5412', 'epoch': '0.03838'}
+{'loss': '1.559', 'grad_norm': '19', 'learning_rate': '9.616e-06', 'num_tokens': '1.413e+05', 'mean_token_accuracy': '0.676', 'epoch': '0.03856'}
+{'loss': '1.574', 'grad_norm': '12.56', 'learning_rate': '9.614e-06', 'num_tokens': '1.42e+05', 'mean_token_accuracy': '0.5888', 'epoch': '0.03873'}
+{'loss': '2.625', 'grad_norm': '12.88', 'learning_rate': '9.613e-06', 'num_tokens': '1.426e+05', 'mean_token_accuracy': '0.5093', 'epoch': '0.0389'}
+{'loss': '1.358', 'grad_norm': '11', 'learning_rate': '9.611e-06', 'num_tokens': '1.435e+05', 'mean_token_accuracy': '0.6381', 'epoch': '0.03908'}
+{'loss': '1.754', 'grad_norm': '21.25', 'learning_rate': '9.609e-06', 'num_tokens': '1.437e+05', 'mean_token_accuracy': '0.6667', 'epoch': '0.03925'}
+{'loss': '1.547', 'grad_norm': '18.12', 'learning_rate': '9.608e-06', 'num_tokens': '1.439e+05', 'mean_token_accuracy': '0.6912', 'epoch': '0.03942'}
+{'loss': '2.686', 'grad_norm': '25.75', 'learning_rate': '9.606e-06', 'num_tokens': '1.441e+05', 'mean_token_accuracy': '0.5244', 'epoch': '0.0396'}
+{'loss': '2.667', 'grad_norm': '12.25', 'learning_rate': '9.604e-06', 'num_tokens': '1.448e+05', 'mean_token_accuracy': '0.4566', 'epoch': '0.03977'}
+{'loss': '1.05', 'grad_norm': '11.75', 'learning_rate': '9.602e-06', 'num_tokens': '1.455e+05', 'mean_token_accuracy': '0.6958', 'epoch': '0.03994'}
+{'loss': '1.376', 'grad_norm': '12.31', 'learning_rate': '9.601e-06', 'num_tokens': '1.463e+05', 'mean_token_accuracy': '0.6202', 'epoch': '0.04012'}
+{'loss': '1.682', 'grad_norm': '18.88', 'learning_rate': '9.599e-06', 'num_tokens': '1.465e+05', 'mean_token_accuracy': '0.6201', 'epoch': '0.04029'}
+{'loss': '2.76', 'grad_norm': '16.25', 'learning_rate': '9.597e-06', 'num_tokens': '1.468e+05', 'mean_token_accuracy': '0.5181', 'epoch': '0.04047'}
+{'loss': '1.428', 'grad_norm': '12.5', 'learning_rate': '9.595e-06', 'num_tokens': '1.475e+05', 'mean_token_accuracy': '0.589', 'epoch': '0.04064'}
+{'loss': '1.653', 'grad_norm': '13.19', 'learning_rate': '9.594e-06', 'num_tokens': '1.482e+05', 'mean_token_accuracy': '0.6054', 'epoch': '0.04081'}
+{'loss': '1.468', 'grad_norm': '10.81', 'learning_rate': '9.592e-06', 'num_tokens': '1.492e+05', 'mean_token_accuracy': '0.6077', 'epoch': '0.04099'}
+{'loss': '1.794', 'grad_norm': '25.38', 'learning_rate': '9.59e-06', 'num_tokens': '1.493e+05', 'mean_token_accuracy': '0.6713', 'epoch': '0.04116'}
+{'loss': '2.095', 'grad_norm': '12', 'learning_rate': '9.588e-06', 'num_tokens': '1.5e+05', 'mean_token_accuracy': '0.5938', 'epoch': '0.04133'}
+{'loss': '1.228', 'grad_norm': '12.19', 'learning_rate': '9.587e-06', 'num_tokens': '1.506e+05', 'mean_token_accuracy': '0.7293', 'epoch': '0.04151'}
+{'loss': '1.575', 'grad_norm': '37', 'learning_rate': '9.585e-06', 'num_tokens': '1.507e+05', 'mean_token_accuracy': '0.6829', 'epoch': '0.04168'}
+{'loss': '1.45', 'grad_norm': '10.44', 'learning_rate': '9.583e-06', 'num_tokens': '1.516e+05', 'mean_token_accuracy': '0.6177', 'epoch': '0.04185'}
+{'loss': '1.32', 'grad_norm': '11.69', 'learning_rate': '9.581e-06', 'num_tokens': '1.524e+05', 'mean_token_accuracy': '0.6463', 'epoch': '0.04203'}
+{'loss': '0.819', 'grad_norm': '8.875', 'learning_rate': '9.58e-06', 'num_tokens': '1.539e+05', 'mean_token_accuracy': '0.7639', 'epoch': '0.0422'}
+{'loss': '1.188', 'grad_norm': '11.56', 'learning_rate': '9.578e-06', 'num_tokens': '1.546e+05', 'mean_token_accuracy': '0.662', 'epoch': '0.04238'}
+{'loss': '1.943', 'grad_norm': '12.38', 'learning_rate': '9.576e-06', 'num_tokens': '1.552e+05', 'mean_token_accuracy': '0.5942', 'epoch': '0.04255'}
+{'loss': '1.363', 'grad_norm': '11.56', 'learning_rate': '9.575e-06', 'num_tokens': '1.56e+05', 'mean_token_accuracy': '0.623', 'epoch': '0.04272'}
+{'loss': '1.758', 'grad_norm': '19.75', 'learning_rate': '9.573e-06', 'num_tokens': '1.562e+05', 'mean_token_accuracy': '0.6333', 'epoch': '0.0429'}
+{'loss': '1.502', 'grad_norm': '16.12', 'learning_rate': '9.571e-06', 'num_tokens': '1.565e+05', 'mean_token_accuracy': '0.6808', 'epoch': '0.04307'}
+{'loss': '2.45', 'grad_norm': '11.75', 'learning_rate': '9.569e-06', 'num_tokens': '1.572e+05', 'mean_token_accuracy': '0.5355', 'epoch': '0.04324'}
+{'loss': '1.972', 'grad_norm': '24.38', 'learning_rate': '9.568e-06', 'num_tokens': '1.573e+05', 'mean_token_accuracy': '0.6403', 'epoch': '0.04342'}
+{'loss': '2.013', 'grad_norm': '14.31', 'learning_rate': '9.566e-06', 'num_tokens': '1.586e+05', 'mean_token_accuracy': '0.5718', 'epoch': '0.04359'}
+{'loss': '1.51', 'grad_norm': '13.38', 'learning_rate': '9.564e-06', 'num_tokens': '1.592e+05', 'mean_token_accuracy': '0.5858', 'epoch': '0.04377'}
+{'loss': '2.558', 'grad_norm': '16.75', 'learning_rate': '9.562e-06', 'num_tokens': '1.598e+05', 'mean_token_accuracy': '0.5087', 'epoch': '0.04394'}
+{'loss': '1.064', 'grad_norm': '11.56', 'learning_rate': '9.561e-06', 'num_tokens': '1.604e+05', 'mean_token_accuracy': '0.672', 'epoch': '0.04411'}
+{'loss': '1.022', 'grad_norm': '13', 'learning_rate': '9.559e-06', 'num_tokens': '1.609e+05', 'mean_token_accuracy': '0.6939', 'epoch': '0.04429'}
+{'loss': '1.488', 'grad_norm': '9.25', 'learning_rate': '9.557e-06', 'num_tokens': '1.621e+05', 'mean_token_accuracy': '0.6584', 'epoch': '0.04446'}
+{'loss': '1.412', 'grad_norm': '10.88', 'learning_rate': '9.555e-06', 'num_tokens': '1.628e+05', 'mean_token_accuracy': '0.621', 'epoch': '0.04463'}
+{'loss': '1.655', 'grad_norm': '12.94', 'learning_rate': '9.554e-06', 'num_tokens': '1.636e+05', 'mean_token_accuracy': '0.5919', 'epoch': '0.04481'}
+{'loss': '2.192', 'grad_norm': '17.38', 'learning_rate': '9.552e-06', 'num_tokens': '1.642e+05', 'mean_token_accuracy': '0.5918', 'epoch': '0.04498'}
+{'loss': '1.55', 'grad_norm': '13.5', 'learning_rate': '9.55e-06', 'num_tokens': '1.648e+05', 'mean_token_accuracy': '0.6351', 'epoch': '0.04515'}
+{'loss': '1.384', 'grad_norm': '18.5', 'learning_rate': '9.548e-06', 'num_tokens': '1.651e+05', 'mean_token_accuracy': '0.6754', 'epoch': '0.04533'}
+{'loss': '0.7988', 'grad_norm': '8.688', 'learning_rate': '9.547e-06', 'num_tokens': '1.665e+05', 'mean_token_accuracy': '0.7885', 'epoch': '0.0455'}
+{'loss': '1.394', 'grad_norm': '10.75', 'learning_rate': '9.545e-06', 'num_tokens': '1.674e+05', 'mean_token_accuracy': '0.6355', 'epoch': '0.04568'}
+{'loss': '1.848', 'grad_norm': '45.5', 'learning_rate': '9.543e-06', 'num_tokens': '1.674e+05', 'mean_token_accuracy': '0.6667', 'epoch': '0.04585'}
+{'loss': '1.752', 'grad_norm': '15.75', 'learning_rate': '9.542e-06', 'num_tokens': '1.682e+05', 'mean_token_accuracy': '0.65', 'epoch': '0.04602'}
+{'loss': '1.666', 'grad_norm': '8.688', 'learning_rate': '9.54e-06', 'num_tokens': '1.694e+05', 'mean_token_accuracy': '0.6491', 'epoch': '0.0462'}
+{'loss': '1.723', 'grad_norm': '12.75', 'learning_rate': '9.538e-06', 'num_tokens': '1.703e+05', 'mean_token_accuracy': '0.6013', 'epoch': '0.04637'}
+{'loss': '1.329', 'grad_norm': '11.69', 'learning_rate': '9.536e-06', 'num_tokens': '1.71e+05', 'mean_token_accuracy': '0.632', 'epoch': '0.04654'}
+{'loss': '1.472', 'grad_norm': '12.75', 'learning_rate': '9.535e-06', 'num_tokens': '1.715e+05', 'mean_token_accuracy': '0.6937', 'epoch': '0.04672'}
+{'loss': '2.376', 'grad_norm': '9.625', 'learning_rate': '9.533e-06', 'num_tokens': '1.726e+05', 'mean_token_accuracy': '0.5416', 'epoch': '0.04689'}
+{'loss': '1.928', 'grad_norm': '19.75', 'learning_rate': '9.531e-06', 'num_tokens': '1.729e+05', 'mean_token_accuracy': '0.6386', 'epoch': '0.04706'}
+{'loss': '2.309', 'grad_norm': '23', 'learning_rate': '9.529e-06', 'num_tokens': '1.731e+05', 'mean_token_accuracy': '0.55', 'epoch': '0.04724'}
+{'loss': '1.54', 'grad_norm': '22.12', 'learning_rate': '9.528e-06', 'num_tokens': '1.733e+05', 'mean_token_accuracy': '0.6973', 'epoch': '0.04741'}
+
  5%|▍         | 274/5758 [01:32<29:07,  3.14it/s]
  5%|▍         | 275/5758 [01:32<30:37,  2.98it/s]
                                                  

  5%|▍         | 275/5758 [01:32<30:37,  2.98it/s]
  5%|▍         | 276/5758 [01:33<30:21,  3.01it/s]
                                                  

  5%|▍         | 276/5758 [01:33<30:21,  3.01it/s]
  5%|▍         | 277/5758 [01:33<31:21,  2.91it/s]
                                                  

  5%|▍         | 277/5758 [01:33<31:21,  2.91it/s]
  5%|▍         | 278/5758 [01:33<31:43,  2.88it/s]
                                                  

  5%|▍         | 278/5758 [01:33<31:43,  2.88it/s]
  5%|▍         | 279/5758 [01:34<31:49,  2.87it/s]
                                                  

  5%|▍         | 279/5758 [01:34<31:49,  2.87it/s]
  5%|▍         | 280/5758 [01:34<30:12,  3.02it/s]
                                                  

  5%|▍         | 280/5758 [01:34<30:12,  3.02it/s]
  5%|▍         | 281/5758 [01:34<34:47,  2.62it/s]
                                                  

  5%|▍         | 281/5758 [01:34<34:47,  2.62it/s]
  5%|▍         | 282/5758 [01:35<33:37,  2.71it/s]
                                                  

  5%|▍         | 282/5758 [01:35<33:37,  2.71it/s]
  5%|▍         | 283/5758 [01:35<33:02,  2.76it/s]
                                                  

  5%|▍         | 283/5758 [01:35<33:02,  2.76it/s]
  5%|▍         | 284/5758 [01:35<31:36,  2.89it/s]
                                                  

  5%|▍         | 284/5758 [01:35<31:36,  2.89it/s]
  5%|▍         | 285/5758 [01:36<31:39,  2.88it/s]
                                                  

  5%|▍         | 285/5758 [01:36<31:39,  2.88it/s]
  5%|▍         | 286/5758 [01:36<30:01,  3.04it/s]
                                                  

  5%|▍         | 286/5758 [01:36<30:01,  3.04it/s]
  5%|▍         | 287/5758 [01:36<29:49,  3.06it/s]
                                                  

  5%|▍         | 287/5758 [01:36<29:49,  3.06it/s]
  5%|▌         | 288/5758 [01:37<29:13,  3.12it/s]
                                                  

  5%|▌         | 288/5758 [01:37<29:13,  3.12it/s]
  5%|▌         | 289/5758 [01:37<29:35,  3.08it/s]
                                                  

  5%|▌         | 289/5758 [01:37<29:35,  3.08it/s]
  5%|▌         | 290/5758 [01:37<30:17,  3.01it/s]
                                                  

  5%|▌         | 290/5758 [01:37<30:17,  3.01it/s]
  5%|▌         | 291/5758 [01:38<29:00,  3.14it/s]
                                                  

  5%|▌         | 291/5758 [01:38<29:00,  3.14it/s]
  5%|▌         | 292/5758 [01:38<28:47,  3.16it/s]
                                                  

  5%|▌         | 292/5758 [01:38<28:47,  3.16it/s]
  5%|▌         | 293/5758 [01:38<30:16,  3.01it/s]
                                                  

  5%|▌         | 293/5758 [01:38<30:16,  3.01it/s]
  5%|▌         | 294/5758 [01:39<29:39,  3.07it/s]
                                                  

  5%|▌         | 294/5758 [01:39<29:39,  3.07it/s]
  5%|▌         | 295/5758 [01:39<30:59,  2.94it/s]
                                                  

  5%|▌         | 295/5758 [01:39<30:59,  2.94it/s]
  5%|▌         | 296/5758 [01:39<29:54,  3.04it/s]
                                                  

  5%|▌         | 296/5758 [01:39<29:54,  3.04it/s]
  5%|▌         | 297/5758 [01:40<30:02,  3.03it/s]
                                                  

  5%|▌         | 297/5758 [01:40<30:02,  3.03it/s]
  5%|▌         | 298/5758 [01:40<30:17,  3.00it/s]
                                                  

  5%|▌         | 298/5758 [01:40<30:17,  3.00it/s]
  5%|▌         | 299/5758 [01:40<29:57,  3.04it/s]
                                                  

  5%|▌         | 299/5758 [01:40<29:57,  3.04it/s]
  5%|▌         | 300/5758 [01:41<29:59,  3.03it/s]
                                                  

  5%|▌         | 300/5758 [01:41<29:59,  3.03it/s]
  5%|▌         | 301/5758 [01:41<29:18,  3.10it/s]
                                                  

  5%|▌         | 301/5758 [01:41<29:18,  3.10it/s]
  5%|▌         | 302/5758 [01:41<28:42,  3.17it/s]
                                                  

  5%|▌         | 302/5758 [01:41<28:42,  3.17it/s]
  5%|▌         | 303/5758 [01:42<30:28,  2.98it/s]
                                                  

  5%|▌         | 303/5758 [01:42<30:28,  2.98it/s]
  5%|▌         | 304/5758 [01:42<30:59,  2.93it/s]
                                                  

  5%|▌         | 304/5758 [01:42<30:59,  2.93it/s]
  5%|▌         | 305/5758 [01:42<30:11,  3.01it/s]
                                                  

  5%|▌         | 305/5758 [01:42<30:11,  3.01it/s]
  5%|▌         | 306/5758 [01:43<28:43,  3.16it/s]
                                                  

  5%|▌         | 306/5758 [01:43<28:43,  3.16it/s]
  5%|▌         | 307/5758 [01:43<28:00,  3.24it/s]
                                                  

  5%|▌         | 307/5758 [01:43<28:00,  3.24it/s]
  5%|▌         | 308/5758 [01:43<30:41,  2.96it/s]
                                                  

  5%|▌         | 308/5758 [01:43<30:41,  2.96it/s]
  5%|▌         | 309/5758 [01:44<31:08,  2.92it/s]
                                                  

  5%|▌         | 309/5758 [01:44<31:08,  2.92it/s]
  5%|▌         | 310/5758 [01:44<29:52,  3.04it/s]
                                                  

  5%|▌         | 310/5758 [01:44<29:52,  3.04it/s]
  5%|▌         | 311/5758 [01:44<29:30,  3.08it/s]
                                                  

  5%|▌         | 311/5758 [01:44<29:30,  3.08it/s]
  5%|▌         | 312/5758 [01:45<30:01,  3.02it/s]
                                                  

  5%|▌         | 312/5758 [01:45<30:01,  3.02it/s]
  5%|▌         | 313/5758 [01:45<30:39,  2.96it/s]
                                                  

  5%|▌         | 313/5758 [01:45<30:39,  2.96it/s]
  5%|▌         | 314/5758 [01:45<29:40,  3.06it/s]
                                                  

  5%|▌         | 314/5758 [01:45<29:40,  3.06it/s]
  5%|▌         | 315/5758 [01:46<28:56,  3.13it/s]
                                                  

  5%|▌         | 315/5758 [01:46<28:56,  3.13it/s]
  5%|▌         | 316/5758 [01:46<28:35,  3.17it/s]
                                                  

  5%|▌         | 316/5758 [01:46<28:35,  3.17it/s]
  6%|▌         | 317/5758 [01:46<28:52,  3.14it/s]
                                                  

  6%|▌         | 317/5758 [01:46<28:52,  3.14it/s]
  6%|▌         | 318/5758 [01:47<32:36,  2.78it/s]
                                                  

  6%|▌         | 318/5758 [01:47<32:36,  2.78it/s]
  6%|▌         | 319/5758 [01:47<31:26,  2.88it/s]
                                                  

  6%|▌         | 319/5758 [01:47<31:26,  2.88it/s]
  6%|▌         | 320/5758 [01:47<31:25,  2.88it/s]
                                                  

  6%|▌         | 320/5758 [01:47<31:25,  2.88it/s]
  6%|▌         | 321/5758 [01:48<30:32,  2.97it/s]
                                                  

  6%|▌         | 321/5758 [01:48<30:32,  2.97it/s]
  6%|▌         | 322/5758 [01:48<31:29,  2.88it/s]
                                                  

  6%|▌         | 322/5758 [01:48<31:29,  2.88it/s]
  6%|▌         | 323/5758 [01:48<31:53,  2.84it/s]
                                                  

  6%|▌         | 323/5758 [01:48<31:53,  2.84it/s]
  6%|▌         | 324/5758 [01:49<32:05,  2.82it/s]
                                                  

  6%|▌         | 324/5758 [01:49<32:05,  2.82it/s]
  6%|▌         | 325/5758 [01:49<32:34,  2.78it/s]
                                                  

  6%|▌         | 325/5758 [01:49<32:34,  2.78it/s]
  6%|▌         | 326/5758 [01:49<30:47,  2.94it/s]
                                                  

  6%|▌         | 326/5758 [01:49<30:47,  2.94it/s]
  6%|▌         | 327/5758 [01:50<29:26,  3.07it/s]
                                                  

  6%|▌         | 327/5758 [01:50<29:26,  3.07it/s]
  6%|▌         | 328/5758 [01:50<28:37,  3.16it/s]
                                                  

  6%|▌         | 328/5758 [01:50<28:37,  3.16it/s]
  6%|▌         | 329/5758 [01:50<28:10,  3.21it/s]
                                                  
{'loss': '2.371', 'grad_norm': '19.75', 'learning_rate': '9.526e-06', 'num_tokens': '1.736e+05', 'mean_token_accuracy': '0.5463', 'epoch': '0.04759'}
+{'loss': '1.852', 'grad_norm': '12.25', 'learning_rate': '9.524e-06', 'num_tokens': '1.741e+05', 'mean_token_accuracy': '0.6351', 'epoch': '0.04776'}
+{'loss': '2.144', 'grad_norm': '23', 'learning_rate': '9.522e-06', 'num_tokens': '1.743e+05', 'mean_token_accuracy': '0.5736', 'epoch': '0.04793'}
+{'loss': '0.943', 'grad_norm': '9', 'learning_rate': '9.521e-06', 'num_tokens': '1.754e+05', 'mean_token_accuracy': '0.74', 'epoch': '0.04811'}
+{'loss': '2.209', 'grad_norm': '28.5', 'learning_rate': '9.519e-06', 'num_tokens': '1.756e+05', 'mean_token_accuracy': '0.5965', 'epoch': '0.04828'}
+{'loss': '2.13', 'grad_norm': '14.56', 'learning_rate': '9.517e-06', 'num_tokens': '1.76e+05', 'mean_token_accuracy': '0.5717', 'epoch': '0.04845'}
+{'loss': '2.161', 'grad_norm': '12.12', 'learning_rate': '9.515e-06', 'num_tokens': '1.767e+05', 'mean_token_accuracy': '0.527', 'epoch': '0.04863'}
+{'loss': '1.17', 'grad_norm': '18.12', 'learning_rate': '9.514e-06', 'num_tokens': '1.769e+05', 'mean_token_accuracy': '0.7363', 'epoch': '0.0488'}
+{'loss': '2.822', 'grad_norm': '20.25', 'learning_rate': '9.512e-06', 'num_tokens': '1.774e+05', 'mean_token_accuracy': '0.4641', 'epoch': '0.04898'}
+{'loss': '1.615', 'grad_norm': '11.06', 'learning_rate': '9.51e-06', 'num_tokens': '1.783e+05', 'mean_token_accuracy': '0.5849', 'epoch': '0.04915'}
+{'loss': '1.166', 'grad_norm': '20.5', 'learning_rate': '9.509e-06', 'num_tokens': '1.785e+05', 'mean_token_accuracy': '0.7283', 'epoch': '0.04932'}
+{'loss': '2.267', 'grad_norm': '19.62', 'learning_rate': '9.507e-06', 'num_tokens': '1.79e+05', 'mean_token_accuracy': '0.5767', 'epoch': '0.0495'}
+{'loss': '2.353', 'grad_norm': '29.75', 'learning_rate': '9.505e-06', 'num_tokens': '1.791e+05', 'mean_token_accuracy': '0.5867', 'epoch': '0.04967'}
+{'loss': '1.265', 'grad_norm': '11.25', 'learning_rate': '9.503e-06', 'num_tokens': '1.799e+05', 'mean_token_accuracy': '0.6739', 'epoch': '0.04984'}
+{'loss': '1.081', 'grad_norm': '12.44', 'learning_rate': '9.502e-06', 'num_tokens': '1.806e+05', 'mean_token_accuracy': '0.6816', 'epoch': '0.05002'}
+{'loss': '1.51', 'grad_norm': '11.12', 'learning_rate': '9.5e-06', 'num_tokens': '1.815e+05', 'mean_token_accuracy': '0.607', 'epoch': '0.05019'}
+{'loss': '1.391', 'grad_norm': '14.06', 'learning_rate': '9.498e-06', 'num_tokens': '1.82e+05', 'mean_token_accuracy': '0.6331', 'epoch': '0.05036'}
+{'loss': '1.314', 'grad_norm': '12.25', 'learning_rate': '9.496e-06', 'num_tokens': '1.826e+05', 'mean_token_accuracy': '0.6507', 'epoch': '0.05054'}
+{'loss': '0.9597', 'grad_norm': '11.56', 'learning_rate': '9.495e-06', 'num_tokens': '1.833e+05', 'mean_token_accuracy': '0.7125', 'epoch': '0.05071'}
+{'loss': '2.236', 'grad_norm': '10.94', 'learning_rate': '9.493e-06', 'num_tokens': '1.844e+05', 'mean_token_accuracy': '0.5737', 'epoch': '0.05089'}
+{'loss': '2.125', 'grad_norm': '17', 'learning_rate': '9.491e-06', 'num_tokens': '1.847e+05', 'mean_token_accuracy': '0.5706', 'epoch': '0.05106'}
+{'loss': '2.41', 'grad_norm': '15.69', 'learning_rate': '9.489e-06', 'num_tokens': '1.853e+05', 'mean_token_accuracy': '0.5284', 'epoch': '0.05123'}
+{'loss': '1.486', 'grad_norm': '14.69', 'learning_rate': '9.488e-06', 'num_tokens': '1.86e+05', 'mean_token_accuracy': '0.6239', 'epoch': '0.05141'}
+{'loss': '1.484', 'grad_norm': '10.38', 'learning_rate': '9.486e-06', 'num_tokens': '1.868e+05', 'mean_token_accuracy': '0.6757', 'epoch': '0.05158'}
+{'loss': '1.581', 'grad_norm': '17.12', 'learning_rate': '9.484e-06', 'num_tokens': '1.871e+05', 'mean_token_accuracy': '0.6723', 'epoch': '0.05175'}
+{'loss': '2.018', 'grad_norm': '40', 'learning_rate': '9.482e-06', 'num_tokens': '1.872e+05', 'mean_token_accuracy': '0.6087', 'epoch': '0.05193'}
+{'loss': '1.288', 'grad_norm': '11.56', 'learning_rate': '9.481e-06', 'num_tokens': '1.88e+05', 'mean_token_accuracy': '0.6632', 'epoch': '0.0521'}
+{'loss': '1.05', 'grad_norm': '13.44', 'learning_rate': '9.479e-06', 'num_tokens': '1.884e+05', 'mean_token_accuracy': '0.7548', 'epoch': '0.05228'}
+{'loss': '1.302', 'grad_norm': '17.25', 'learning_rate': '9.477e-06', 'num_tokens': '1.887e+05', 'mean_token_accuracy': '0.6766', 'epoch': '0.05245'}
+{'loss': '1.583', 'grad_norm': '14.5', 'learning_rate': '9.476e-06', 'num_tokens': '1.893e+05', 'mean_token_accuracy': '0.5754', 'epoch': '0.05262'}
+{'loss': '1.682', 'grad_norm': '30', 'learning_rate': '9.474e-06', 'num_tokens': '1.894e+05', 'mean_token_accuracy': '0.6', 'epoch': '0.0528'}
+{'loss': '1.323', 'grad_norm': '12.5', 'learning_rate': '9.472e-06', 'num_tokens': '1.902e+05', 'mean_token_accuracy': '0.6388', 'epoch': '0.05297'}
+{'loss': '1.29', 'grad_norm': '42.5', 'learning_rate': '9.47e-06', 'num_tokens': '1.902e+05', 'mean_token_accuracy': '0.7049', 'epoch': '0.05314'}
+{'loss': '1.306', 'grad_norm': '12.88', 'learning_rate': '9.469e-06', 'num_tokens': '1.909e+05', 'mean_token_accuracy': '0.62', 'epoch': '0.05332'}
+{'loss': '1.904', 'grad_norm': '21.5', 'learning_rate': '9.467e-06', 'num_tokens': '1.912e+05', 'mean_token_accuracy': '0.5795', 'epoch': '0.05349'}
+{'loss': '2.191', 'grad_norm': '10.06', 'learning_rate': '9.465e-06', 'num_tokens': '1.922e+05', 'mean_token_accuracy': '0.5457', 'epoch': '0.05366'}
+{'loss': '2.85', 'grad_norm': '13.25', 'learning_rate': '9.463e-06', 'num_tokens': '1.929e+05', 'mean_token_accuracy': '0.4738', 'epoch': '0.05384'}
+{'loss': '1.188', 'grad_norm': '10.94', 'learning_rate': '9.462e-06', 'num_tokens': '1.936e+05', 'mean_token_accuracy': '0.6841', 'epoch': '0.05401'}
+{'loss': '1.791', 'grad_norm': '12.31', 'learning_rate': '9.46e-06', 'num_tokens': '1.946e+05', 'mean_token_accuracy': '0.6506', 'epoch': '0.05419'}
+{'loss': '1.622', 'grad_norm': '12.75', 'learning_rate': '9.458e-06', 'num_tokens': '1.95e+05', 'mean_token_accuracy': '0.666', 'epoch': '0.05436'}
+{'loss': '1.282', 'grad_norm': '13.69', 'learning_rate': '9.456e-06', 'num_tokens': '1.957e+05', 'mean_token_accuracy': '0.6188', 'epoch': '0.05453'}
+{'loss': '2.471', 'grad_norm': '15.88', 'learning_rate': '9.455e-06', 'num_tokens': '1.963e+05', 'mean_token_accuracy': '0.5392', 'epoch': '0.05471'}
+{'loss': '1.49', 'grad_norm': '41', 'learning_rate': '9.453e-06', 'num_tokens': '1.964e+05', 'mean_token_accuracy': '0.6951', 'epoch': '0.05488'}
+{'loss': '2.034', 'grad_norm': '10.94', 'learning_rate': '9.451e-06', 'num_tokens': '1.973e+05', 'mean_token_accuracy': '0.6005', 'epoch': '0.05505'}
+{'loss': '1.724', 'grad_norm': '10.44', 'learning_rate': '9.449e-06', 'num_tokens': '1.984e+05', 'mean_token_accuracy': '0.5703', 'epoch': '0.05523'}
+{'loss': '2.037', 'grad_norm': '22.75', 'learning_rate': '9.448e-06', 'num_tokens': '1.987e+05', 'mean_token_accuracy': '0.5688', 'epoch': '0.0554'}
+{'loss': '2.38', 'grad_norm': '12.44', 'learning_rate': '9.446e-06', 'num_tokens': '1.992e+05', 'mean_token_accuracy': '0.5458', 'epoch': '0.05557'}
+{'loss': '1.776', 'grad_norm': '14.75', 'learning_rate': '9.444e-06', 'num_tokens': '1.996e+05', 'mean_token_accuracy': '0.6419', 'epoch': '0.05575'}
+{'loss': '1.232', 'grad_norm': '21.38', 'learning_rate': '9.443e-06', 'num_tokens': '1.999e+05', 'mean_token_accuracy': '0.7116', 'epoch': '0.05592'}
+{'loss': '1.206', 'grad_norm': '11.81', 'learning_rate': '9.441e-06', 'num_tokens': '2.004e+05', 'mean_token_accuracy': '0.7126', 'epoch': '0.0561'}
+{'loss': '2.089', 'grad_norm': '15.31', 'learning_rate': '9.439e-06', 'num_tokens': '2.009e+05', 'mean_token_accuracy': '0.5618', 'epoch': '0.05627'}
+{'loss': '1.917', 'grad_norm': '15.5', 'learning_rate': '9.437e-06', 'num_tokens': '2.014e+05', 'mean_token_accuracy': '0.5888', 'epoch': '0.05644'}
+{'loss': '1.438', 'grad_norm': '13.25', 'learning_rate': '9.436e-06', 'num_tokens': '2.021e+05', 'mean_token_accuracy': '0.6026', 'epoch': '0.05662'}
+{'loss': '0.9139', 'grad_norm': '17.88', 'learning_rate': '9.434e-06', 'num_tokens': '2.022e+05', 'mean_token_accuracy': '0.8397', 'epoch': '0.05679'}
+{'loss': '0.7476', 'grad_norm': '8.438', 'learning_rate': '9.432e-06', 'num_tokens': '2.035e+05', 'mean_token_accuracy': '0.7871', 'epoch': '0.05696'}
+
  6%|▌         | 329/5758 [01:50<28:10,  3.21it/s]
  6%|▌         | 330/5758 [01:51<27:27,  3.29it/s]
                                                  

  6%|▌         | 330/5758 [01:51<27:27,  3.29it/s]
  6%|▌         | 331/5758 [01:51<28:16,  3.20it/s]
                                                  

  6%|▌         | 331/5758 [01:51<28:16,  3.20it/s]
  6%|▌         | 332/5758 [01:51<27:56,  3.24it/s]
                                                  

  6%|▌         | 332/5758 [01:51<27:56,  3.24it/s]
  6%|▌         | 333/5758 [01:52<28:53,  3.13it/s]
                                                  

  6%|▌         | 333/5758 [01:52<28:53,  3.13it/s]
  6%|▌         | 334/5758 [01:52<31:35,  2.86it/s]
                                                  

  6%|▌         | 334/5758 [01:52<31:35,  2.86it/s]
  6%|▌         | 335/5758 [01:52<31:45,  2.85it/s]
                                                  

  6%|▌         | 335/5758 [01:52<31:45,  2.85it/s]
  6%|▌         | 336/5758 [01:53<30:13,  2.99it/s]
                                                  

  6%|▌         | 336/5758 [01:53<30:13,  2.99it/s]
  6%|▌         | 337/5758 [01:53<29:19,  3.08it/s]
                                                  

  6%|▌         | 337/5758 [01:53<29:19,  3.08it/s]
  6%|▌         | 338/5758 [01:53<30:42,  2.94it/s]
                                                  

  6%|▌         | 338/5758 [01:53<30:42,  2.94it/s]
  6%|▌         | 339/5758 [01:54<30:02,  3.01it/s]
                                                  

  6%|▌         | 339/5758 [01:54<30:02,  3.01it/s]
  6%|▌         | 340/5758 [01:54<31:00,  2.91it/s]
                                                  

  6%|▌         | 340/5758 [01:54<31:00,  2.91it/s]
  6%|▌         | 341/5758 [01:54<31:46,  2.84it/s]
                                                  

  6%|▌         | 341/5758 [01:54<31:46,  2.84it/s]
  6%|▌         | 342/5758 [01:55<34:13,  2.64it/s]
                                                  

  6%|▌         | 342/5758 [01:55<34:13,  2.64it/s]
  6%|▌         | 343/5758 [01:55<33:08,  2.72it/s]
                                                  

  6%|▌         | 343/5758 [01:55<33:08,  2.72it/s]
  6%|▌         | 344/5758 [01:55<33:10,  2.72it/s]
                                                  

  6%|▌         | 344/5758 [01:55<33:10,  2.72it/s]
  6%|▌         | 345/5758 [01:56<31:26,  2.87it/s]
                                                  

  6%|▌         | 345/5758 [01:56<31:26,  2.87it/s]
  6%|▌         | 346/5758 [01:56<29:58,  3.01it/s]
                                                  

  6%|▌         | 346/5758 [01:56<29:58,  3.01it/s]
  6%|▌         | 347/5758 [01:56<30:31,  2.95it/s]
                                                  

  6%|▌         | 347/5758 [01:56<30:31,  2.95it/s]
  6%|▌         | 348/5758 [01:57<30:36,  2.95it/s]
                                                  

  6%|▌         | 348/5758 [01:57<30:36,  2.95it/s]
  6%|▌         | 349/5758 [01:57<29:31,  3.05it/s]
                                                  

  6%|▌         | 349/5758 [01:57<29:31,  3.05it/s]
  6%|▌         | 350/5758 [01:57<30:10,  2.99it/s]
                                                  

  6%|▌         | 350/5758 [01:57<30:10,  2.99it/s]
  6%|▌         | 351/5758 [01:58<29:57,  3.01it/s]
                                                  

  6%|▌         | 351/5758 [01:58<29:57,  3.01it/s]
  6%|▌         | 352/5758 [01:58<31:11,  2.89it/s]
                                                  

  6%|▌         | 352/5758 [01:58<31:11,  2.89it/s]
  6%|▌         | 353/5758 [01:58<30:23,  2.96it/s]
                                                  

  6%|▌         | 353/5758 [01:58<30:23,  2.96it/s]
  6%|▌         | 354/5758 [01:59<29:44,  3.03it/s]
                                                  

  6%|▌         | 354/5758 [01:59<29:44,  3.03it/s]
  6%|▌         | 355/5758 [01:59<28:58,  3.11it/s]
                                                  

  6%|▌         | 355/5758 [01:59<28:58,  3.11it/s]
  6%|▌         | 356/5758 [01:59<29:33,  3.05it/s]
                                                  

  6%|▌         | 356/5758 [01:59<29:33,  3.05it/s]
  6%|▌         | 357/5758 [02:00<31:55,  2.82it/s]
                                                  

  6%|▌         | 357/5758 [02:00<31:55,  2.82it/s]
  6%|▌         | 358/5758 [02:00<31:15,  2.88it/s]
                                                  

  6%|▌         | 358/5758 [02:00<31:15,  2.88it/s]
  6%|▌         | 359/5758 [02:00<30:39,  2.93it/s]
                                                  

  6%|▌         | 359/5758 [02:00<30:39,  2.93it/s]
  6%|▋         | 360/5758 [02:01<30:49,  2.92it/s]
                                                  

  6%|▋         | 360/5758 [02:01<30:49,  2.92it/s]
  6%|▋         | 361/5758 [02:01<29:21,  3.06it/s]
                                                  

  6%|▋         | 361/5758 [02:01<29:21,  3.06it/s]
  6%|▋         | 362/5758 [02:01<29:00,  3.10it/s]
                                                  

  6%|▋         | 362/5758 [02:01<29:00,  3.10it/s]
  6%|▋         | 363/5758 [02:02<28:49,  3.12it/s]
                                                  

  6%|▋         | 363/5758 [02:02<28:49,  3.12it/s]
  6%|▋         | 364/5758 [02:02<27:54,  3.22it/s]
                                                  

  6%|▋         | 364/5758 [02:02<27:54,  3.22it/s]
  6%|▋         | 365/5758 [02:02<29:14,  3.07it/s]
                                                  

  6%|▋         | 365/5758 [02:02<29:14,  3.07it/s]
  6%|▋         | 366/5758 [02:03<28:42,  3.13it/s]
                                                  

  6%|▋         | 366/5758 [02:03<28:42,  3.13it/s]
  6%|▋         | 367/5758 [02:03<28:47,  3.12it/s]
                                                  

  6%|▋         | 367/5758 [02:03<28:47,  3.12it/s]
  6%|▋         | 368/5758 [02:03<28:22,  3.17it/s]
                                                  

  6%|▋         | 368/5758 [02:03<28:22,  3.17it/s]
  6%|▋         | 369/5758 [02:04<28:28,  3.15it/s]
                                                  

  6%|▋         | 369/5758 [02:04<28:28,  3.15it/s]
  6%|▋         | 370/5758 [02:04<28:46,  3.12it/s]
                                                  

  6%|▋         | 370/5758 [02:04<28:46,  3.12it/s]
  6%|▋         | 371/5758 [02:04<29:40,  3.03it/s]
                                                  

  6%|▋         | 371/5758 [02:04<29:40,  3.03it/s]
  6%|▋         | 372/5758 [02:05<31:28,  2.85it/s]
                                                  

  6%|▋         | 372/5758 [02:05<31:28,  2.85it/s]
  6%|▋         | 373/5758 [02:05<31:47,  2.82it/s]
                                                  

  6%|▋         | 373/5758 [02:05<31:47,  2.82it/s]
  6%|▋         | 374/5758 [02:05<32:27,  2.77it/s]
                                                  

  6%|▋         | 374/5758 [02:05<32:27,  2.77it/s]
  7%|▋         | 375/5758 [02:06<33:10,  2.70it/s]
                                                  

  7%|▋         | 375/5758 [02:06<33:10,  2.70it/s]
  7%|▋         | 376/5758 [02:06<32:12,  2.79it/s]
                                                  

  7%|▋         | 376/5758 [02:06<32:12,  2.79it/s]
  7%|▋         | 377/5758 [02:07<32:43,  2.74it/s]
                                                  

  7%|▋         | 377/5758 [02:07<32:43,  2.74it/s]
  7%|▋         | 378/5758 [02:07<30:46,  2.91it/s]
                                                  

  7%|▋         | 378/5758 [02:07<30:46,  2.91it/s]
  7%|▋         | 379/5758 [02:07<29:43,  3.02it/s]
                                                  

  7%|▋         | 379/5758 [02:07<29:43,  3.02it/s]
  7%|▋         | 380/5758 [02:07<28:38,  3.13it/s]
                                                  

  7%|▋         | 380/5758 [02:07<28:38,  3.13it/s]
  7%|▋         | 381/5758 [02:08<28:06,  3.19it/s]
                                                  

  7%|▋         | 381/5758 [02:08<28:06,  3.19it/s]
  7%|▋         | 382/5758 [02:08<27:30,  3.26it/s]
                                                  

  7%|▋         | 382/5758 [02:08<27:30,  3.26it/s]
  7%|▋         | 383/5758 [02:08<31:18,  2.86it/s]
                                                  
{'loss': '1.496', 'grad_norm': '12.94', 'learning_rate': '9.43e-06', 'num_tokens': '2.042e+05', 'mean_token_accuracy': '0.6089', 'epoch': '0.05714'}
+{'loss': '1.176', 'grad_norm': '27.5', 'learning_rate': '9.429e-06', 'num_tokens': '2.043e+05', 'mean_token_accuracy': '0.7248', 'epoch': '0.05731'}
+{'loss': '1.564', 'grad_norm': '16.38', 'learning_rate': '9.427e-06', 'num_tokens': '2.048e+05', 'mean_token_accuracy': '0.5873', 'epoch': '0.05749'}
+{'loss': '1.993', 'grad_norm': '12', 'learning_rate': '9.425e-06', 'num_tokens': '2.055e+05', 'mean_token_accuracy': '0.6272', 'epoch': '0.05766'}
+{'loss': '2.409', 'grad_norm': '18.88', 'learning_rate': '9.423e-06', 'num_tokens': '2.059e+05', 'mean_token_accuracy': '0.5353', 'epoch': '0.05783'}
+{'loss': '0.7366', 'grad_norm': '12.75', 'learning_rate': '9.422e-06', 'num_tokens': '2.066e+05', 'mean_token_accuracy': '0.812', 'epoch': '0.05801'}
+{'loss': '1.234', 'grad_norm': '11.69', 'learning_rate': '9.42e-06', 'num_tokens': '2.076e+05', 'mean_token_accuracy': '0.6537', 'epoch': '0.05818'}
+{'loss': '0.8733', 'grad_norm': '8.688', 'learning_rate': '9.418e-06', 'num_tokens': '2.089e+05', 'mean_token_accuracy': '0.7258', 'epoch': '0.05835'}
+{'loss': '1.555', 'grad_norm': '24.5', 'learning_rate': '9.416e-06', 'num_tokens': '2.091e+05', 'mean_token_accuracy': '0.6946', 'epoch': '0.05853'}
+{'loss': '1.947', 'grad_norm': '18.25', 'learning_rate': '9.415e-06', 'num_tokens': '2.094e+05', 'mean_token_accuracy': '0.5949', 'epoch': '0.0587'}
+{'loss': '1.385', 'grad_norm': '11.75', 'learning_rate': '9.413e-06', 'num_tokens': '2.102e+05', 'mean_token_accuracy': '0.6258', 'epoch': '0.05887'}
+{'loss': '2.045', 'grad_norm': '15.19', 'learning_rate': '9.411e-06', 'num_tokens': '2.107e+05', 'mean_token_accuracy': '0.5774', 'epoch': '0.05905'}
+{'loss': '1.256', 'grad_norm': '9.125', 'learning_rate': '9.41e-06', 'num_tokens': '2.117e+05', 'mean_token_accuracy': '0.6989', 'epoch': '0.05922'}
+{'loss': '1.356', 'grad_norm': '7.656', 'learning_rate': '9.408e-06', 'num_tokens': '2.138e+05', 'mean_token_accuracy': '0.6683', 'epoch': '0.0594'}
+{'loss': '1.306', 'grad_norm': '11.19', 'learning_rate': '9.406e-06', 'num_tokens': '2.147e+05', 'mean_token_accuracy': '0.634', 'epoch': '0.05957'}
+{'loss': '1.261', 'grad_norm': '12.5', 'learning_rate': '9.404e-06', 'num_tokens': '2.152e+05', 'mean_token_accuracy': '0.6981', 'epoch': '0.05974'}
+{'loss': '2.44', 'grad_norm': '19.25', 'learning_rate': '9.403e-06', 'num_tokens': '2.155e+05', 'mean_token_accuracy': '0.5', 'epoch': '0.05992'}
+{'loss': '1.152', 'grad_norm': '13.44', 'learning_rate': '9.401e-06', 'num_tokens': '2.161e+05', 'mean_token_accuracy': '0.6502', 'epoch': '0.06009'}
+{'loss': '1.531', 'grad_norm': '32.25', 'learning_rate': '9.399e-06', 'num_tokens': '2.162e+05', 'mean_token_accuracy': '0.6697', 'epoch': '0.06026'}
+{'loss': '1.279', 'grad_norm': '13.31', 'learning_rate': '9.397e-06', 'num_tokens': '2.167e+05', 'mean_token_accuracy': '0.7424', 'epoch': '0.06044'}
+{'loss': '1.302', 'grad_norm': '12.81', 'learning_rate': '9.396e-06', 'num_tokens': '2.174e+05', 'mean_token_accuracy': '0.6467', 'epoch': '0.06061'}
+{'loss': '0.9318', 'grad_norm': '12.69', 'learning_rate': '9.394e-06', 'num_tokens': '2.178e+05', 'mean_token_accuracy': '0.7578', 'epoch': '0.06078'}
+{'loss': '1.416', 'grad_norm': '11.94', 'learning_rate': '9.392e-06', 'num_tokens': '2.187e+05', 'mean_token_accuracy': '0.5973', 'epoch': '0.06096'}
+{'loss': '1.001', 'grad_norm': '24.62', 'learning_rate': '9.39e-06', 'num_tokens': '2.188e+05', 'mean_token_accuracy': '0.8', 'epoch': '0.06113'}
+{'loss': '1.142', 'grad_norm': '12', 'learning_rate': '9.389e-06', 'num_tokens': '2.195e+05', 'mean_token_accuracy': '0.6566', 'epoch': '0.06131'}
+{'loss': '2.434', 'grad_norm': '12.5', 'learning_rate': '9.387e-06', 'num_tokens': '2.202e+05', 'mean_token_accuracy': '0.5365', 'epoch': '0.06148'}
+{'loss': '1.568', 'grad_norm': '12.69', 'learning_rate': '9.385e-06', 'num_tokens': '2.209e+05', 'mean_token_accuracy': '0.6197', 'epoch': '0.06165'}
+{'loss': '2.243', 'grad_norm': '9.438', 'learning_rate': '9.383e-06', 'num_tokens': '2.227e+05', 'mean_token_accuracy': '0.5314', 'epoch': '0.06183'}
+{'loss': '1.359', 'grad_norm': '13', 'learning_rate': '9.382e-06', 'num_tokens': '2.234e+05', 'mean_token_accuracy': '0.6252', 'epoch': '0.062'}
+{'loss': '0.7355', 'grad_norm': '7.812', 'learning_rate': '9.38e-06', 'num_tokens': '2.25e+05', 'mean_token_accuracy': '0.7783', 'epoch': '0.06217'}
+{'loss': '0.6773', 'grad_norm': '7.875', 'learning_rate': '9.378e-06', 'num_tokens': '2.266e+05', 'mean_token_accuracy': '0.8019', 'epoch': '0.06235'}
+{'loss': '2.237', 'grad_norm': '12', 'learning_rate': '9.377e-06', 'num_tokens': '2.276e+05', 'mean_token_accuracy': '0.5358', 'epoch': '0.06252'}
+{'loss': '1.344', 'grad_norm': '17.5', 'learning_rate': '9.375e-06', 'num_tokens': '2.279e+05', 'mean_token_accuracy': '0.7043', 'epoch': '0.0627'}
+{'loss': '0.6078', 'grad_norm': '7.438', 'learning_rate': '9.373e-06', 'num_tokens': '2.293e+05', 'mean_token_accuracy': '0.8116', 'epoch': '0.06287'}
+{'loss': '0.7092', 'grad_norm': '8.812', 'learning_rate': '9.371e-06', 'num_tokens': '2.309e+05', 'mean_token_accuracy': '0.7911', 'epoch': '0.06304'}
+{'loss': '1.552', 'grad_norm': '43.75', 'learning_rate': '9.37e-06', 'num_tokens': '2.309e+05', 'mean_token_accuracy': '0.6866', 'epoch': '0.06322'}
+{'loss': '1.159', 'grad_norm': '12.88', 'learning_rate': '9.368e-06', 'num_tokens': '2.314e+05', 'mean_token_accuracy': '0.7375', 'epoch': '0.06339'}
+{'loss': '1.413', 'grad_norm': '13.06', 'learning_rate': '9.366e-06', 'num_tokens': '2.321e+05', 'mean_token_accuracy': '0.637', 'epoch': '0.06356'}
+{'loss': '2.01', 'grad_norm': '12.88', 'learning_rate': '9.364e-06', 'num_tokens': '2.329e+05', 'mean_token_accuracy': '0.5718', 'epoch': '0.06374'}
+{'loss': '1.302', 'grad_norm': '23.12', 'learning_rate': '9.363e-06', 'num_tokens': '2.331e+05', 'mean_token_accuracy': '0.7368', 'epoch': '0.06391'}
+{'loss': '1.039', 'grad_norm': '10.38', 'learning_rate': '9.361e-06', 'num_tokens': '2.339e+05', 'mean_token_accuracy': '0.7303', 'epoch': '0.06408'}
+{'loss': '2.132', 'grad_norm': '15.5', 'learning_rate': '9.359e-06', 'num_tokens': '2.343e+05', 'mean_token_accuracy': '0.5642', 'epoch': '0.06426'}
+{'loss': '2.551', 'grad_norm': '19.62', 'learning_rate': '9.357e-06', 'num_tokens': '2.348e+05', 'mean_token_accuracy': '0.502', 'epoch': '0.06443'}
+{'loss': '2.08', 'grad_norm': '28.75', 'learning_rate': '9.356e-06', 'num_tokens': '2.349e+05', 'mean_token_accuracy': '0.6', 'epoch': '0.06461'}
+{'loss': '1.272', 'grad_norm': '9.938', 'learning_rate': '9.354e-06', 'num_tokens': '2.36e+05', 'mean_token_accuracy': '0.6871', 'epoch': '0.06478'}
+{'loss': '1.312', 'grad_norm': '13.56', 'learning_rate': '9.352e-06', 'num_tokens': '2.365e+05', 'mean_token_accuracy': '0.645', 'epoch': '0.06495'}
+{'loss': '2.063', 'grad_norm': '22.5', 'learning_rate': '9.35e-06', 'num_tokens': '2.368e+05', 'mean_token_accuracy': '0.5533', 'epoch': '0.06513'}
+{'loss': '1.327', 'grad_norm': '18.12', 'learning_rate': '9.349e-06', 'num_tokens': '2.37e+05', 'mean_token_accuracy': '0.7256', 'epoch': '0.0653'}
+{'loss': '1.602', 'grad_norm': '14.62', 'learning_rate': '9.347e-06', 'num_tokens': '2.376e+05', 'mean_token_accuracy': '0.5724', 'epoch': '0.06547'}
+{'loss': '1.347', 'grad_norm': '13.5', 'learning_rate': '9.345e-06', 'num_tokens': '2.382e+05', 'mean_token_accuracy': '0.6465', 'epoch': '0.06565'}
+{'loss': '0.7195', 'grad_norm': '12.25', 'learning_rate': '9.344e-06', 'num_tokens': '2.389e+05', 'mean_token_accuracy': '0.8144', 'epoch': '0.06582'}
+{'loss': '2.573', 'grad_norm': '18.5', 'learning_rate': '9.342e-06', 'num_tokens': '2.395e+05', 'mean_token_accuracy': '0.5297', 'epoch': '0.066'}
+{'loss': '0.8162', 'grad_norm': '8.5', 'learning_rate': '9.34e-06', 'num_tokens': '2.408e+05', 'mean_token_accuracy': '0.7599', 'epoch': '0.06617'}
+{'loss': '2.444', 'grad_norm': '12.44', 'learning_rate': '9.338e-06', 'num_tokens': '2.414e+05', 'mean_token_accuracy': '0.5726', 'epoch': '0.06634'}
+
  7%|▋         | 383/5758 [02:08<31:18,  2.86it/s]
  7%|▋         | 384/5758 [02:09<31:50,  2.81it/s]
                                                  

  7%|▋         | 384/5758 [02:09<31:50,  2.81it/s]
  7%|▋         | 385/5758 [02:09<31:41,  2.83it/s]
                                                  

  7%|▋         | 385/5758 [02:09<31:41,  2.83it/s]
  7%|▋         | 386/5758 [02:09<30:23,  2.95it/s]
                                                  

  7%|▋         | 386/5758 [02:10<30:23,  2.95it/s]
  7%|▋         | 387/5758 [02:10<29:18,  3.05it/s]
                                                  

  7%|▋         | 387/5758 [02:10<29:18,  3.05it/s]
  7%|▋         | 388/5758 [02:10<30:44,  2.91it/s]
                                                  

  7%|▋         | 388/5758 [02:10<30:44,  2.91it/s]
  7%|▋         | 389/5758 [02:10<29:34,  3.03it/s]
                                                  

  7%|▋         | 389/5758 [02:10<29:34,  3.03it/s]
  7%|▋         | 390/5758 [02:11<28:46,  3.11it/s]
                                                  

  7%|▋         | 390/5758 [02:11<28:46,  3.11it/s]
  7%|▋         | 391/5758 [02:11<28:44,  3.11it/s]
                                                  

  7%|▋         | 391/5758 [02:11<28:44,  3.11it/s]
  7%|▋         | 392/5758 [02:11<28:25,  3.15it/s]
                                                  

  7%|▋         | 392/5758 [02:11<28:25,  3.15it/s]
  7%|▋         | 393/5758 [02:12<29:16,  3.06it/s]
                                                  

  7%|▋         | 393/5758 [02:12<29:16,  3.06it/s]
  7%|▋         | 394/5758 [02:12<29:35,  3.02it/s]
                                                  

  7%|▋         | 394/5758 [02:12<29:35,  3.02it/s]
  7%|▋         | 395/5758 [02:12<28:36,  3.12it/s]
                                                  

  7%|▋         | 395/5758 [02:12<28:36,  3.12it/s]
  7%|▋         | 396/5758 [02:13<28:44,  3.11it/s]
                                                  

  7%|▋         | 396/5758 [02:13<28:44,  3.11it/s]
  7%|▋         | 397/5758 [02:13<28:53,  3.09it/s]
                                                  

  7%|▋         | 397/5758 [02:13<28:53,  3.09it/s]
  7%|▋         | 398/5758 [02:13<30:26,  2.93it/s]
                                                  

  7%|▋         | 398/5758 [02:13<30:26,  2.93it/s]
  7%|▋         | 399/5758 [02:14<29:34,  3.02it/s]
                                                  

  7%|▋         | 399/5758 [02:14<29:34,  3.02it/s]
  7%|▋         | 400/5758 [02:14<28:33,  3.13it/s]
                                                  

  7%|▋         | 400/5758 [02:14<28:33,  3.13it/s]
  7%|▋         | 401/5758 [02:14<30:09,  2.96it/s]
                                                  

  7%|▋         | 401/5758 [02:14<30:09,  2.96it/s]
  7%|▋         | 402/5758 [02:15<29:01,  3.08it/s]
                                                  

  7%|▋         | 402/5758 [02:15<29:01,  3.08it/s]
  7%|▋         | 403/5758 [02:15<28:36,  3.12it/s]
                                                  

  7%|▋         | 403/5758 [02:15<28:36,  3.12it/s]
  7%|▋         | 404/5758 [02:15<30:04,  2.97it/s]
                                                  

  7%|▋         | 404/5758 [02:15<30:04,  2.97it/s]
  7%|▋         | 405/5758 [02:16<31:21,  2.84it/s]
                                                  

  7%|▋         | 405/5758 [02:16<31:21,  2.84it/s]
  7%|▋         | 406/5758 [02:16<31:09,  2.86it/s]
                                                  

  7%|▋         | 406/5758 [02:16<31:09,  2.86it/s]
  7%|▋         | 407/5758 [02:17<32:13,  2.77it/s]
                                                  

  7%|▋         | 407/5758 [02:17<32:13,  2.77it/s]
  7%|▋         | 408/5758 [02:17<32:36,  2.73it/s]
                                                  

  7%|▋         | 408/5758 [02:17<32:36,  2.73it/s]
  7%|▋         | 409/5758 [02:17<35:33,  2.51it/s]
                                                  

  7%|▋         | 409/5758 [02:17<35:33,  2.51it/s]
  7%|▋         | 410/5758 [02:18<33:14,  2.68it/s]
                                                  

  7%|▋         | 410/5758 [02:18<33:14,  2.68it/s]
  7%|▋         | 411/5758 [02:18<32:41,  2.73it/s]
                                                  

  7%|▋         | 411/5758 [02:18<32:41,  2.73it/s]
  7%|▋         | 412/5758 [02:18<31:24,  2.84it/s]
                                                  

  7%|▋         | 412/5758 [02:18<31:24,  2.84it/s]
  7%|▋         | 413/5758 [02:19<32:17,  2.76it/s]
                                                  

  7%|▋         | 413/5758 [02:19<32:17,  2.76it/s]
  7%|▋         | 414/5758 [02:19<30:24,  2.93it/s]
                                                  

  7%|▋         | 414/5758 [02:19<30:24,  2.93it/s]
  7%|▋         | 415/5758 [02:19<31:27,  2.83it/s]
                                                  

  7%|▋         | 415/5758 [02:19<31:27,  2.83it/s]
  7%|▋         | 416/5758 [02:20<29:49,  2.98it/s]
                                                  

  7%|▋         | 416/5758 [02:20<29:49,  2.98it/s]
  7%|▋         | 417/5758 [02:20<28:59,  3.07it/s]
                                                  

  7%|▋         | 417/5758 [02:20<28:59,  3.07it/s]
  7%|▋         | 418/5758 [02:20<28:32,  3.12it/s]
                                                  

  7%|▋         | 418/5758 [02:20<28:32,  3.12it/s]
  7%|▋         | 419/5758 [02:21<30:14,  2.94it/s]
                                                  

  7%|▋         | 419/5758 [02:21<30:14,  2.94it/s]
  7%|▋         | 420/5758 [02:21<30:03,  2.96it/s]
                                                  

  7%|▋         | 420/5758 [02:21<30:03,  2.96it/s]
  7%|▋         | 421/5758 [02:21<30:16,  2.94it/s]
                                                  

  7%|▋         | 421/5758 [02:21<30:16,  2.94it/s]
  7%|▋         | 422/5758 [02:22<29:11,  3.05it/s]
                                                  

  7%|▋         | 422/5758 [02:22<29:11,  3.05it/s]
  7%|▋         | 423/5758 [02:22<30:19,  2.93it/s]
                                                  

  7%|▋         | 423/5758 [02:22<30:19,  2.93it/s]
  7%|▋         | 424/5758 [02:22<29:56,  2.97it/s]
                                                  

  7%|▋         | 424/5758 [02:22<29:56,  2.97it/s]
  7%|▋         | 425/5758 [02:23<29:01,  3.06it/s]
                                                  

  7%|▋         | 425/5758 [02:23<29:01,  3.06it/s]
  7%|▋         | 426/5758 [02:23<29:03,  3.06it/s]
                                                  

  7%|▋         | 426/5758 [02:23<29:03,  3.06it/s]
  7%|▋         | 427/5758 [02:23<29:31,  3.01it/s]
                                                  

  7%|▋         | 427/5758 [02:23<29:31,  3.01it/s]
  7%|▋         | 428/5758 [02:24<30:25,  2.92it/s]
                                                  

  7%|▋         | 428/5758 [02:24<30:25,  2.92it/s]
  7%|▋         | 429/5758 [02:24<28:56,  3.07it/s]
                                                  

  7%|▋         | 429/5758 [02:24<28:56,  3.07it/s]
  7%|▋         | 430/5758 [02:24<29:18,  3.03it/s]
                                                  

  7%|▋         | 430/5758 [02:24<29:18,  3.03it/s]
  7%|▋         | 431/5758 [02:25<30:15,  2.93it/s]
                                                  

  7%|▋         | 431/5758 [02:25<30:15,  2.93it/s]
  8%|▊         | 432/5758 [02:25<29:22,  3.02it/s]
                                                  

  8%|▊         | 432/5758 [02:25<29:22,  3.02it/s]
  8%|▊         | 433/5758 [02:25<30:29,  2.91it/s]
                                                  

  8%|▊         | 433/5758 [02:25<30:29,  2.91it/s]
  8%|▊         | 434/5758 [02:26<31:04,  2.86it/s]
                                                  

  8%|▊         | 434/5758 [02:26<31:04,  2.86it/s]
  8%|▊         | 435/5758 [02:26<29:26,  3.01it/s]
                                                  

  8%|▊         | 435/5758 [02:26<29:26,  3.01it/s]
  8%|▊         | 436/5758 [02:26<28:39,  3.10it/s]
                                                  

  8%|▊         | 436/5758 [02:26<28:39,  3.10it/s]
  8%|▊         | 437/5758 [02:27<27:50,  3.19it/s]
                                                  
{'loss': '1.386', 'grad_norm': '11', 'learning_rate': '9.337e-06', 'num_tokens': '2.422e+05', 'mean_token_accuracy': '0.6411', 'epoch': '0.06652'}
+{'loss': '2.047', 'grad_norm': '18.88', 'learning_rate': '9.335e-06', 'num_tokens': '2.425e+05', 'mean_token_accuracy': '0.5889', 'epoch': '0.06669'}
+{'loss': '2.133', 'grad_norm': '7.594', 'learning_rate': '9.333e-06', 'num_tokens': '2.444e+05', 'mean_token_accuracy': '0.5598', 'epoch': '0.06686'}
+{'loss': '1.394', 'grad_norm': '11.81', 'learning_rate': '9.331e-06', 'num_tokens': '2.451e+05', 'mean_token_accuracy': '0.6081', 'epoch': '0.06704'}
+{'loss': '1.404', 'grad_norm': '12.12', 'learning_rate': '9.33e-06', 'num_tokens': '2.457e+05', 'mean_token_accuracy': '0.6661', 'epoch': '0.06721'}
+{'loss': '1.555', 'grad_norm': '13.94', 'learning_rate': '9.328e-06', 'num_tokens': '2.463e+05', 'mean_token_accuracy': '0.6028', 'epoch': '0.06738'}
+{'loss': '1.517', 'grad_norm': '12.75', 'learning_rate': '9.326e-06', 'num_tokens': '2.469e+05', 'mean_token_accuracy': '0.6056', 'epoch': '0.06756'}
+{'loss': '0.759', 'grad_norm': '7.875', 'learning_rate': '9.324e-06', 'num_tokens': '2.483e+05', 'mean_token_accuracy': '0.7768', 'epoch': '0.06773'}
+{'loss': '2.424', 'grad_norm': '9.875', 'learning_rate': '9.323e-06', 'num_tokens': '2.498e+05', 'mean_token_accuracy': '0.5526', 'epoch': '0.06791'}
+{'loss': '1.435', 'grad_norm': '11.56', 'learning_rate': '9.321e-06', 'num_tokens': '2.505e+05', 'mean_token_accuracy': '0.6094', 'epoch': '0.06808'}
+{'loss': '1.544', 'grad_norm': '11.25', 'learning_rate': '9.319e-06', 'num_tokens': '2.515e+05', 'mean_token_accuracy': '0.5928', 'epoch': '0.06825'}
+{'loss': '1.584', 'grad_norm': '14.31', 'learning_rate': '9.317e-06', 'num_tokens': '2.519e+05', 'mean_token_accuracy': '0.6508', 'epoch': '0.06843'}
+{'loss': '2.19', 'grad_norm': '8.188', 'learning_rate': '9.316e-06', 'num_tokens': '2.532e+05', 'mean_token_accuracy': '0.5476', 'epoch': '0.0686'}
+{'loss': '1.955', 'grad_norm': '18.75', 'learning_rate': '9.314e-06', 'num_tokens': '2.536e+05', 'mean_token_accuracy': '0.6325', 'epoch': '0.06877'}
+{'loss': '1.165', 'grad_norm': '27.5', 'learning_rate': '9.312e-06', 'num_tokens': '2.537e+05', 'mean_token_accuracy': '0.7097', 'epoch': '0.06895'}
+{'loss': '1.155', 'grad_norm': '13.38', 'learning_rate': '9.311e-06', 'num_tokens': '2.542e+05', 'mean_token_accuracy': '0.7127', 'epoch': '0.06912'}
+{'loss': '1.455', 'grad_norm': '13.12', 'learning_rate': '9.309e-06', 'num_tokens': '2.549e+05', 'mean_token_accuracy': '0.6138', 'epoch': '0.06929'}
+{'loss': '2.077', 'grad_norm': '19.62', 'learning_rate': '9.307e-06', 'num_tokens': '2.553e+05', 'mean_token_accuracy': '0.5802', 'epoch': '0.06947'}
+{'loss': '1.475', 'grad_norm': '13.5', 'learning_rate': '9.305e-06', 'num_tokens': '2.558e+05', 'mean_token_accuracy': '0.6162', 'epoch': '0.06964'}
+{'loss': '1.351', 'grad_norm': '22.5', 'learning_rate': '9.304e-06', 'num_tokens': '2.56e+05', 'mean_token_accuracy': '0.7055', 'epoch': '0.06982'}
+{'loss': '1.207', 'grad_norm': '20.25', 'learning_rate': '9.302e-06', 'num_tokens': '2.562e+05', 'mean_token_accuracy': '0.7198', 'epoch': '0.06999'}
+{'loss': '1.364', 'grad_norm': '7.312', 'learning_rate': '9.3e-06', 'num_tokens': '2.584e+05', 'mean_token_accuracy': '0.657', 'epoch': '0.07016'}
+{'loss': '1.298', 'grad_norm': '9.5', 'learning_rate': '9.298e-06', 'num_tokens': '2.595e+05', 'mean_token_accuracy': '0.7069', 'epoch': '0.07034'}
+{'loss': '1.97', 'grad_norm': '16.38', 'learning_rate': '9.297e-06', 'num_tokens': '2.6e+05', 'mean_token_accuracy': '0.5461', 'epoch': '0.07051'}
+{'loss': '1.449', 'grad_norm': '12.94', 'learning_rate': '9.295e-06', 'num_tokens': '2.606e+05', 'mean_token_accuracy': '0.6081', 'epoch': '0.07068'}
+{'loss': '1.984', 'grad_norm': '9.812', 'learning_rate': '9.293e-06', 'num_tokens': '2.617e+05', 'mean_token_accuracy': '0.6275', 'epoch': '0.07086'}
+{'loss': '2.063', 'grad_norm': '30.12', 'learning_rate': '9.291e-06', 'num_tokens': '2.618e+05', 'mean_token_accuracy': '0.5575', 'epoch': '0.07103'}
+{'loss': '1.471', 'grad_norm': '12.62', 'learning_rate': '9.29e-06', 'num_tokens': '2.626e+05', 'mean_token_accuracy': '0.591', 'epoch': '0.07121'}
+{'loss': '1.587', 'grad_norm': '10.38', 'learning_rate': '9.288e-06', 'num_tokens': '2.635e+05', 'mean_token_accuracy': '0.6663', 'epoch': '0.07138'}
+{'loss': '1.644', 'grad_norm': '21.62', 'learning_rate': '9.286e-06', 'num_tokens': '2.637e+05', 'mean_token_accuracy': '0.6911', 'epoch': '0.07155'}
+{'loss': '1.924', 'grad_norm': '33', 'learning_rate': '9.284e-06', 'num_tokens': '2.638e+05', 'mean_token_accuracy': '0.5496', 'epoch': '0.07173'}
+{'loss': '1.891', 'grad_norm': '24.5', 'learning_rate': '9.283e-06', 'num_tokens': '2.64e+05', 'mean_token_accuracy': '0.5714', 'epoch': '0.0719'}
+{'loss': '2.637', 'grad_norm': '18.25', 'learning_rate': '9.281e-06', 'num_tokens': '2.646e+05', 'mean_token_accuracy': '0.5096', 'epoch': '0.07207'}
+{'loss': '1.494', 'grad_norm': '14.5', 'learning_rate': '9.279e-06', 'num_tokens': '2.652e+05', 'mean_token_accuracy': '0.6315', 'epoch': '0.07225'}
+{'loss': '1.304', 'grad_norm': '13.06', 'learning_rate': '9.278e-06', 'num_tokens': '2.659e+05', 'mean_token_accuracy': '0.6491', 'epoch': '0.07242'}
+{'loss': '1.292', 'grad_norm': '35.75', 'learning_rate': '9.276e-06', 'num_tokens': '2.66e+05', 'mean_token_accuracy': '0.7439', 'epoch': '0.07259'}
+{'loss': '1.422', 'grad_norm': '5.938', 'learning_rate': '9.274e-06', 'num_tokens': '2.685e+05', 'mean_token_accuracy': '0.6703', 'epoch': '0.07277'}
+{'loss': '1.264', 'grad_norm': '29.88', 'learning_rate': '9.272e-06', 'num_tokens': '2.686e+05', 'mean_token_accuracy': '0.7396', 'epoch': '0.07294'}
+{'loss': '2.525', 'grad_norm': '18.25', 'learning_rate': '9.271e-06', 'num_tokens': '2.691e+05', 'mean_token_accuracy': '0.4967', 'epoch': '0.07312'}
+{'loss': '1.557', 'grad_norm': '12.69', 'learning_rate': '9.269e-06', 'num_tokens': '2.697e+05', 'mean_token_accuracy': '0.6575', 'epoch': '0.07329'}
+{'loss': '2.23', 'grad_norm': '9.562', 'learning_rate': '9.267e-06', 'num_tokens': '2.708e+05', 'mean_token_accuracy': '0.5633', 'epoch': '0.07346'}
+{'loss': '1.246', 'grad_norm': '11.75', 'learning_rate': '9.265e-06', 'num_tokens': '2.717e+05', 'mean_token_accuracy': '0.6612', 'epoch': '0.07364'}
+{'loss': '0.9376', 'grad_norm': '10.44', 'learning_rate': '9.264e-06', 'num_tokens': '2.731e+05', 'mean_token_accuracy': '0.7342', 'epoch': '0.07381'}
+{'loss': '2.016', 'grad_norm': '11.38', 'learning_rate': '9.262e-06', 'num_tokens': '2.739e+05', 'mean_token_accuracy': '0.5884', 'epoch': '0.07398'}
+{'loss': '1.286', 'grad_norm': '31.12', 'learning_rate': '9.26e-06', 'num_tokens': '2.74e+05', 'mean_token_accuracy': '0.6893', 'epoch': '0.07416'}
+{'loss': '1.54', 'grad_norm': '31.75', 'learning_rate': '9.258e-06', 'num_tokens': '2.742e+05', 'mean_token_accuracy': '0.6529', 'epoch': '0.07433'}
+{'loss': '2.455', 'grad_norm': '24.12', 'learning_rate': '9.257e-06', 'num_tokens': '2.745e+05', 'mean_token_accuracy': '0.5601', 'epoch': '0.07451'}
+{'loss': '2.435', 'grad_norm': '14.25', 'learning_rate': '9.255e-06', 'num_tokens': '2.749e+05', 'mean_token_accuracy': '0.5361', 'epoch': '0.07468'}
+{'loss': '2.258', 'grad_norm': '9.438', 'learning_rate': '9.253e-06', 'num_tokens': '2.76e+05', 'mean_token_accuracy': '0.5317', 'epoch': '0.07485'}
+{'loss': '1.159', 'grad_norm': '15.56', 'learning_rate': '9.251e-06', 'num_tokens': '2.763e+05', 'mean_token_accuracy': '0.7332', 'epoch': '0.07503'}
+{'loss': '1.279', 'grad_norm': '13', 'learning_rate': '9.25e-06', 'num_tokens': '2.769e+05', 'mean_token_accuracy': '0.6757', 'epoch': '0.0752'}
+{'loss': '2.05', 'grad_norm': '11.69', 'learning_rate': '9.248e-06', 'num_tokens': '2.774e+05', 'mean_token_accuracy': '0.5896', 'epoch': '0.07537'}
+{'loss': '1.43', 'grad_norm': '13.56', 'learning_rate': '9.246e-06', 'num_tokens': '2.781e+05', 'mean_token_accuracy': '0.6424', 'epoch': '0.07555'}
+{'loss': '2.349', 'grad_norm': '18.12', 'learning_rate': '9.245e-06', 'num_tokens': '2.784e+05', 'mean_token_accuracy': '0.536', 'epoch': '0.07572'}
+
  8%|▊         | 437/5758 [02:27<27:50,  3.19it/s]
  8%|▊         | 438/5758 [02:27<30:36,  2.90it/s]
                                                  

  8%|▊         | 438/5758 [02:27<30:36,  2.90it/s]
  8%|▊         | 439/5758 [02:27<31:17,  2.83it/s]
                                                  

  8%|▊         | 439/5758 [02:27<31:17,  2.83it/s]
  8%|▊         | 440/5758 [02:28<31:00,  2.86it/s]
                                                  

  8%|▊         | 440/5758 [02:28<31:00,  2.86it/s]
  8%|▊         | 441/5758 [02:28<30:35,  2.90it/s]
                                                  

  8%|▊         | 441/5758 [02:28<30:35,  2.90it/s]
  8%|▊         | 442/5758 [02:28<30:30,  2.90it/s]
                                                  

  8%|▊         | 442/5758 [02:28<30:30,  2.90it/s]
  8%|▊         | 443/5758 [02:29<30:49,  2.87it/s]
                                                  

  8%|▊         | 443/5758 [02:29<30:49,  2.87it/s]
  8%|▊         | 444/5758 [02:29<30:16,  2.93it/s]
                                                  

  8%|▊         | 444/5758 [02:29<30:16,  2.93it/s]
  8%|▊         | 445/5758 [02:29<29:05,  3.04it/s]
                                                  

  8%|▊         | 445/5758 [02:29<29:05,  3.04it/s]
  8%|▊         | 446/5758 [02:30<30:10,  2.93it/s]
                                                  

  8%|▊         | 446/5758 [02:30<30:10,  2.93it/s]
  8%|▊         | 447/5758 [02:30<31:03,  2.85it/s]
                                                  

  8%|▊         | 447/5758 [02:30<31:03,  2.85it/s]
  8%|▊         | 448/5758 [02:30<30:04,  2.94it/s]
                                                  

  8%|▊         | 448/5758 [02:30<30:04,  2.94it/s]
  8%|▊         | 449/5758 [02:31<28:51,  3.07it/s]
                                                  

  8%|▊         | 449/5758 [02:31<28:51,  3.07it/s]
  8%|▊         | 450/5758 [02:31<28:22,  3.12it/s]
                                                  

  8%|▊         | 450/5758 [02:31<28:22,  3.12it/s]
  8%|▊         | 451/5758 [02:31<28:40,  3.08it/s]
                                                  

  8%|▊         | 451/5758 [02:31<28:40,  3.08it/s]
  8%|▊         | 452/5758 [02:32<27:58,  3.16it/s]
                                                  

  8%|▊         | 452/5758 [02:32<27:58,  3.16it/s]
  8%|▊         | 453/5758 [02:32<28:17,  3.13it/s]
                                                  

  8%|▊         | 453/5758 [02:32<28:17,  3.13it/s]
  8%|▊         | 454/5758 [02:32<29:44,  2.97it/s]
                                                  

  8%|▊         | 454/5758 [02:32<29:44,  2.97it/s]
  8%|▊         | 455/5758 [02:33<31:00,  2.85it/s]
                                                  

  8%|▊         | 455/5758 [02:33<31:00,  2.85it/s]
  8%|▊         | 456/5758 [02:33<30:36,  2.89it/s]
                                                  

  8%|▊         | 456/5758 [02:33<30:36,  2.89it/s]
  8%|▊         | 457/5758 [02:33<30:01,  2.94it/s]
                                                  

  8%|▊         | 457/5758 [02:33<30:01,  2.94it/s]
  8%|▊         | 458/5758 [02:34<29:11,  3.03it/s]
                                                  

  8%|▊         | 458/5758 [02:34<29:11,  3.03it/s]
  8%|▊         | 459/5758 [02:34<28:44,  3.07it/s]
                                                  

  8%|▊         | 459/5758 [02:34<28:44,  3.07it/s]
  8%|▊         | 460/5758 [02:34<27:54,  3.16it/s]
                                                  

  8%|▊         | 460/5758 [02:34<27:54,  3.16it/s]
  8%|▊         | 461/5758 [02:35<28:18,  3.12it/s]
                                                  

  8%|▊         | 461/5758 [02:35<28:18,  3.12it/s]
  8%|▊         | 462/5758 [02:35<27:47,  3.18it/s]
                                                  

  8%|▊         | 462/5758 [02:35<27:47,  3.18it/s]
  8%|▊         | 463/5758 [02:35<30:35,  2.88it/s]
                                                  

  8%|▊         | 463/5758 [02:35<30:35,  2.88it/s]
  8%|▊         | 464/5758 [02:36<32:25,  2.72it/s]
                                                  

  8%|▊         | 464/5758 [02:36<32:25,  2.72it/s]
  8%|▊         | 465/5758 [02:36<32:33,  2.71it/s]
                                                  

  8%|▊         | 465/5758 [02:36<32:33,  2.71it/s]
  8%|▊         | 466/5758 [02:37<30:52,  2.86it/s]
                                                  

  8%|▊         | 466/5758 [02:37<30:52,  2.86it/s]
  8%|▊         | 467/5758 [02:37<29:49,  2.96it/s]
                                                  

  8%|▊         | 467/5758 [02:37<29:49,  2.96it/s]
  8%|▊         | 468/5758 [02:37<29:34,  2.98it/s]
                                                  

  8%|▊         | 468/5758 [02:37<29:34,  2.98it/s]
  8%|▊         | 469/5758 [02:37<28:49,  3.06it/s]
                                                  

  8%|▊         | 469/5758 [02:37<28:49,  3.06it/s]
  8%|▊         | 470/5758 [02:38<29:09,  3.02it/s]
                                                  

  8%|▊         | 470/5758 [02:38<29:09,  3.02it/s]
  8%|▊         | 471/5758 [02:38<29:24,  3.00it/s]
                                                  

  8%|▊         | 471/5758 [02:38<29:24,  3.00it/s]
  8%|▊         | 472/5758 [02:38<28:46,  3.06it/s]
                                                  

  8%|▊         | 472/5758 [02:38<28:46,  3.06it/s]
  8%|▊         | 473/5758 [02:39<28:06,  3.13it/s]
                                                  

  8%|▊         | 473/5758 [02:39<28:06,  3.13it/s]
  8%|▊         | 474/5758 [02:39<28:31,  3.09it/s]
                                                  

  8%|▊         | 474/5758 [02:39<28:31,  3.09it/s]
  8%|▊         | 475/5758 [02:39<28:14,  3.12it/s]
                                                  

  8%|▊         | 475/5758 [02:39<28:14,  3.12it/s]
  8%|▊         | 476/5758 [02:40<27:59,  3.14it/s]
                                                  

  8%|▊         | 476/5758 [02:40<27:59,  3.14it/s]
  8%|▊         | 477/5758 [02:40<28:25,  3.10it/s]
                                                  

  8%|▊         | 477/5758 [02:40<28:25,  3.10it/s]
  8%|▊         | 478/5758 [02:40<29:22,  3.00it/s]
                                                  

  8%|▊         | 478/5758 [02:40<29:22,  3.00it/s]
  8%|▊         | 479/5758 [02:41<29:18,  3.00it/s]
                                                  

  8%|▊         | 479/5758 [02:41<29:18,  3.00it/s]
  8%|▊         | 480/5758 [02:41<29:17,  3.00it/s]
                                                  

  8%|▊         | 480/5758 [02:41<29:17,  3.00it/s]
  8%|▊         | 481/5758 [02:41<29:16,  3.00it/s]
                                                  

  8%|▊         | 481/5758 [02:41<29:16,  3.00it/s]
  8%|▊         | 482/5758 [02:42<29:19,  3.00it/s]
                                                  

  8%|▊         | 482/5758 [02:42<29:19,  3.00it/s]
  8%|▊         | 483/5758 [02:42<28:39,  3.07it/s]
                                                  

  8%|▊         | 483/5758 [02:42<28:39,  3.07it/s]
  8%|▊         | 484/5758 [02:42<27:58,  3.14it/s]
                                                  

  8%|▊         | 484/5758 [02:42<27:58,  3.14it/s]
  8%|▊         | 485/5758 [02:43<28:57,  3.04it/s]
                                                  

  8%|▊         | 485/5758 [02:43<28:57,  3.04it/s]
  8%|▊         | 486/5758 [02:43<28:22,  3.10it/s]
                                                  

  8%|▊         | 486/5758 [02:43<28:22,  3.10it/s]
  8%|▊         | 487/5758 [02:43<27:41,  3.17it/s]
                                                  

  8%|▊         | 487/5758 [02:43<27:41,  3.17it/s]
  8%|▊         | 488/5758 [02:44<30:03,  2.92it/s]
                                                  

  8%|▊         | 488/5758 [02:44<30:03,  2.92it/s]
  8%|▊         | 489/5758 [02:44<32:38,  2.69it/s]
                                                  

  8%|▊         | 489/5758 [02:44<32:38,  2.69it/s]
  9%|▊         | 490/5758 [02:44<31:16,  2.81it/s]
                                                  

  9%|▊         | 490/5758 [02:44<31:16,  2.81it/s]
  9%|▊         | 491/5758 [02:45<30:15,  2.90it/s]
                                                  
{'loss': '1.397', 'grad_norm': '11.44', 'learning_rate': '9.243e-06', 'num_tokens': '2.79e+05', 'mean_token_accuracy': '0.6894', 'epoch': '0.07589'}
+{'loss': '1.382', 'grad_norm': '7.812', 'learning_rate': '9.241e-06', 'num_tokens': '2.804e+05', 'mean_token_accuracy': '0.7289', 'epoch': '0.07607'}
+{'loss': '2.161', 'grad_norm': '18', 'learning_rate': '9.239e-06', 'num_tokens': '2.809e+05', 'mean_token_accuracy': '0.5604', 'epoch': '0.07624'}
+{'loss': '1.584', 'grad_norm': '13.5', 'learning_rate': '9.238e-06', 'num_tokens': '2.813e+05', 'mean_token_accuracy': '0.6636', 'epoch': '0.07642'}
+{'loss': '2.394', 'grad_norm': '14.06', 'learning_rate': '9.236e-06', 'num_tokens': '2.822e+05', 'mean_token_accuracy': '0.5213', 'epoch': '0.07659'}
+{'loss': '1.418', 'grad_norm': '10.94', 'learning_rate': '9.234e-06', 'num_tokens': '2.83e+05', 'mean_token_accuracy': '0.6094', 'epoch': '0.07676'}
+{'loss': '1.459', 'grad_norm': '21.5', 'learning_rate': '9.232e-06', 'num_tokens': '2.833e+05', 'mean_token_accuracy': '0.6763', 'epoch': '0.07694'}
+{'loss': '0.8779', 'grad_norm': '17.25', 'learning_rate': '9.231e-06', 'num_tokens': '2.835e+05', 'mean_token_accuracy': '0.809', 'epoch': '0.07711'}
+{'loss': '1.187', 'grad_norm': '12', 'learning_rate': '9.229e-06', 'num_tokens': '2.842e+05', 'mean_token_accuracy': '0.6715', 'epoch': '0.07728'}
+{'loss': '1.531', 'grad_norm': '19.25', 'learning_rate': '9.227e-06', 'num_tokens': '2.844e+05', 'mean_token_accuracy': '0.7052', 'epoch': '0.07746'}
+{'loss': '0.9798', 'grad_norm': '18.38', 'learning_rate': '9.225e-06', 'num_tokens': '2.847e+05', 'mean_token_accuracy': '0.7536', 'epoch': '0.07763'}
+{'loss': '0.6351', 'grad_norm': '7.125', 'learning_rate': '9.224e-06', 'num_tokens': '2.862e+05', 'mean_token_accuracy': '0.7961', 'epoch': '0.0778'}
+{'loss': '1.312', 'grad_norm': '36', 'learning_rate': '9.222e-06', 'num_tokens': '2.863e+05', 'mean_token_accuracy': '0.7286', 'epoch': '0.07798'}
+{'loss': '1.261', 'grad_norm': '12.25', 'learning_rate': '9.22e-06', 'num_tokens': '2.87e+05', 'mean_token_accuracy': '0.6502', 'epoch': '0.07815'}
+{'loss': '1.767', 'grad_norm': '21.75', 'learning_rate': '9.218e-06', 'num_tokens': '2.872e+05', 'mean_token_accuracy': '0.6465', 'epoch': '0.07833'}
+{'loss': '2.203', 'grad_norm': '18.12', 'learning_rate': '9.217e-06', 'num_tokens': '2.879e+05', 'mean_token_accuracy': '0.568', 'epoch': '0.0785'}
+{'loss': '1.218', 'grad_norm': '10.94', 'learning_rate': '9.215e-06', 'num_tokens': '2.887e+05', 'mean_token_accuracy': '0.6783', 'epoch': '0.07867'}
+{'loss': '1.611', 'grad_norm': '10.56', 'learning_rate': '9.213e-06', 'num_tokens': '2.898e+05', 'mean_token_accuracy': '0.5989', 'epoch': '0.07885'}
+{'loss': '2.254', 'grad_norm': '17.12', 'learning_rate': '9.212e-06', 'num_tokens': '2.904e+05', 'mean_token_accuracy': '0.5664', 'epoch': '0.07902'}
+{'loss': '2.008', 'grad_norm': '37', 'learning_rate': '9.21e-06', 'num_tokens': '2.905e+05', 'mean_token_accuracy': '0.61', 'epoch': '0.07919'}
+{'loss': '1.917', 'grad_norm': '18.62', 'learning_rate': '9.208e-06', 'num_tokens': '2.909e+05', 'mean_token_accuracy': '0.6123', 'epoch': '0.07937'}
+{'loss': '1.082', 'grad_norm': '11.75', 'learning_rate': '9.206e-06', 'num_tokens': '2.916e+05', 'mean_token_accuracy': '0.6789', 'epoch': '0.07954'}
+{'loss': '1.237', 'grad_norm': '12.31', 'learning_rate': '9.205e-06', 'num_tokens': '2.924e+05', 'mean_token_accuracy': '0.6399', 'epoch': '0.07972'}
+{'loss': '1.621', 'grad_norm': '31.38', 'learning_rate': '9.203e-06', 'num_tokens': '2.925e+05', 'mean_token_accuracy': '0.6536', 'epoch': '0.07989'}
+{'loss': '1.991', 'grad_norm': '11.94', 'learning_rate': '9.201e-06', 'num_tokens': '2.934e+05', 'mean_token_accuracy': '0.6079', 'epoch': '0.08006'}
+{'loss': '2.455', 'grad_norm': '11.44', 'learning_rate': '9.199e-06', 'num_tokens': '2.941e+05', 'mean_token_accuracy': '0.5023', 'epoch': '0.08024'}
+{'loss': '1.114', 'grad_norm': '11.44', 'learning_rate': '9.198e-06', 'num_tokens': '2.947e+05', 'mean_token_accuracy': '0.7251', 'epoch': '0.08041'}
+{'loss': '1.34', 'grad_norm': '28.5', 'learning_rate': '9.196e-06', 'num_tokens': '2.949e+05', 'mean_token_accuracy': '0.6986', 'epoch': '0.08058'}
+{'loss': '1.345', 'grad_norm': '14.38', 'learning_rate': '9.194e-06', 'num_tokens': '2.954e+05', 'mean_token_accuracy': '0.6485', 'epoch': '0.08076'}
+{'loss': '2.122', 'grad_norm': '8.875', 'learning_rate': '9.192e-06', 'num_tokens': '2.968e+05', 'mean_token_accuracy': '0.5743', 'epoch': '0.08093'}
+{'loss': '0.7271', 'grad_norm': '8.062', 'learning_rate': '9.191e-06', 'num_tokens': '2.982e+05', 'mean_token_accuracy': '0.7812', 'epoch': '0.0811'}
+{'loss': '1.765', 'grad_norm': '35', 'learning_rate': '9.189e-06', 'num_tokens': '2.983e+05', 'mean_token_accuracy': '0.6196', 'epoch': '0.08128'}
+{'loss': '0.9349', 'grad_norm': '11.19', 'learning_rate': '9.187e-06', 'num_tokens': '2.99e+05', 'mean_token_accuracy': '0.7122', 'epoch': '0.08145'}
+{'loss': '1.994', 'grad_norm': '14.31', 'learning_rate': '9.185e-06', 'num_tokens': '2.994e+05', 'mean_token_accuracy': '0.5917', 'epoch': '0.08163'}
+{'loss': '1.277', 'grad_norm': '11.62', 'learning_rate': '9.184e-06', 'num_tokens': '3.003e+05', 'mean_token_accuracy': '0.6537', 'epoch': '0.0818'}
+{'loss': '2.242', 'grad_norm': '20.25', 'learning_rate': '9.182e-06', 'num_tokens': '3.007e+05', 'mean_token_accuracy': '0.569', 'epoch': '0.08197'}
+{'loss': '1.809', 'grad_norm': '9.625', 'learning_rate': '9.18e-06', 'num_tokens': '3.02e+05', 'mean_token_accuracy': '0.6089', 'epoch': '0.08215'}
+{'loss': '1.739', 'grad_norm': '8.25', 'learning_rate': '9.179e-06', 'num_tokens': '3.037e+05', 'mean_token_accuracy': '0.6336', 'epoch': '0.08232'}
+{'loss': '1.159', 'grad_norm': '11.69', 'learning_rate': '9.177e-06', 'num_tokens': '3.044e+05', 'mean_token_accuracy': '0.6708', 'epoch': '0.08249'}
+{'loss': '1.519', 'grad_norm': '12.31', 'learning_rate': '9.175e-06', 'num_tokens': '3.051e+05', 'mean_token_accuracy': '0.6176', 'epoch': '0.08267'}
+{'loss': '1.975', 'grad_norm': '8.125', 'learning_rate': '9.173e-06', 'num_tokens': '3.068e+05', 'mean_token_accuracy': '0.6001', 'epoch': '0.08284'}
+{'loss': '1.721', 'grad_norm': '19.62', 'learning_rate': '9.172e-06', 'num_tokens': '3.07e+05', 'mean_token_accuracy': '0.6787', 'epoch': '0.08301'}
+{'loss': '2.315', 'grad_norm': '14.38', 'learning_rate': '9.17e-06', 'num_tokens': '3.075e+05', 'mean_token_accuracy': '0.5145', 'epoch': '0.08319'}
+{'loss': '2.315', 'grad_norm': '11.5', 'learning_rate': '9.168e-06', 'num_tokens': '3.083e+05', 'mean_token_accuracy': '0.5369', 'epoch': '0.08336'}
+{'loss': '1.356', 'grad_norm': '10.88', 'learning_rate': '9.166e-06', 'num_tokens': '3.092e+05', 'mean_token_accuracy': '0.6423', 'epoch': '0.08354'}
+{'loss': '1.462', 'grad_norm': '21.25', 'learning_rate': '9.165e-06', 'num_tokens': '3.094e+05', 'mean_token_accuracy': '0.7097', 'epoch': '0.08371'}
+{'loss': '1.448', 'grad_norm': '12.88', 'learning_rate': '9.163e-06', 'num_tokens': '3.101e+05', 'mean_token_accuracy': '0.6197', 'epoch': '0.08388'}
+{'loss': '0.612', 'grad_norm': '37', 'learning_rate': '9.161e-06', 'num_tokens': '3.103e+05', 'mean_token_accuracy': '0.8704', 'epoch': '0.08406'}
+{'loss': '1.839', 'grad_norm': '43.75', 'learning_rate': '9.159e-06', 'num_tokens': '3.104e+05', 'mean_token_accuracy': '0.6903', 'epoch': '0.08423'}
+{'loss': '1.541', 'grad_norm': '8.438', 'learning_rate': '9.158e-06', 'num_tokens': '3.118e+05', 'mean_token_accuracy': '0.6434', 'epoch': '0.0844'}
+{'loss': '0.8791', 'grad_norm': '8.375', 'learning_rate': '9.156e-06', 'num_tokens': '3.131e+05', 'mean_token_accuracy': '0.7488', 'epoch': '0.08458'}
+{'loss': '0.8085', 'grad_norm': '7.438', 'learning_rate': '9.154e-06', 'num_tokens': '3.147e+05', 'mean_token_accuracy': '0.7571', 'epoch': '0.08475'}
+{'loss': '1.619', 'grad_norm': '14.38', 'learning_rate': '9.152e-06', 'num_tokens': '3.151e+05', 'mean_token_accuracy': '0.6474', 'epoch': '0.08493'}
+{'loss': '1.5', 'grad_norm': '11.62', 'learning_rate': '9.151e-06', 'num_tokens': '3.159e+05', 'mean_token_accuracy': '0.599', 'epoch': '0.0851'}
+
  9%|▊         | 491/5758 [02:45<30:15,  2.90it/s]
  9%|▊         | 492/5758 [02:45<31:05,  2.82it/s]
                                                  

  9%|▊         | 492/5758 [02:45<31:05,  2.82it/s]
  9%|▊         | 493/5758 [02:45<29:28,  2.98it/s]
                                                  

  9%|▊         | 493/5758 [02:45<29:28,  2.98it/s]
  9%|▊         | 494/5758 [02:46<29:49,  2.94it/s]
                                                  

  9%|▊         | 494/5758 [02:46<29:49,  2.94it/s]
  9%|▊         | 495/5758 [02:46<29:48,  2.94it/s]
                                                  

  9%|▊         | 495/5758 [02:46<29:48,  2.94it/s]
  9%|▊         | 496/5758 [02:47<30:02,  2.92it/s]
                                                  

  9%|▊         | 496/5758 [02:47<30:02,  2.92it/s]
  9%|▊         | 497/5758 [02:47<29:09,  3.01it/s]
                                                  

  9%|▊         | 497/5758 [02:47<29:09,  3.01it/s]
  9%|▊         | 498/5758 [02:47<28:33,  3.07it/s]
                                                  

  9%|▊         | 498/5758 [02:47<28:33,  3.07it/s]
  9%|▊         | 499/5758 [02:47<29:32,  2.97it/s]
                                                  

  9%|▊         | 499/5758 [02:47<29:32,  2.97it/s]
  9%|▊         | 500/5758 [02:48<28:26,  3.08it/s]
                                                  

  9%|▊         | 500/5758 [02:48<28:26,  3.08it/s]{'loss': '1.894', 'grad_norm': '35', 'learning_rate': '9.149e-06', 'num_tokens': '3.16e+05', 'mean_token_accuracy': '0.6667', 'epoch': '0.08527'}
+{'loss': '2.192', 'grad_norm': '16.12', 'learning_rate': '9.147e-06', 'num_tokens': '3.166e+05', 'mean_token_accuracy': '0.5643', 'epoch': '0.08545'}
+{'loss': '0.8194', 'grad_norm': '8.875', 'learning_rate': '9.146e-06', 'num_tokens': '3.179e+05', 'mean_token_accuracy': '0.7558', 'epoch': '0.08562'}
+{'loss': '1.06', 'grad_norm': '17.5', 'learning_rate': '9.144e-06', 'num_tokens': '3.181e+05', 'mean_token_accuracy': '0.7845', 'epoch': '0.08579'}
+{'loss': '1.02', 'grad_norm': '12.75', 'learning_rate': '9.142e-06', 'num_tokens': '3.186e+05', 'mean_token_accuracy': '0.7639', 'epoch': '0.08597'}
+{'loss': '1.085', 'grad_norm': '13.44', 'learning_rate': '9.14e-06', 'num_tokens': '3.19e+05', 'mean_token_accuracy': '0.729', 'epoch': '0.08614'}
+{'loss': '1.537', 'grad_norm': '12.62', 'learning_rate': '9.139e-06', 'num_tokens': '3.198e+05', 'mean_token_accuracy': '0.6128', 'epoch': '0.08631'}
+{'loss': '1.763', 'grad_norm': '17.25', 'learning_rate': '9.137e-06', 'num_tokens': '3.201e+05', 'mean_token_accuracy': '0.6532', 'epoch': '0.08649'}
+{'loss': '2.39', 'grad_norm': '19.25', 'learning_rate': '9.135e-06', 'num_tokens': '3.206e+05', 'mean_token_accuracy': '0.5368', 'epoch': '0.08666'}
+{'loss': '1.357', 'grad_norm': '13.38', 'learning_rate': '9.133e-06', 'num_tokens': '3.213e+05', 'mean_token_accuracy': '0.643', 'epoch': '0.08684'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.20it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.19it/s]
+
  9%|▊         | 501/5758 [02:50<1:22:49,  1.06it/s]
                                                    

  9%|▊         | 501/5758 [02:50<1:22:49,  1.06it/s]
  9%|▊         | 502/5758 [02:51<1:06:49,  1.31it/s]
                                                    

  9%|▊         | 502/5758 [02:51<1:06:49,  1.31it/s]
  9%|▊         | 503/5758 [02:51<54:48,  1.60it/s]  
                                                  

  9%|▊         | 503/5758 [02:51<54:48,  1.60it/s]
  9%|▉         | 504/5758 [02:51<48:11,  1.82it/s]
                                                  

  9%|▉         | 504/5758 [02:51<48:11,  1.82it/s]
  9%|▉         | 505/5758 [02:52<42:32,  2.06it/s]
                                                  

  9%|▉         | 505/5758 [02:52<42:32,  2.06it/s]
  9%|▉         | 506/5758 [02:52<37:55,  2.31it/s]
                                                  

  9%|▉         | 506/5758 [02:52<37:55,  2.31it/s]
  9%|▉         | 507/5758 [02:52<36:37,  2.39it/s]
                                                  

  9%|▉         | 507/5758 [02:52<36:37,  2.39it/s]
  9%|▉         | 508/5758 [02:53<35:38,  2.45it/s]
                                                  

  9%|▉         | 508/5758 [02:53<35:38,  2.45it/s]
  9%|▉         | 509/5758 [02:53<33:30,  2.61it/s]
                                                  

  9%|▉         | 509/5758 [02:53<33:30,  2.61it/s]
  9%|▉         | 510/5758 [02:53<31:56,  2.74it/s]
                                                  

  9%|▉         | 510/5758 [02:53<31:56,  2.74it/s]
  9%|▉         | 511/5758 [02:54<31:33,  2.77it/s]
                                                  

  9%|▉         | 511/5758 [02:54<31:33,  2.77it/s]
  9%|▉         | 512/5758 [02:54<32:10,  2.72it/s]
                                                  

  9%|▉         | 512/5758 [02:54<32:10,  2.72it/s]
  9%|▉         | 513/5758 [02:54<32:25,  2.70it/s]
                                                  

  9%|▉         | 513/5758 [02:54<32:25,  2.70it/s]
  9%|▉         | 514/5758 [02:55<31:50,  2.75it/s]
                                                  

  9%|▉         | 514/5758 [02:55<31:50,  2.75it/s]
  9%|▉         | 515/5758 [02:55<31:22,  2.78it/s]
                                                  

  9%|▉         | 515/5758 [02:55<31:22,  2.78it/s]
  9%|▉         | 516/5758 [02:55<32:53,  2.66it/s]
                                                  

  9%|▉         | 516/5758 [02:55<32:53,  2.66it/s]
  9%|▉         | 517/5758 [02:56<31:57,  2.73it/s]
                                                  

  9%|▉         | 517/5758 [02:56<31:57,  2.73it/s]
  9%|▉         | 518/5758 [02:56<30:29,  2.86it/s]
                                                  

  9%|▉         | 518/5758 [02:56<30:29,  2.86it/s]
  9%|▉         | 519/5758 [02:56<28:56,  3.02it/s]
                                                  

  9%|▉         | 519/5758 [02:56<28:56,  3.02it/s]
  9%|▉         | 520/5758 [02:57<27:46,  3.14it/s]
                                                  

  9%|▉         | 520/5758 [02:57<27:46,  3.14it/s]
  9%|▉         | 521/5758 [02:57<27:53,  3.13it/s]
                                                  

  9%|▉         | 521/5758 [02:57<27:53,  3.13it/s]
  9%|▉         | 522/5758 [02:57<27:13,  3.20it/s]
                                                  

  9%|▉         | 522/5758 [02:57<27:13,  3.20it/s]
  9%|▉         | 523/5758 [02:58<28:42,  3.04it/s]
                                                  

  9%|▉         | 523/5758 [02:58<28:42,  3.04it/s]
  9%|▉         | 524/5758 [02:58<28:40,  3.04it/s]
                                                  

  9%|▉         | 524/5758 [02:58<28:40,  3.04it/s]
  9%|▉         | 525/5758 [02:58<30:05,  2.90it/s]
                                                  

  9%|▉         | 525/5758 [02:58<30:05,  2.90it/s]
  9%|▉         | 526/5758 [02:59<29:07,  2.99it/s]
                                                  

  9%|▉         | 526/5758 [02:59<29:07,  2.99it/s]
  9%|▉         | 527/5758 [02:59<31:42,  2.75it/s]
                                                  

  9%|▉         | 527/5758 [02:59<31:42,  2.75it/s]
  9%|▉         | 528/5758 [03:00<31:58,  2.73it/s]
                                                  

  9%|▉         | 528/5758 [03:00<31:58,  2.73it/s]
  9%|▉         | 529/5758 [03:00<29:54,  2.91it/s]
                                                  

  9%|▉         | 529/5758 [03:00<29:54,  2.91it/s]
  9%|▉         | 530/5758 [03:00<29:05,  2.99it/s]
                                                  

  9%|▉         | 530/5758 [03:00<29:05,  2.99it/s]
  9%|▉         | 531/5758 [03:01<30:15,  2.88it/s]
                                                  

  9%|▉         | 531/5758 [03:01<30:15,  2.88it/s]
  9%|▉         | 532/5758 [03:01<31:05,  2.80it/s]
                                                  

  9%|▉         | 532/5758 [03:01<31:05,  2.80it/s]
  9%|▉         | 533/5758 [03:01<29:39,  2.94it/s]
                                                  

  9%|▉         | 533/5758 [03:01<29:39,  2.94it/s]
  9%|▉         | 534/5758 [03:02<29:50,  2.92it/s]
                                                  

  9%|▉         | 534/5758 [03:02<29:50,  2.92it/s]
  9%|▉         | 535/5758 [03:02<28:50,  3.02it/s]
                                                  

  9%|▉         | 535/5758 [03:02<28:50,  3.02it/s]
  9%|▉         | 536/5758 [03:02<27:36,  3.15it/s]
                                                  

  9%|▉         | 536/5758 [03:02<27:36,  3.15it/s]
  9%|▉         | 537/5758 [03:02<27:16,  3.19it/s]
                                                  

  9%|▉         | 537/5758 [03:02<27:16,  3.19it/s]
  9%|▉         | 538/5758 [03:03<26:56,  3.23it/s]
                                                  

  9%|▉         | 538/5758 [03:03<26:56,  3.23it/s]
  9%|▉         | 539/5758 [03:03<27:23,  3.18it/s]
                                                  

  9%|▉         | 539/5758 [03:03<27:23,  3.18it/s]
  9%|▉         | 540/5758 [03:03<28:18,  3.07it/s]
                                                  

  9%|▉         | 540/5758 [03:03<28:18,  3.07it/s]
  9%|▉         | 541/5758 [03:04<31:30,  2.76it/s]
                                                  

  9%|▉         | 541/5758 [03:04<31:30,  2.76it/s]
  9%|▉         | 542/5758 [03:04<30:14,  2.87it/s]
                                                  

  9%|▉         | 542/5758 [03:04<30:14,  2.87it/s]
  9%|▉         | 543/5758 [03:04<28:49,  3.01it/s]
                                                  

  9%|▉         | 543/5758 [03:04<28:49,  3.01it/s]
  9%|▉         | 544/5758 [03:05<28:35,  3.04it/s]
                                                  

  9%|▉         | 544/5758 [03:05<28:35,  3.04it/s]
  9%|▉         | 545/5758 [03:05<30:11,  2.88it/s]
                                                  

  9%|▉         | 545/5758 [03:05<30:11,  2.88it/s]
  9%|▉         | 546/5758 [03:06<31:05,  2.79it/s]
                                                  

  9%|▉         | 546/5758 [03:06<31:05,  2.79it/s]
  9%|▉         | 547/5758 [03:06<29:27,  2.95it/s]
                                                  

  9%|▉         | 547/5758 [03:06<29:27,  2.95it/s]
 10%|▉         | 548/5758 [03:06<28:36,  3.04it/s]
                                                  

 10%|▉         | 548/5758 [03:06<28:36,  3.04it/s]
 10%|▉         | 549/5758 [03:06<28:11,  3.08it/s]
                                                  

 10%|▉         | 549/5758 [03:06<28:11,  3.08it/s]
 10%|▉         | 550/5758 [03:07<29:31,  2.94it/s]
                                                  

 10%|▉         | 550/5758 [03:07<29:31,  2.94it/s]
 10%|▉         | 551/5758 [03:07<30:40,  2.83it/s]
                                                  

 10%|▉         | 551/5758 [03:07<30:40,  2.83it/s]
 10%|▉         | 552/5758 [03:08<29:02,  2.99it/s]
                                                  

 10%|▉         | 552/5758 [03:08<29:02,  2.99it/s]
 10%|▉         | 553/5758 [03:08<29:58,  2.89it/s]
                                                  

 10%|▉         | 553/5758 [03:08<29:58,  2.89it/s]
 10%|▉         | 554/5758 [03:08<30:02,  2.89it/s]
                                                  

 10%|▉         | 554/5758 [03:08<30:02,  2.89it/s]
 10%|▉         | 555/5758 [03:09<29:41,  2.92it/s]
                                                  
{'loss': '1.174', 'grad_norm': '28.12', 'learning_rate': '9.132e-06', 'num_tokens': '3.214e+05', 'mean_token_accuracy': '0.729', 'epoch': '0.08701'}
+{'loss': '1.107', 'grad_norm': '29.38', 'learning_rate': '9.13e-06', 'num_tokens': '3.215e+05', 'mean_token_accuracy': '0.7172', 'epoch': '0.08718'}
+{'loss': '1.898', 'grad_norm': '19.25', 'learning_rate': '9.128e-06', 'num_tokens': '3.219e+05', 'mean_token_accuracy': '0.6188', 'epoch': '0.08736'}
+{'loss': '1.775', 'grad_norm': '20.62', 'learning_rate': '9.126e-06', 'num_tokens': '3.221e+05', 'mean_token_accuracy': '0.635', 'epoch': '0.08753'}
+{'loss': '1.569', 'grad_norm': '11.81', 'learning_rate': '9.125e-06', 'num_tokens': '3.23e+05', 'mean_token_accuracy': '0.6706', 'epoch': '0.0877'}
+{'loss': '1.977', 'grad_norm': '15.38', 'learning_rate': '9.123e-06', 'num_tokens': '3.237e+05', 'mean_token_accuracy': '0.5769', 'epoch': '0.08788'}
+{'loss': '2.166', 'grad_norm': '16.5', 'learning_rate': '9.121e-06', 'num_tokens': '3.243e+05', 'mean_token_accuracy': '0.5496', 'epoch': '0.08805'}
+{'loss': '2.313', 'grad_norm': '16.38', 'learning_rate': '9.119e-06', 'num_tokens': '3.248e+05', 'mean_token_accuracy': '0.5712', 'epoch': '0.08823'}
+{'loss': '1.17', 'grad_norm': '11.75', 'learning_rate': '9.118e-06', 'num_tokens': '3.256e+05', 'mean_token_accuracy': '0.66', 'epoch': '0.0884'}
+{'loss': '1.288', 'grad_norm': '11', 'learning_rate': '9.116e-06', 'num_tokens': '3.264e+05', 'mean_token_accuracy': '0.66', 'epoch': '0.08857'}
+{'loss': '1.258', 'grad_norm': '10.81', 'learning_rate': '9.114e-06', 'num_tokens': '3.274e+05', 'mean_token_accuracy': '0.6692', 'epoch': '0.08875'}
+{'loss': '0.7768', 'grad_norm': '10.19', 'learning_rate': '9.113e-06', 'num_tokens': '3.285e+05', 'mean_token_accuracy': '0.7809', 'epoch': '0.08892'}
+{'loss': '1.767', 'grad_norm': '36.25', 'learning_rate': '9.111e-06', 'num_tokens': '3.286e+05', 'mean_token_accuracy': '0.6083', 'epoch': '0.08909'}
+{'loss': '1.834', 'grad_norm': '17.25', 'learning_rate': '9.109e-06', 'num_tokens': '3.291e+05', 'mean_token_accuracy': '0.5938', 'epoch': '0.08927'}
+{'loss': '2.251', 'grad_norm': '21.5', 'learning_rate': '9.107e-06', 'num_tokens': '3.295e+05', 'mean_token_accuracy': '0.5754', 'epoch': '0.08944'}
+{'loss': '1.436', 'grad_norm': '13.06', 'learning_rate': '9.106e-06', 'num_tokens': '3.302e+05', 'mean_token_accuracy': '0.5997', 'epoch': '0.08961'}
+{'loss': '0.7757', 'grad_norm': '7.469', 'learning_rate': '9.104e-06', 'num_tokens': '3.319e+05', 'mean_token_accuracy': '0.7687', 'epoch': '0.08979'}
+{'loss': '1.366', 'grad_norm': '13.88', 'learning_rate': '9.102e-06', 'num_tokens': '3.327e+05', 'mean_token_accuracy': '0.6497', 'epoch': '0.08996'}
+{'loss': '1.162', 'grad_norm': '13.19', 'learning_rate': '9.1e-06', 'num_tokens': '3.333e+05', 'mean_token_accuracy': '0.6882', 'epoch': '0.09014'}
+{'loss': '1.187', 'grad_norm': '36.75', 'learning_rate': '9.099e-06', 'num_tokens': '3.333e+05', 'mean_token_accuracy': '0.7794', 'epoch': '0.09031'}
+{'loss': '0.5972', 'grad_norm': '7.031', 'learning_rate': '9.097e-06', 'num_tokens': '3.349e+05', 'mean_token_accuracy': '0.8235', 'epoch': '0.09048'}
+{'loss': '1.503', 'grad_norm': '13.06', 'learning_rate': '9.095e-06', 'num_tokens': '3.356e+05', 'mean_token_accuracy': '0.6009', 'epoch': '0.09066'}
+{'loss': '1.973', 'grad_norm': '14.12', 'learning_rate': '9.093e-06', 'num_tokens': '3.361e+05', 'mean_token_accuracy': '0.593', 'epoch': '0.09083'}
+{'loss': '0.671', 'grad_norm': '6.781', 'learning_rate': '9.092e-06', 'num_tokens': '3.378e+05', 'mean_token_accuracy': '0.8005', 'epoch': '0.091'}
+{'loss': '1.345', 'grad_norm': '13.5', 'learning_rate': '9.09e-06', 'num_tokens': '3.384e+05', 'mean_token_accuracy': '0.6362', 'epoch': '0.09118'}
+{'loss': '1.991', 'grad_norm': '23.12', 'learning_rate': '9.088e-06', 'num_tokens': '3.385e+05', 'mean_token_accuracy': '0.5763', 'epoch': '0.09135'}
+{'loss': '1.369', 'grad_norm': '7.594', 'learning_rate': '9.086e-06', 'num_tokens': '3.405e+05', 'mean_token_accuracy': '0.6827', 'epoch': '0.09152'}
+{'loss': '0.9894', 'grad_norm': '31.88', 'learning_rate': '9.085e-06', 'num_tokens': '3.406e+05', 'mean_token_accuracy': '0.7742', 'epoch': '0.0917'}
+{'loss': '1.311', 'grad_norm': '13.06', 'learning_rate': '9.083e-06', 'num_tokens': '3.412e+05', 'mean_token_accuracy': '0.6361', 'epoch': '0.09187'}
+{'loss': '1.401', 'grad_norm': '12.44', 'learning_rate': '9.081e-06', 'num_tokens': '3.42e+05', 'mean_token_accuracy': '0.628', 'epoch': '0.09205'}
+{'loss': '2.763', 'grad_norm': '36.25', 'learning_rate': '9.08e-06', 'num_tokens': '3.421e+05', 'mean_token_accuracy': '0.5159', 'epoch': '0.09222'}
+{'loss': '2.29', 'grad_norm': '18', 'learning_rate': '9.078e-06', 'num_tokens': '3.426e+05', 'mean_token_accuracy': '0.5373', 'epoch': '0.09239'}
+{'loss': '1.37', 'grad_norm': '12.25', 'learning_rate': '9.076e-06', 'num_tokens': '3.433e+05', 'mean_token_accuracy': '0.6198', 'epoch': '0.09257'}
+{'loss': '0.8607', 'grad_norm': '9.75', 'learning_rate': '9.074e-06', 'num_tokens': '3.442e+05', 'mean_token_accuracy': '0.7693', 'epoch': '0.09274'}
+{'loss': '1.929', 'grad_norm': '16.62', 'learning_rate': '9.073e-06', 'num_tokens': '3.449e+05', 'mean_token_accuracy': '0.6024', 'epoch': '0.09291'}
+{'loss': '1.45', 'grad_norm': '41.25', 'learning_rate': '9.071e-06', 'num_tokens': '3.45e+05', 'mean_token_accuracy': '0.7273', 'epoch': '0.09309'}
+{'loss': '0.9514', 'grad_norm': '34', 'learning_rate': '9.069e-06', 'num_tokens': '3.451e+05', 'mean_token_accuracy': '0.8182', 'epoch': '0.09326'}
+{'loss': '1.454', 'grad_norm': '12.38', 'learning_rate': '9.067e-06', 'num_tokens': '3.457e+05', 'mean_token_accuracy': '0.637', 'epoch': '0.09344'}
+{'loss': '2.185', 'grad_norm': '15.25', 'learning_rate': '9.066e-06', 'num_tokens': '3.461e+05', 'mean_token_accuracy': '0.5842', 'epoch': '0.09361'}
+{'loss': '1.559', 'grad_norm': '19.25', 'learning_rate': '9.064e-06', 'num_tokens': '3.464e+05', 'mean_token_accuracy': '0.6875', 'epoch': '0.09378'}
+{'loss': '1.601', 'grad_norm': '11.62', 'learning_rate': '9.062e-06', 'num_tokens': '3.472e+05', 'mean_token_accuracy': '0.5958', 'epoch': '0.09396'}
+{'loss': '1.31', 'grad_norm': '12.38', 'learning_rate': '9.06e-06', 'num_tokens': '3.479e+05', 'mean_token_accuracy': '0.6433', 'epoch': '0.09413'}
+{'loss': '1.168', 'grad_norm': '12.56', 'learning_rate': '9.059e-06', 'num_tokens': '3.485e+05', 'mean_token_accuracy': '0.6672', 'epoch': '0.0943'}
+{'loss': '2.389', 'grad_norm': '13.19', 'learning_rate': '9.057e-06', 'num_tokens': '3.493e+05', 'mean_token_accuracy': '0.529', 'epoch': '0.09448'}
+{'loss': '1.961', 'grad_norm': '10.44', 'learning_rate': '9.055e-06', 'num_tokens': '3.505e+05', 'mean_token_accuracy': '0.5789', 'epoch': '0.09465'}
+{'loss': '2.257', 'grad_norm': '17.12', 'learning_rate': '9.053e-06', 'num_tokens': '3.511e+05', 'mean_token_accuracy': '0.5552', 'epoch': '0.09482'}
+{'loss': '1.402', 'grad_norm': '13.06', 'learning_rate': '9.052e-06', 'num_tokens': '3.517e+05', 'mean_token_accuracy': '0.6345', 'epoch': '0.095'}
+{'loss': '0.9906', 'grad_norm': '15.75', 'learning_rate': '9.05e-06', 'num_tokens': '3.521e+05', 'mean_token_accuracy': '0.8085', 'epoch': '0.09517'}
+{'loss': '1.378', 'grad_norm': '12.25', 'learning_rate': '9.048e-06', 'num_tokens': '3.528e+05', 'mean_token_accuracy': '0.6029', 'epoch': '0.09535'}
+{'loss': '2.301', 'grad_norm': '10.19', 'learning_rate': '9.047e-06', 'num_tokens': '3.539e+05', 'mean_token_accuracy': '0.5097', 'epoch': '0.09552'}
+{'loss': '0.9448', 'grad_norm': '28.38', 'learning_rate': '9.045e-06', 'num_tokens': '3.54e+05', 'mean_token_accuracy': '0.7846', 'epoch': '0.09569'}
+{'loss': '1.848', 'grad_norm': '33', 'learning_rate': '9.043e-06', 'num_tokens': '3.542e+05', 'mean_token_accuracy': '0.6093', 'epoch': '0.09587'}
+{'loss': '1.734', 'grad_norm': '28.25', 'learning_rate': '9.041e-06', 'num_tokens': '3.545e+05', 'mean_token_accuracy': '0.5165', 'epoch': '0.09604'}
+{'loss': '1.914', 'grad_norm': '15.12', 'learning_rate': '9.04e-06', 'num_tokens': '3.549e+05', 'mean_token_accuracy': '0.5878', 'epoch': '0.09621'}
+
 10%|▉         | 555/5758 [03:09<29:41,  2.92it/s]
 10%|▉         | 556/5758 [03:09<29:02,  2.99it/s]
                                                  

 10%|▉         | 556/5758 [03:09<29:02,  2.99it/s]
 10%|▉         | 557/5758 [03:09<30:46,  2.82it/s]
                                                  

 10%|▉         | 557/5758 [03:09<30:46,  2.82it/s]
 10%|▉         | 558/5758 [03:10<30:34,  2.83it/s]
                                                  

 10%|▉         | 558/5758 [03:10<30:34,  2.83it/s]
 10%|▉         | 559/5758 [03:10<31:19,  2.77it/s]
                                                  

 10%|▉         | 559/5758 [03:10<31:19,  2.77it/s]
 10%|▉         | 560/5758 [03:10<29:24,  2.95it/s]
                                                  

 10%|▉         | 560/5758 [03:10<29:24,  2.95it/s]
 10%|▉         | 561/5758 [03:11<30:33,  2.83it/s]
                                                  

 10%|▉         | 561/5758 [03:11<30:33,  2.83it/s]
 10%|▉         | 562/5758 [03:11<31:14,  2.77it/s]
                                                  

 10%|▉         | 562/5758 [03:11<31:14,  2.77it/s]
 10%|▉         | 563/5758 [03:11<30:23,  2.85it/s]
                                                  

 10%|▉         | 563/5758 [03:11<30:23,  2.85it/s]
 10%|▉         | 564/5758 [03:12<29:07,  2.97it/s]
                                                  

 10%|▉         | 564/5758 [03:12<29:07,  2.97it/s]
 10%|▉         | 565/5758 [03:12<31:37,  2.74it/s]
                                                  

 10%|▉         | 565/5758 [03:12<31:37,  2.74it/s]
 10%|▉         | 566/5758 [03:12<29:49,  2.90it/s]
                                                  

 10%|▉         | 566/5758 [03:12<29:49,  2.90it/s]
 10%|▉         | 567/5758 [03:13<29:52,  2.90it/s]
                                                  

 10%|▉         | 567/5758 [03:13<29:52,  2.90it/s]
 10%|▉         | 568/5758 [03:13<28:23,  3.05it/s]
                                                  

 10%|▉         | 568/5758 [03:13<28:23,  3.05it/s]
 10%|▉         | 569/5758 [03:13<27:38,  3.13it/s]
                                                  

 10%|▉         | 569/5758 [03:13<27:38,  3.13it/s]
 10%|▉         | 570/5758 [03:14<28:29,  3.03it/s]
                                                  

 10%|▉         | 570/5758 [03:14<28:29,  3.03it/s]
 10%|▉         | 571/5758 [03:14<28:46,  3.00it/s]
                                                  

 10%|▉         | 571/5758 [03:14<28:46,  3.00it/s]
 10%|▉         | 572/5758 [03:14<28:21,  3.05it/s]
                                                  

 10%|▉         | 572/5758 [03:14<28:21,  3.05it/s]
 10%|▉         | 573/5758 [03:15<30:01,  2.88it/s]
                                                  

 10%|▉         | 573/5758 [03:15<30:01,  2.88it/s]
 10%|▉         | 574/5758 [03:15<31:01,  2.78it/s]
                                                  

 10%|▉         | 574/5758 [03:15<31:01,  2.78it/s]
 10%|▉         | 575/5758 [03:16<31:05,  2.78it/s]
                                                  

 10%|▉         | 575/5758 [03:16<31:05,  2.78it/s]
 10%|█         | 576/5758 [03:16<30:22,  2.84it/s]
                                                  

 10%|█         | 576/5758 [03:16<30:22,  2.84it/s]
 10%|█         | 577/5758 [03:16<29:18,  2.95it/s]
                                                  

 10%|█         | 577/5758 [03:16<29:18,  2.95it/s]
 10%|█         | 578/5758 [03:16<28:35,  3.02it/s]
                                                  

 10%|█         | 578/5758 [03:16<28:35,  3.02it/s]
 10%|█         | 579/5758 [03:17<27:36,  3.13it/s]
                                                  

 10%|█         | 579/5758 [03:17<27:36,  3.13it/s]
 10%|█         | 580/5758 [03:17<27:24,  3.15it/s]
                                                  

 10%|█         | 580/5758 [03:17<27:24,  3.15it/s]
 10%|█         | 581/5758 [03:17<27:51,  3.10it/s]
                                                  

 10%|█         | 581/5758 [03:17<27:51,  3.10it/s]
 10%|█         | 582/5758 [03:18<28:34,  3.02it/s]
                                                  

 10%|█         | 582/5758 [03:18<28:34,  3.02it/s]
 10%|█         | 583/5758 [03:18<28:23,  3.04it/s]
                                                  

 10%|█         | 583/5758 [03:18<28:23,  3.04it/s]
 10%|█         | 584/5758 [03:18<27:44,  3.11it/s]
                                                  

 10%|█         | 584/5758 [03:18<27:44,  3.11it/s]
 10%|█         | 585/5758 [03:19<29:08,  2.96it/s]
                                                  

 10%|█         | 585/5758 [03:19<29:08,  2.96it/s]
 10%|█         | 586/5758 [03:19<28:33,  3.02it/s]
                                                  

 10%|█         | 586/5758 [03:19<28:33,  3.02it/s]
 10%|█         | 587/5758 [03:19<28:48,  2.99it/s]
                                                  

 10%|█         | 587/5758 [03:19<28:48,  2.99it/s]
 10%|█         | 588/5758 [03:20<27:59,  3.08it/s]
                                                  

 10%|█         | 588/5758 [03:20<27:59,  3.08it/s]
 10%|█         | 589/5758 [03:20<27:40,  3.11it/s]
                                                  

 10%|█         | 589/5758 [03:20<27:40,  3.11it/s]
 10%|█         | 590/5758 [03:20<28:00,  3.08it/s]
                                                  

 10%|█         | 590/5758 [03:20<28:00,  3.08it/s]
 10%|█         | 591/5758 [03:21<27:36,  3.12it/s]
                                                  

 10%|█         | 591/5758 [03:21<27:36,  3.12it/s]
 10%|█         | 592/5758 [03:21<29:23,  2.93it/s]
                                                  

 10%|█         | 592/5758 [03:21<29:23,  2.93it/s]
 10%|█         | 593/5758 [03:21<28:57,  2.97it/s]
                                                  

 10%|█         | 593/5758 [03:21<28:57,  2.97it/s]
 10%|█         | 594/5758 [03:22<31:12,  2.76it/s]
                                                  

 10%|█         | 594/5758 [03:22<31:12,  2.76it/s]
 10%|█         | 595/5758 [03:22<31:57,  2.69it/s]
                                                  

 10%|█         | 595/5758 [03:22<31:57,  2.69it/s]
 10%|█         | 596/5758 [03:23<30:58,  2.78it/s]
                                                  

 10%|█         | 596/5758 [03:23<30:58,  2.78it/s]
 10%|█         | 597/5758 [03:23<29:29,  2.92it/s]
                                                  

 10%|█         | 597/5758 [03:23<29:29,  2.92it/s]
 10%|█         | 598/5758 [03:23<30:02,  2.86it/s]
                                                  

 10%|█         | 598/5758 [03:23<30:02,  2.86it/s]
 10%|█         | 599/5758 [03:24<30:51,  2.79it/s]
                                                  

 10%|█         | 599/5758 [03:24<30:51,  2.79it/s]
 10%|█         | 600/5758 [03:24<29:43,  2.89it/s]
                                                  

 10%|█         | 600/5758 [03:24<29:43,  2.89it/s]
 10%|█         | 601/5758 [03:24<28:18,  3.04it/s]
                                                  

 10%|█         | 601/5758 [03:24<28:18,  3.04it/s]
 10%|█         | 602/5758 [03:25<27:35,  3.11it/s]
                                                  

 10%|█         | 602/5758 [03:25<27:35,  3.11it/s]
 10%|█         | 603/5758 [03:25<29:21,  2.93it/s]
                                                  

 10%|█         | 603/5758 [03:25<29:21,  2.93it/s]
 10%|█         | 604/5758 [03:25<29:46,  2.89it/s]
                                                  

 10%|█         | 604/5758 [03:25<29:46,  2.89it/s]
 11%|█         | 605/5758 [03:26<30:15,  2.84it/s]
                                                  

 11%|█         | 605/5758 [03:26<30:15,  2.84it/s]
 11%|█         | 606/5758 [03:26<30:15,  2.84it/s]
                                                  

 11%|█         | 606/5758 [03:26<30:15,  2.84it/s]
 11%|█         | 607/5758 [03:26<29:04,  2.95it/s]
                                                  

 11%|█         | 607/5758 [03:26<29:04,  2.95it/s]
 11%|█         | 608/5758 [03:27<29:06,  2.95it/s]
                                                  

 11%|█         | 608/5758 [03:27<29:06,  2.95it/s]
 11%|█         | 609/5758 [03:27<28:46,  2.98it/s]
                                                  

 11%|█         | 609/5758 [03:27<28:46,  2.98it/s]
 11%|█         | 610/5758 [03:27<28:33,  3.01it/s]
                                                  
{'loss': '1.22', 'grad_norm': '21', 'learning_rate': '9.038e-06', 'num_tokens': '3.551e+05', 'mean_token_accuracy': '0.7512', 'epoch': '0.09639'}
+{'loss': '1.012', 'grad_norm': '24.12', 'learning_rate': '9.036e-06', 'num_tokens': '3.553e+05', 'mean_token_accuracy': '0.7663', 'epoch': '0.09656'}
+{'loss': '1.307', 'grad_norm': '6.188', 'learning_rate': '9.034e-06', 'num_tokens': '3.583e+05', 'mean_token_accuracy': '0.6969', 'epoch': '0.09673'}
+{'loss': '0.5417', 'grad_norm': '6.125', 'learning_rate': '9.033e-06', 'num_tokens': '3.602e+05', 'mean_token_accuracy': '0.841', 'epoch': '0.09691'}
+{'loss': '1.712', 'grad_norm': '29.25', 'learning_rate': '9.031e-06', 'num_tokens': '3.603e+05', 'mean_token_accuracy': '0.6953', 'epoch': '0.09708'}
+{'loss': '1.165', 'grad_norm': '12.5', 'learning_rate': '9.029e-06', 'num_tokens': '3.609e+05', 'mean_token_accuracy': '0.7131', 'epoch': '0.09726'}
+{'loss': '1.007', 'grad_norm': '11.88', 'learning_rate': '9.027e-06', 'num_tokens': '3.615e+05', 'mean_token_accuracy': '0.7526', 'epoch': '0.09743'}
+{'loss': '2.127', 'grad_norm': '13.56', 'learning_rate': '9.026e-06', 'num_tokens': '3.62e+05', 'mean_token_accuracy': '0.5542', 'epoch': '0.0976'}
+{'loss': '2.313', 'grad_norm': '10.81', 'learning_rate': '9.024e-06', 'num_tokens': '3.628e+05', 'mean_token_accuracy': '0.5417', 'epoch': '0.09778'}
+{'loss': '1.352', 'grad_norm': '12.56', 'learning_rate': '9.022e-06', 'num_tokens': '3.635e+05', 'mean_token_accuracy': '0.6384', 'epoch': '0.09795'}
+{'loss': '1.332', 'grad_norm': '12.06', 'learning_rate': '9.02e-06', 'num_tokens': '3.642e+05', 'mean_token_accuracy': '0.6402', 'epoch': '0.09812'}
+{'loss': '1.134', 'grad_norm': '27.38', 'learning_rate': '9.019e-06', 'num_tokens': '3.644e+05', 'mean_token_accuracy': '0.7044', 'epoch': '0.0983'}
+{'loss': '1.134', 'grad_norm': '17.62', 'learning_rate': '9.017e-06', 'num_tokens': '3.646e+05', 'mean_token_accuracy': '0.7913', 'epoch': '0.09847'}
+{'loss': '1.095', 'grad_norm': '42.75', 'learning_rate': '9.015e-06', 'num_tokens': '3.647e+05', 'mean_token_accuracy': '0.7353', 'epoch': '0.09865'}
+{'loss': '2.147', 'grad_norm': '26.38', 'learning_rate': '9.014e-06', 'num_tokens': '3.649e+05', 'mean_token_accuracy': '0.5928', 'epoch': '0.09882'}
+{'loss': '2.273', 'grad_norm': '20.62', 'learning_rate': '9.012e-06', 'num_tokens': '3.651e+05', 'mean_token_accuracy': '0.5274', 'epoch': '0.09899'}
+{'loss': '0.769', 'grad_norm': '7.594', 'learning_rate': '9.01e-06', 'num_tokens': '3.668e+05', 'mean_token_accuracy': '0.7643', 'epoch': '0.09917'}
+{'loss': '1.411', 'grad_norm': '12.5', 'learning_rate': '9.008e-06', 'num_tokens': '3.675e+05', 'mean_token_accuracy': '0.6294', 'epoch': '0.09934'}
+{'loss': '1.231', 'grad_norm': '12.56', 'learning_rate': '9.007e-06', 'num_tokens': '3.681e+05', 'mean_token_accuracy': '0.6701', 'epoch': '0.09951'}
+{'loss': '1.977', 'grad_norm': '28.5', 'learning_rate': '9.005e-06', 'num_tokens': '3.682e+05', 'mean_token_accuracy': '0.5814', 'epoch': '0.09969'}
+{'loss': '1.788', 'grad_norm': '21.38', 'learning_rate': '9.003e-06', 'num_tokens': '3.685e+05', 'mean_token_accuracy': '0.629', 'epoch': '0.09986'}
+{'loss': '1.487', 'grad_norm': '11.38', 'learning_rate': '9.001e-06', 'num_tokens': '3.693e+05', 'mean_token_accuracy': '0.616', 'epoch': '0.1'}
+{'loss': '1.664', 'grad_norm': '40.25', 'learning_rate': '9e-06', 'num_tokens': '3.694e+05', 'mean_token_accuracy': '0.6203', 'epoch': '0.1002'}
+{'loss': '1.076', 'grad_norm': '22.38', 'learning_rate': '8.998e-06', 'num_tokens': '3.696e+05', 'mean_token_accuracy': '0.7717', 'epoch': '0.1004'}
+{'loss': '1.339', 'grad_norm': '9.062', 'learning_rate': '8.996e-06', 'num_tokens': '3.708e+05', 'mean_token_accuracy': '0.6637', 'epoch': '0.1006'}
+{'loss': '1.182', 'grad_norm': '12.62', 'learning_rate': '8.994e-06', 'num_tokens': '3.715e+05', 'mean_token_accuracy': '0.6662', 'epoch': '0.1007'}
+{'loss': '1.273', 'grad_norm': '20.38', 'learning_rate': '8.993e-06', 'num_tokens': '3.717e+05', 'mean_token_accuracy': '0.7143', 'epoch': '0.1009'}
+{'loss': '1.439', 'grad_norm': '28.12', 'learning_rate': '8.991e-06', 'num_tokens': '3.718e+05', 'mean_token_accuracy': '0.7103', 'epoch': '0.1011'}
+{'loss': '0.6579', 'grad_norm': '7.062', 'learning_rate': '8.989e-06', 'num_tokens': '3.734e+05', 'mean_token_accuracy': '0.7946', 'epoch': '0.1013'}
+{'loss': '1.267', 'grad_norm': '12.19', 'learning_rate': '8.987e-06', 'num_tokens': '3.741e+05', 'mean_token_accuracy': '0.6542', 'epoch': '0.1014'}
+{'loss': '0.9145', 'grad_norm': '15.88', 'learning_rate': '8.986e-06', 'num_tokens': '3.743e+05', 'mean_token_accuracy': '0.8139', 'epoch': '0.1016'}
+{'loss': '1.758', 'grad_norm': '17', 'learning_rate': '8.984e-06', 'num_tokens': '3.747e+05', 'mean_token_accuracy': '0.6274', 'epoch': '0.1018'}
+{'loss': '1.394', 'grad_norm': '7.562', 'learning_rate': '8.982e-06', 'num_tokens': '3.765e+05', 'mean_token_accuracy': '0.6864', 'epoch': '0.1019'}
+{'loss': '1.609', 'grad_norm': '12', 'learning_rate': '8.981e-06', 'num_tokens': '3.771e+05', 'mean_token_accuracy': '0.6385', 'epoch': '0.1021'}
+{'loss': '1.091', 'grad_norm': '35.5', 'learning_rate': '8.979e-06', 'num_tokens': '3.772e+05', 'mean_token_accuracy': '0.8125', 'epoch': '0.1023'}
+{'loss': '1.328', 'grad_norm': '8.75', 'learning_rate': '8.977e-06', 'num_tokens': '3.787e+05', 'mean_token_accuracy': '0.6928', 'epoch': '0.1025'}
+{'loss': '1.3', 'grad_norm': '13.38', 'learning_rate': '8.975e-06', 'num_tokens': '3.793e+05', 'mean_token_accuracy': '0.6416', 'epoch': '0.1026'}
+{'loss': '2.471', 'grad_norm': '13.31', 'learning_rate': '8.974e-06', 'num_tokens': '3.799e+05', 'mean_token_accuracy': '0.5234', 'epoch': '0.1028'}
+{'loss': '1.531', 'grad_norm': '26.38', 'learning_rate': '8.972e-06', 'num_tokens': '3.801e+05', 'mean_token_accuracy': '0.6968', 'epoch': '0.103'}
+{'loss': '0.8505', 'grad_norm': '16.62', 'learning_rate': '8.97e-06', 'num_tokens': '3.804e+05', 'mean_token_accuracy': '0.7935', 'epoch': '0.1032'}
+{'loss': '1.134', 'grad_norm': '13.31', 'learning_rate': '8.968e-06', 'num_tokens': '3.81e+05', 'mean_token_accuracy': '0.6798', 'epoch': '0.1033'}
+{'loss': '2.047', 'grad_norm': '24.5', 'learning_rate': '8.967e-06', 'num_tokens': '3.812e+05', 'mean_token_accuracy': '0.6019', 'epoch': '0.1035'}
+{'loss': '2.285', 'grad_norm': '17.5', 'learning_rate': '8.965e-06', 'num_tokens': '3.819e+05', 'mean_token_accuracy': '0.5464', 'epoch': '0.1037'}
+{'loss': '1.893', 'grad_norm': '33.25', 'learning_rate': '8.963e-06', 'num_tokens': '3.82e+05', 'mean_token_accuracy': '0.6154', 'epoch': '0.1039'}
+{'loss': '0.644', 'grad_norm': '8.812', 'learning_rate': '8.961e-06', 'num_tokens': '3.831e+05', 'mean_token_accuracy': '0.8014', 'epoch': '0.104'}
+{'loss': '2.097', 'grad_norm': '17.88', 'learning_rate': '8.96e-06', 'num_tokens': '3.835e+05', 'mean_token_accuracy': '0.5595', 'epoch': '0.1042'}
+{'loss': '1.043', 'grad_norm': '11.44', 'learning_rate': '8.958e-06', 'num_tokens': '3.841e+05', 'mean_token_accuracy': '0.7114', 'epoch': '0.1044'}
+{'loss': '1.368', 'grad_norm': '22.38', 'learning_rate': '8.956e-06', 'num_tokens': '3.843e+05', 'mean_token_accuracy': '0.7427', 'epoch': '0.1046'}
+{'loss': '2.535', 'grad_norm': '23.12', 'learning_rate': '8.954e-06', 'num_tokens': '3.846e+05', 'mean_token_accuracy': '0.5625', 'epoch': '0.1047'}
+{'loss': '1.375', 'grad_norm': '17.12', 'learning_rate': '8.953e-06', 'num_tokens': '3.851e+05', 'mean_token_accuracy': '0.6413', 'epoch': '0.1049'}
+{'loss': '1.223', 'grad_norm': '6', 'learning_rate': '8.951e-06', 'num_tokens': '3.892e+05', 'mean_token_accuracy': '0.6867', 'epoch': '0.1051'}
+{'loss': '1.408', 'grad_norm': '15.38', 'learning_rate': '8.949e-06', 'num_tokens': '3.897e+05', 'mean_token_accuracy': '0.6792', 'epoch': '0.1052'}
+{'loss': '1.26', 'grad_norm': '13', 'learning_rate': '8.948e-06', 'num_tokens': '3.904e+05', 'mean_token_accuracy': '0.6344', 'epoch': '0.1054'}
+{'loss': '1.342', 'grad_norm': '8.062', 'learning_rate': '8.946e-06', 'num_tokens': '3.921e+05', 'mean_token_accuracy': '0.6813', 'epoch': '0.1056'}
+{'loss': '1.351', 'grad_norm': '12.12', 'learning_rate': '8.944e-06', 'num_tokens': '3.929e+05', 'mean_token_accuracy': '0.6971', 'epoch': '0.1058'}
+
 11%|█         | 610/5758 [03:27<28:33,  3.01it/s]
 11%|█         | 611/5758 [03:28<30:05,  2.85it/s]
                                                  

 11%|█         | 611/5758 [03:28<30:05,  2.85it/s]
 11%|█         | 612/5758 [03:28<29:49,  2.88it/s]
                                                  

 11%|█         | 612/5758 [03:28<29:49,  2.88it/s]
 11%|█         | 613/5758 [03:28<30:24,  2.82it/s]
                                                  

 11%|█         | 613/5758 [03:28<30:24,  2.82it/s]
 11%|█         | 614/5758 [03:29<30:38,  2.80it/s]
                                                  

 11%|█         | 614/5758 [03:29<30:38,  2.80it/s]
 11%|█         | 615/5758 [03:29<29:56,  2.86it/s]
                                                  

 11%|█         | 615/5758 [03:29<29:56,  2.86it/s]
 11%|█         | 616/5758 [03:29<30:22,  2.82it/s]
                                                  

 11%|█         | 616/5758 [03:29<30:22,  2.82it/s]
 11%|█         | 617/5758 [03:30<30:43,  2.79it/s]
                                                  

 11%|█         | 617/5758 [03:30<30:43,  2.79it/s]
 11%|█         | 618/5758 [03:30<30:10,  2.84it/s]
                                                  

 11%|█         | 618/5758 [03:30<30:10,  2.84it/s]
 11%|█         | 619/5758 [03:30<28:41,  2.99it/s]
                                                  

 11%|█         | 619/5758 [03:30<28:41,  2.99it/s]
 11%|█         | 620/5758 [03:31<30:49,  2.78it/s]
                                                  

 11%|█         | 620/5758 [03:31<30:49,  2.78it/s]
 11%|█         | 621/5758 [03:31<29:32,  2.90it/s]
                                                  

 11%|█         | 621/5758 [03:31<29:32,  2.90it/s]
 11%|█         | 622/5758 [03:31<28:04,  3.05it/s]
                                                  

 11%|█         | 622/5758 [03:31<28:04,  3.05it/s]
 11%|█         | 623/5758 [03:32<28:50,  2.97it/s]
                                                  

 11%|█         | 623/5758 [03:32<28:50,  2.97it/s]
 11%|█         | 624/5758 [03:32<28:08,  3.04it/s]
                                                  

 11%|█         | 624/5758 [03:32<28:08,  3.04it/s]
 11%|█         | 625/5758 [03:32<28:47,  2.97it/s]
                                                  

 11%|█         | 625/5758 [03:32<28:47,  2.97it/s]
 11%|█         | 626/5758 [03:33<29:09,  2.93it/s]
                                                  

 11%|█         | 626/5758 [03:33<29:09,  2.93it/s]
 11%|█         | 627/5758 [03:33<27:55,  3.06it/s]
                                                  

 11%|█         | 627/5758 [03:33<27:55,  3.06it/s]
 11%|█         | 628/5758 [03:33<27:19,  3.13it/s]
                                                  

 11%|█         | 628/5758 [03:33<27:19,  3.13it/s]
 11%|█         | 629/5758 [03:34<28:32,  3.00it/s]
                                                  

 11%|█         | 629/5758 [03:34<28:32,  3.00it/s]
 11%|█         | 630/5758 [03:34<27:30,  3.11it/s]
                                                  

 11%|█         | 630/5758 [03:34<27:30,  3.11it/s]
 11%|█         | 631/5758 [03:34<28:59,  2.95it/s]
                                                  

 11%|█         | 631/5758 [03:34<28:59,  2.95it/s]
 11%|█         | 632/5758 [03:35<27:53,  3.06it/s]
                                                  

 11%|█         | 632/5758 [03:35<27:53,  3.06it/s]
 11%|█         | 633/5758 [03:35<27:14,  3.14it/s]
                                                  

 11%|█         | 633/5758 [03:35<27:14,  3.14it/s]
 11%|█         | 634/5758 [03:35<27:32,  3.10it/s]
                                                  

 11%|█         | 634/5758 [03:35<27:32,  3.10it/s]
 11%|█         | 635/5758 [03:36<26:43,  3.19it/s]
                                                  

 11%|█         | 635/5758 [03:36<26:43,  3.19it/s]
 11%|█         | 636/5758 [03:36<26:43,  3.19it/s]
                                                  

 11%|█         | 636/5758 [03:36<26:43,  3.19it/s]
 11%|█         | 637/5758 [03:36<28:27,  3.00it/s]
                                                  

 11%|█         | 637/5758 [03:36<28:27,  3.00it/s]
 11%|█         | 638/5758 [03:37<29:17,  2.91it/s]
                                                  

 11%|█         | 638/5758 [03:37<29:17,  2.91it/s]
 11%|█         | 639/5758 [03:37<28:37,  2.98it/s]
                                                  

 11%|█         | 639/5758 [03:37<28:37,  2.98it/s]
 11%|█         | 640/5758 [03:37<27:35,  3.09it/s]
                                                  

 11%|█         | 640/5758 [03:37<27:35,  3.09it/s]
 11%|█         | 641/5758 [03:38<26:51,  3.17it/s]
                                                  

 11%|█         | 641/5758 [03:38<26:51,  3.17it/s]
 11%|█         | 642/5758 [03:38<27:16,  3.13it/s]
                                                  

 11%|█         | 642/5758 [03:38<27:16,  3.13it/s]
 11%|█         | 643/5758 [03:38<26:57,  3.16it/s]
                                                  

 11%|█         | 643/5758 [03:38<26:57,  3.16it/s]
 11%|█         | 644/5758 [03:39<27:16,  3.13it/s]
                                                  

 11%|█         | 644/5758 [03:39<27:16,  3.13it/s]
 11%|█         | 645/5758 [03:39<26:49,  3.18it/s]
                                                  

 11%|█         | 645/5758 [03:39<26:49,  3.18it/s]
 11%|█         | 646/5758 [03:39<31:39,  2.69it/s]
                                                  

 11%|█         | 646/5758 [03:39<31:39,  2.69it/s]
 11%|█         | 647/5758 [03:40<29:44,  2.86it/s]
                                                  

 11%|█         | 647/5758 [03:40<29:44,  2.86it/s]
 11%|█▏        | 648/5758 [03:40<28:31,  2.99it/s]
                                                  

 11%|█▏        | 648/5758 [03:40<28:31,  2.99it/s]
 11%|█▏        | 649/5758 [03:40<29:34,  2.88it/s]
                                                  

 11%|█▏        | 649/5758 [03:40<29:34,  2.88it/s]
 11%|█▏        | 650/5758 [03:41<28:54,  2.94it/s]
                                                  

 11%|█▏        | 650/5758 [03:41<28:54,  2.94it/s]
 11%|█▏        | 651/5758 [03:41<30:03,  2.83it/s]
                                                  

 11%|█▏        | 651/5758 [03:41<30:03,  2.83it/s]
 11%|█▏        | 652/5758 [03:41<29:53,  2.85it/s]
                                                  

 11%|█▏        | 652/5758 [03:41<29:53,  2.85it/s]
 11%|█▏        | 653/5758 [03:42<28:37,  2.97it/s]
                                                  

 11%|█▏        | 653/5758 [03:42<28:37,  2.97it/s]
 11%|█▏        | 654/5758 [03:42<27:54,  3.05it/s]
                                                  

 11%|█▏        | 654/5758 [03:42<27:54,  3.05it/s]
 11%|█▏        | 655/5758 [03:42<26:57,  3.16it/s]
                                                  

 11%|█▏        | 655/5758 [03:42<26:57,  3.16it/s]
 11%|█▏        | 656/5758 [03:43<27:50,  3.05it/s]
                                                  

 11%|█▏        | 656/5758 [03:43<27:50,  3.05it/s]
 11%|█▏        | 657/5758 [03:43<27:43,  3.07it/s]
                                                  

 11%|█▏        | 657/5758 [03:43<27:43,  3.07it/s]
 11%|█▏        | 658/5758 [03:43<29:04,  2.92it/s]
                                                  

 11%|█▏        | 658/5758 [03:43<29:04,  2.92it/s]
 11%|█▏        | 659/5758 [03:44<28:21,  3.00it/s]
                                                  

 11%|█▏        | 659/5758 [03:44<28:21,  3.00it/s]
 11%|█▏        | 660/5758 [03:44<29:32,  2.88it/s]
                                                  

 11%|█▏        | 660/5758 [03:44<29:32,  2.88it/s]
 11%|█▏        | 661/5758 [03:44<28:41,  2.96it/s]
                                                  

 11%|█▏        | 661/5758 [03:44<28:41,  2.96it/s]
 11%|█▏        | 662/5758 [03:45<28:36,  2.97it/s]
                                                  

 11%|█▏        | 662/5758 [03:45<28:36,  2.97it/s]
 12%|█▏        | 663/5758 [03:45<27:54,  3.04it/s]
                                                  

 12%|█▏        | 663/5758 [03:45<27:54,  3.04it/s]
 12%|█▏        | 664/5758 [03:45<27:38,  3.07it/s]
                                                  

 12%|█▏        | 664/5758 [03:45<27:38,  3.07it/s]
 12%|█▏        | 665/5758 [03:46<28:08,  3.02it/s]
                                                  
{'loss': '1.553', 'grad_norm': '21.75', 'learning_rate': '8.942e-06', 'num_tokens': '3.931e+05', 'mean_token_accuracy': '0.6946', 'epoch': '0.1059'}
+{'loss': '1.142', 'grad_norm': '13', 'learning_rate': '8.941e-06', 'num_tokens': '3.937e+05', 'mean_token_accuracy': '0.6831', 'epoch': '0.1061'}
+{'loss': '1.359', 'grad_norm': '29.62', 'learning_rate': '8.939e-06', 'num_tokens': '3.938e+05', 'mean_token_accuracy': '0.6863', 'epoch': '0.1063'}
+{'loss': '1.367', 'grad_norm': '20.38', 'learning_rate': '8.937e-06', 'num_tokens': '3.941e+05', 'mean_token_accuracy': '0.7416', 'epoch': '0.1065'}
+{'loss': '1.305', 'grad_norm': '6.094', 'learning_rate': '8.935e-06', 'num_tokens': '3.982e+05', 'mean_token_accuracy': '0.7133', 'epoch': '0.1066'}
+{'loss': '0.762', 'grad_norm': '7.844', 'learning_rate': '8.934e-06', 'num_tokens': '3.998e+05', 'mean_token_accuracy': '0.7696', 'epoch': '0.1068'}
+{'loss': '1.506', 'grad_norm': '10.5', 'learning_rate': '8.932e-06', 'num_tokens': '4.008e+05', 'mean_token_accuracy': '0.5923', 'epoch': '0.107'}
+{'loss': '1.123', 'grad_norm': '9.188', 'learning_rate': '8.93e-06', 'num_tokens': '4.019e+05', 'mean_token_accuracy': '0.7379', 'epoch': '0.1072'}
+{'loss': '1.237', 'grad_norm': '11', 'learning_rate': '8.928e-06', 'num_tokens': '4.028e+05', 'mean_token_accuracy': '0.6444', 'epoch': '0.1073'}
+{'loss': '1.351', 'grad_norm': '12.88', 'learning_rate': '8.927e-06', 'num_tokens': '4.034e+05', 'mean_token_accuracy': '0.645', 'epoch': '0.1075'}
+{'loss': '1.511', 'grad_norm': '13.12', 'learning_rate': '8.925e-06', 'num_tokens': '4.041e+05', 'mean_token_accuracy': '0.6272', 'epoch': '0.1077'}
+{'loss': '1.386', 'grad_norm': '13.56', 'learning_rate': '8.923e-06', 'num_tokens': '4.048e+05', 'mean_token_accuracy': '0.6265', 'epoch': '0.1078'}
+{'loss': '0.7965', 'grad_norm': '21.88', 'learning_rate': '8.922e-06', 'num_tokens': '4.05e+05', 'mean_token_accuracy': '0.8411', 'epoch': '0.108'}
+{'loss': '1.911', 'grad_norm': '32.5', 'learning_rate': '8.92e-06', 'num_tokens': '4.051e+05', 'mean_token_accuracy': '0.641', 'epoch': '0.1082'}
+{'loss': '1.686', 'grad_norm': '15.88', 'learning_rate': '8.918e-06', 'num_tokens': '4.054e+05', 'mean_token_accuracy': '0.6462', 'epoch': '0.1084'}
+{'loss': '1.956', 'grad_norm': '16.25', 'learning_rate': '8.916e-06', 'num_tokens': '4.059e+05', 'mean_token_accuracy': '0.6098', 'epoch': '0.1085'}
+{'loss': '0.7952', 'grad_norm': '14.12', 'learning_rate': '8.915e-06', 'num_tokens': '4.061e+05', 'mean_token_accuracy': '0.8636', 'epoch': '0.1087'}
+{'loss': '1.875', 'grad_norm': '14.75', 'learning_rate': '8.913e-06', 'num_tokens': '4.068e+05', 'mean_token_accuracy': '0.5997', 'epoch': '0.1089'}
+{'loss': '1.131', 'grad_norm': '12.56', 'learning_rate': '8.911e-06', 'num_tokens': '4.075e+05', 'mean_token_accuracy': '0.6758', 'epoch': '0.1091'}
+{'loss': '1.748', 'grad_norm': '29.75', 'learning_rate': '8.909e-06', 'num_tokens': '4.076e+05', 'mean_token_accuracy': '0.661', 'epoch': '0.1092'}
+{'loss': '1.616', 'grad_norm': '13.69', 'learning_rate': '8.908e-06', 'num_tokens': '4.082e+05', 'mean_token_accuracy': '0.5779', 'epoch': '0.1094'}
+{'loss': '1.525', 'grad_norm': '10.38', 'learning_rate': '8.906e-06', 'num_tokens': '4.093e+05', 'mean_token_accuracy': '0.6733', 'epoch': '0.1096'}
+{'loss': '1.44', 'grad_norm': '13.5', 'learning_rate': '8.904e-06', 'num_tokens': '4.099e+05', 'mean_token_accuracy': '0.6234', 'epoch': '0.1098'}
+{'loss': '1.249', 'grad_norm': '13', 'learning_rate': '8.902e-06', 'num_tokens': '4.106e+05', 'mean_token_accuracy': '0.6712', 'epoch': '0.1099'}
+{'loss': '1.53', 'grad_norm': '12.31', 'learning_rate': '8.901e-06', 'num_tokens': '4.114e+05', 'mean_token_accuracy': '0.5906', 'epoch': '0.1101'}
+{'loss': '1.137', 'grad_norm': '12.44', 'learning_rate': '8.899e-06', 'num_tokens': '4.12e+05', 'mean_token_accuracy': '0.6817', 'epoch': '0.1103'}
+{'loss': '1.559', 'grad_norm': '12.62', 'learning_rate': '8.897e-06', 'num_tokens': '4.128e+05', 'mean_token_accuracy': '0.5947', 'epoch': '0.1105'}
+{'loss': '2.23', 'grad_norm': '32.75', 'learning_rate': '8.895e-06', 'num_tokens': '4.129e+05', 'mean_token_accuracy': '0.656', 'epoch': '0.1106'}
+{'loss': '0.8161', 'grad_norm': '12.81', 'learning_rate': '8.894e-06', 'num_tokens': '4.134e+05', 'mean_token_accuracy': '0.787', 'epoch': '0.1108'}
+{'loss': '1.31', 'grad_norm': '17.75', 'learning_rate': '8.892e-06', 'num_tokens': '4.141e+05', 'mean_token_accuracy': '0.6971', 'epoch': '0.111'}
+{'loss': '1.982', 'grad_norm': '12', 'learning_rate': '8.89e-06', 'num_tokens': '4.148e+05', 'mean_token_accuracy': '0.6218', 'epoch': '0.1111'}
+{'loss': '0.8254', 'grad_norm': '8.75', 'learning_rate': '8.889e-06', 'num_tokens': '4.16e+05', 'mean_token_accuracy': '0.755', 'epoch': '0.1113'}
+{'loss': '1.331', 'grad_norm': '15.56', 'learning_rate': '8.887e-06', 'num_tokens': '4.164e+05', 'mean_token_accuracy': '0.6722', 'epoch': '0.1115'}
+{'loss': '1.266', 'grad_norm': '13', 'learning_rate': '8.885e-06', 'num_tokens': '4.17e+05', 'mean_token_accuracy': '0.6493', 'epoch': '0.1117'}
+{'loss': '0.6029', 'grad_norm': '7.125', 'learning_rate': '8.883e-06', 'num_tokens': '4.187e+05', 'mean_token_accuracy': '0.8309', 'epoch': '0.1118'}
+{'loss': '0.9209', 'grad_norm': '28.25', 'learning_rate': '8.882e-06', 'num_tokens': '4.188e+05', 'mean_token_accuracy': '0.8204', 'epoch': '0.112'}
+{'loss': '1.311', 'grad_norm': '10.69', 'learning_rate': '8.88e-06', 'num_tokens': '4.2e+05', 'mean_token_accuracy': '0.6356', 'epoch': '0.1122'}
+{'loss': '0.9198', 'grad_norm': '25.75', 'learning_rate': '8.878e-06', 'num_tokens': '4.201e+05', 'mean_token_accuracy': '0.7879', 'epoch': '0.1124'}
+{'loss': '1.822', 'grad_norm': '11.69', 'learning_rate': '8.876e-06', 'num_tokens': '4.208e+05', 'mean_token_accuracy': '0.6308', 'epoch': '0.1125'}
+{'loss': '1.182', 'grad_norm': '18', 'learning_rate': '8.875e-06', 'num_tokens': '4.211e+05', 'mean_token_accuracy': '0.7445', 'epoch': '0.1127'}
+{'loss': '1.282', 'grad_norm': '11.75', 'learning_rate': '8.873e-06', 'num_tokens': '4.219e+05', 'mean_token_accuracy': '0.6477', 'epoch': '0.1129'}
+{'loss': '1.198', 'grad_norm': '14.06', 'learning_rate': '8.871e-06', 'num_tokens': '4.225e+05', 'mean_token_accuracy': '0.6765', 'epoch': '0.1131'}
+{'loss': '1.632', 'grad_norm': '19.88', 'learning_rate': '8.869e-06', 'num_tokens': '4.227e+05', 'mean_token_accuracy': '0.6723', 'epoch': '0.1132'}
+{'loss': '0.7267', 'grad_norm': '9', 'learning_rate': '8.868e-06', 'num_tokens': '4.24e+05', 'mean_token_accuracy': '0.7902', 'epoch': '0.1134'}
+{'loss': '0.6506', 'grad_norm': '7.562', 'learning_rate': '8.866e-06', 'num_tokens': '4.255e+05', 'mean_token_accuracy': '0.8043', 'epoch': '0.1136'}
+{'loss': '0.7292', 'grad_norm': '8.562', 'learning_rate': '8.864e-06', 'num_tokens': '4.267e+05', 'mean_token_accuracy': '0.7925', 'epoch': '0.1138'}
+{'loss': '1.166', 'grad_norm': '15.81', 'learning_rate': '8.862e-06', 'num_tokens': '4.271e+05', 'mean_token_accuracy': '0.6827', 'epoch': '0.1139'}
+{'loss': '1.02', 'grad_norm': '20.38', 'learning_rate': '8.861e-06', 'num_tokens': '4.273e+05', 'mean_token_accuracy': '0.7892', 'epoch': '0.1141'}
+{'loss': '2.399', 'grad_norm': '16', 'learning_rate': '8.859e-06', 'num_tokens': '4.279e+05', 'mean_token_accuracy': '0.5362', 'epoch': '0.1143'}
+{'loss': '0.7042', 'grad_norm': '7.25', 'learning_rate': '8.857e-06', 'num_tokens': '4.293e+05', 'mean_token_accuracy': '0.7902', 'epoch': '0.1144'}
+{'loss': '1.69', 'grad_norm': '19.62', 'learning_rate': '8.856e-06', 'num_tokens': '4.296e+05', 'mean_token_accuracy': '0.6572', 'epoch': '0.1146'}
+{'loss': '0.942', 'grad_norm': '21.38', 'learning_rate': '8.854e-06', 'num_tokens': '4.298e+05', 'mean_token_accuracy': '0.7989', 'epoch': '0.1148'}
+{'loss': '0.848', 'grad_norm': '19.88', 'learning_rate': '8.852e-06', 'num_tokens': '4.3e+05', 'mean_token_accuracy': '0.8068', 'epoch': '0.115'}
+{'loss': '1.982', 'grad_norm': '11.69', 'learning_rate': '8.85e-06', 'num_tokens': '4.307e+05', 'mean_token_accuracy': '0.6365', 'epoch': '0.1151'}
+{'loss': '2.06', 'grad_norm': '9.438', 'learning_rate': '8.849e-06', 'num_tokens': '4.322e+05', 'mean_token_accuracy': '0.5446', 'epoch': '0.1153'}
+
 12%|█▏        | 665/5758 [03:46<28:08,  3.02it/s]
 12%|█▏        | 666/5758 [03:46<28:04,  3.02it/s]
                                                  

 12%|█▏        | 666/5758 [03:46<28:04,  3.02it/s]
 12%|█▏        | 667/5758 [03:46<27:47,  3.05it/s]
                                                  

 12%|█▏        | 667/5758 [03:46<27:47,  3.05it/s]
 12%|█▏        | 668/5758 [03:47<26:55,  3.15it/s]
                                                  

 12%|█▏        | 668/5758 [03:47<26:55,  3.15it/s]
 12%|█▏        | 669/5758 [03:47<26:31,  3.20it/s]
                                                  

 12%|█▏        | 669/5758 [03:47<26:31,  3.20it/s]
 12%|█▏        | 670/5758 [03:47<28:16,  3.00it/s]
                                                  

 12%|█▏        | 670/5758 [03:47<28:16,  3.00it/s]
 12%|█▏        | 671/5758 [03:48<27:54,  3.04it/s]
                                                  

 12%|█▏        | 671/5758 [03:48<27:54,  3.04it/s]
 12%|█▏        | 672/5758 [03:48<29:58,  2.83it/s]
                                                  

 12%|█▏        | 672/5758 [03:48<29:58,  2.83it/s]
 12%|█▏        | 673/5758 [03:48<29:54,  2.83it/s]
                                                  

 12%|█▏        | 673/5758 [03:48<29:54,  2.83it/s]
 12%|█▏        | 674/5758 [03:49<29:06,  2.91it/s]
                                                  

 12%|█▏        | 674/5758 [03:49<29:06,  2.91it/s]
 12%|█▏        | 675/5758 [03:49<30:13,  2.80it/s]
                                                  

 12%|█▏        | 675/5758 [03:49<30:13,  2.80it/s]
 12%|█▏        | 676/5758 [03:49<28:55,  2.93it/s]
                                                  

 12%|█▏        | 676/5758 [03:49<28:55,  2.93it/s]
 12%|█▏        | 677/5758 [03:50<27:44,  3.05it/s]
                                                  

 12%|█▏        | 677/5758 [03:50<27:44,  3.05it/s]
 12%|█▏        | 678/5758 [03:50<26:54,  3.15it/s]
                                                  

 12%|█▏        | 678/5758 [03:50<26:54,  3.15it/s]
 12%|█▏        | 679/5758 [03:50<28:34,  2.96it/s]
                                                  

 12%|█▏        | 679/5758 [03:50<28:34,  2.96it/s]
 12%|█▏        | 680/5758 [03:51<30:34,  2.77it/s]
                                                  

 12%|█▏        | 680/5758 [03:51<30:34,  2.77it/s]
 12%|█▏        | 681/5758 [03:51<29:44,  2.85it/s]
                                                  

 12%|█▏        | 681/5758 [03:51<29:44,  2.85it/s]
 12%|█▏        | 682/5758 [03:52<29:46,  2.84it/s]
                                                  

 12%|█▏        | 682/5758 [03:52<29:46,  2.84it/s]
 12%|█▏        | 683/5758 [03:52<30:42,  2.75it/s]
                                                  

 12%|█▏        | 683/5758 [03:52<30:42,  2.75it/s]
 12%|█▏        | 684/5758 [03:52<29:14,  2.89it/s]
                                                  

 12%|█▏        | 684/5758 [03:52<29:14,  2.89it/s]
 12%|█▏        | 685/5758 [03:53<27:57,  3.02it/s]
                                                  

 12%|█▏        | 685/5758 [03:53<27:57,  3.02it/s]
 12%|█▏        | 686/5758 [03:53<27:27,  3.08it/s]
                                                  

 12%|█▏        | 686/5758 [03:53<27:27,  3.08it/s]
 12%|█▏        | 687/5758 [03:53<26:55,  3.14it/s]
                                                  

 12%|█▏        | 687/5758 [03:53<26:55,  3.14it/s]
 12%|█▏        | 688/5758 [03:53<27:58,  3.02it/s]
                                                  

 12%|█▏        | 688/5758 [03:54<27:58,  3.02it/s]
 12%|█▏        | 689/5758 [03:54<28:20,  2.98it/s]
                                                  

 12%|█▏        | 689/5758 [03:54<28:20,  2.98it/s]
 12%|█▏        | 690/5758 [03:54<27:16,  3.10it/s]
                                                  

 12%|█▏        | 690/5758 [03:54<27:16,  3.10it/s]
 12%|█▏        | 691/5758 [03:54<26:39,  3.17it/s]
                                                  

 12%|█▏        | 691/5758 [03:54<26:39,  3.17it/s]
 12%|█▏        | 692/5758 [03:55<26:17,  3.21it/s]
                                                  

 12%|█▏        | 692/5758 [03:55<26:17,  3.21it/s]
 12%|█▏        | 693/5758 [03:55<26:58,  3.13it/s]
                                                  

 12%|█▏        | 693/5758 [03:55<26:58,  3.13it/s]
 12%|█▏        | 694/5758 [03:55<26:17,  3.21it/s]
                                                  

 12%|█▏        | 694/5758 [03:55<26:17,  3.21it/s]
 12%|█▏        | 695/5758 [03:56<27:48,  3.03it/s]
                                                  

 12%|█▏        | 695/5758 [03:56<27:48,  3.03it/s]
 12%|█▏        | 696/5758 [03:56<27:22,  3.08it/s]
                                                  

 12%|█▏        | 696/5758 [03:56<27:22,  3.08it/s]
 12%|█▏        | 697/5758 [03:56<27:40,  3.05it/s]
                                                  

 12%|█▏        | 697/5758 [03:56<27:40,  3.05it/s]
 12%|█▏        | 698/5758 [03:57<31:07,  2.71it/s]
                                                  

 12%|█▏        | 698/5758 [03:57<31:07,  2.71it/s]
 12%|█▏        | 699/5758 [03:57<30:45,  2.74it/s]
                                                  

 12%|█▏        | 699/5758 [03:57<30:45,  2.74it/s]
 12%|█▏        | 700/5758 [03:58<30:13,  2.79it/s]
                                                  

 12%|█▏        | 700/5758 [03:58<30:13,  2.79it/s]
 12%|█▏        | 701/5758 [03:58<30:28,  2.77it/s]
                                                  

 12%|█▏        | 701/5758 [03:58<30:28,  2.77it/s]
 12%|█▏        | 702/5758 [03:58<30:24,  2.77it/s]
                                                  

 12%|█▏        | 702/5758 [03:58<30:24,  2.77it/s]
 12%|█▏        | 703/5758 [03:59<30:58,  2.72it/s]
                                                  

 12%|█▏        | 703/5758 [03:59<30:58,  2.72it/s]
 12%|█▏        | 704/5758 [03:59<30:50,  2.73it/s]
                                                  

 12%|█▏        | 704/5758 [03:59<30:50,  2.73it/s]
 12%|█▏        | 705/5758 [03:59<29:24,  2.86it/s]
                                                  

 12%|█▏        | 705/5758 [03:59<29:24,  2.86it/s]
 12%|█▏        | 706/5758 [04:00<27:59,  3.01it/s]
                                                  

 12%|█▏        | 706/5758 [04:00<27:59,  3.01it/s]
 12%|█▏        | 707/5758 [04:00<27:59,  3.01it/s]
                                                  

 12%|█▏        | 707/5758 [04:00<27:59,  3.01it/s]
 12%|█▏        | 708/5758 [04:00<27:51,  3.02it/s]
                                                  

 12%|█▏        | 708/5758 [04:00<27:51,  3.02it/s]
 12%|█▏        | 709/5758 [04:01<28:53,  2.91it/s]
                                                  

 12%|█▏        | 709/5758 [04:01<28:53,  2.91it/s]
 12%|█▏        | 710/5758 [04:01<28:36,  2.94it/s]
                                                  

 12%|█▏        | 710/5758 [04:01<28:36,  2.94it/s]
 12%|█▏        | 711/5758 [04:01<28:03,  3.00it/s]
                                                  

 12%|█▏        | 711/5758 [04:01<28:03,  3.00it/s]
 12%|█▏        | 712/5758 [04:02<28:09,  2.99it/s]
                                                  

 12%|█▏        | 712/5758 [04:02<28:09,  2.99it/s]
 12%|█▏        | 713/5758 [04:02<27:31,  3.05it/s]
                                                  

 12%|█▏        | 713/5758 [04:02<27:31,  3.05it/s]
 12%|█▏        | 714/5758 [04:02<27:50,  3.02it/s]
                                                  

 12%|█▏        | 714/5758 [04:02<27:50,  3.02it/s]
 12%|█▏        | 715/5758 [04:03<27:53,  3.01it/s]
                                                  

 12%|█▏        | 715/5758 [04:03<27:53,  3.01it/s]
 12%|█▏        | 716/5758 [04:03<29:19,  2.86it/s]
                                                  

 12%|█▏        | 716/5758 [04:03<29:19,  2.86it/s]
 12%|█▏        | 717/5758 [04:03<28:39,  2.93it/s]
                                                  

 12%|█▏        | 717/5758 [04:03<28:39,  2.93it/s]
 12%|█▏        | 718/5758 [04:04<27:33,  3.05it/s]
                                                  

 12%|█▏        | 718/5758 [04:04<27:33,  3.05it/s]
 12%|█▏        | 719/5758 [04:04<28:19,  2.96it/s]
                                                  

 12%|█▏        | 719/5758 [04:04<28:19,  2.96it/s]
 13%|█▎        | 720/5758 [04:04<27:47,  3.02it/s]
                                                  
{'loss': '2.371', 'grad_norm': '17.12', 'learning_rate': '8.847e-06', 'num_tokens': '4.327e+05', 'mean_token_accuracy': '0.5461', 'epoch': '0.1155'}
+{'loss': '1.385', 'grad_norm': '12.38', 'learning_rate': '8.845e-06', 'num_tokens': '4.336e+05', 'mean_token_accuracy': '0.6275', 'epoch': '0.1157'}
+{'loss': '1.014', 'grad_norm': '14.5', 'learning_rate': '8.843e-06', 'num_tokens': '4.34e+05', 'mean_token_accuracy': '0.7825', 'epoch': '0.1158'}
+{'loss': '1.276', 'grad_norm': '12.94', 'learning_rate': '8.842e-06', 'num_tokens': '4.346e+05', 'mean_token_accuracy': '0.6634', 'epoch': '0.116'}
+{'loss': '1.846', 'grad_norm': '59.25', 'learning_rate': '8.84e-06', 'num_tokens': '4.346e+05', 'mean_token_accuracy': '0.726', 'epoch': '0.1162'}
+{'loss': '1.14', 'grad_norm': '13.69', 'learning_rate': '8.838e-06', 'num_tokens': '4.352e+05', 'mean_token_accuracy': '0.6714', 'epoch': '0.1164'}
+{'loss': '1.073', 'grad_norm': '12.88', 'learning_rate': '8.836e-06', 'num_tokens': '4.36e+05', 'mean_token_accuracy': '0.686', 'epoch': '0.1165'}
+{'loss': '2.207', 'grad_norm': '9.25', 'learning_rate': '8.835e-06', 'num_tokens': '4.372e+05', 'mean_token_accuracy': '0.5707', 'epoch': '0.1167'}
+{'loss': '2.262', 'grad_norm': '24.25', 'learning_rate': '8.833e-06', 'num_tokens': '4.375e+05', 'mean_token_accuracy': '0.5256', 'epoch': '0.1169'}
+{'loss': '0.5978', 'grad_norm': '7', 'learning_rate': '8.831e-06', 'num_tokens': '4.389e+05', 'mean_token_accuracy': '0.8182', 'epoch': '0.1171'}
+{'loss': '2.449', 'grad_norm': '13.56', 'learning_rate': '8.829e-06', 'num_tokens': '4.395e+05', 'mean_token_accuracy': '0.5494', 'epoch': '0.1172'}
+{'loss': '1.158', 'grad_norm': '34.25', 'learning_rate': '8.828e-06', 'num_tokens': '4.396e+05', 'mean_token_accuracy': '0.72', 'epoch': '0.1174'}
+{'loss': '0.7102', 'grad_norm': '7.938', 'learning_rate': '8.826e-06', 'num_tokens': '4.408e+05', 'mean_token_accuracy': '0.7989', 'epoch': '0.1176'}
+{'loss': '1.406', 'grad_norm': '13.62', 'learning_rate': '8.824e-06', 'num_tokens': '4.414e+05', 'mean_token_accuracy': '0.624', 'epoch': '0.1177'}
+{'loss': '1.776', 'grad_norm': '29.88', 'learning_rate': '8.823e-06', 'num_tokens': '4.416e+05', 'mean_token_accuracy': '0.6614', 'epoch': '0.1179'}
+{'loss': '2.097', 'grad_norm': '31.38', 'learning_rate': '8.821e-06', 'num_tokens': '4.417e+05', 'mean_token_accuracy': '0.585', 'epoch': '0.1181'}
+{'loss': '1.227', 'grad_norm': '11.31', 'learning_rate': '8.819e-06', 'num_tokens': '4.425e+05', 'mean_token_accuracy': '0.6561', 'epoch': '0.1183'}
+{'loss': '1.446', 'grad_norm': '19.38', 'learning_rate': '8.817e-06', 'num_tokens': '4.428e+05', 'mean_token_accuracy': '0.7227', 'epoch': '0.1184'}
+{'loss': '1.517', 'grad_norm': '15', 'learning_rate': '8.816e-06', 'num_tokens': '4.433e+05', 'mean_token_accuracy': '0.6254', 'epoch': '0.1186'}
+{'loss': '1.819', 'grad_norm': '16.75', 'learning_rate': '8.814e-06', 'num_tokens': '4.437e+05', 'mean_token_accuracy': '0.6303', 'epoch': '0.1188'}
+{'loss': '0.995', 'grad_norm': '18.5', 'learning_rate': '8.812e-06', 'num_tokens': '4.44e+05', 'mean_token_accuracy': '0.7819', 'epoch': '0.119'}
+{'loss': '1.518', 'grad_norm': '14.06', 'learning_rate': '8.81e-06', 'num_tokens': '4.447e+05', 'mean_token_accuracy': '0.6291', 'epoch': '0.1191'}
+{'loss': '0.5199', 'grad_norm': '6.812', 'learning_rate': '8.809e-06', 'num_tokens': '4.46e+05', 'mean_token_accuracy': '0.8478', 'epoch': '0.1193'}
+{'loss': '1.343', 'grad_norm': '11.06', 'learning_rate': '8.807e-06', 'num_tokens': '4.47e+05', 'mean_token_accuracy': '0.6994', 'epoch': '0.1195'}
+{'loss': '0.9861', 'grad_norm': '14.12', 'learning_rate': '8.805e-06', 'num_tokens': '4.474e+05', 'mean_token_accuracy': '0.7692', 'epoch': '0.1197'}
+{'loss': '1.006', 'grad_norm': '11.62', 'learning_rate': '8.803e-06', 'num_tokens': '4.481e+05', 'mean_token_accuracy': '0.7789', 'epoch': '0.1198'}
+{'loss': '1.14', 'grad_norm': '23.88', 'learning_rate': '8.802e-06', 'num_tokens': '4.482e+05', 'mean_token_accuracy': '0.7544', 'epoch': '0.12'}
+{'loss': '1.38', 'grad_norm': '18.38', 'learning_rate': '8.8e-06', 'num_tokens': '4.486e+05', 'mean_token_accuracy': '0.727', 'epoch': '0.1202'}
+{'loss': '0.6179', 'grad_norm': '30.5', 'learning_rate': '8.798e-06', 'num_tokens': '4.487e+05', 'mean_token_accuracy': '0.8469', 'epoch': '0.1204'}
+{'loss': '2.337', 'grad_norm': '9.625', 'learning_rate': '8.796e-06', 'num_tokens': '4.499e+05', 'mean_token_accuracy': '0.5368', 'epoch': '0.1205'}
+{'loss': '2.045', 'grad_norm': '10.12', 'learning_rate': '8.795e-06', 'num_tokens': '4.51e+05', 'mean_token_accuracy': '0.5976', 'epoch': '0.1207'}
+{'loss': '1.738', 'grad_norm': '14.44', 'learning_rate': '8.793e-06', 'num_tokens': '4.517e+05', 'mean_token_accuracy': '0.6416', 'epoch': '0.1209'}
+{'loss': '1.371', 'grad_norm': '11.31', 'learning_rate': '8.791e-06', 'num_tokens': '4.526e+05', 'mean_token_accuracy': '0.6354', 'epoch': '0.121'}
+{'loss': '1.194', 'grad_norm': '16', 'learning_rate': '8.79e-06', 'num_tokens': '4.53e+05', 'mean_token_accuracy': '0.6967', 'epoch': '0.1212'}
+{'loss': '1.612', 'grad_norm': '17', 'learning_rate': '8.788e-06', 'num_tokens': '4.535e+05', 'mean_token_accuracy': '0.6266', 'epoch': '0.1214'}
+{'loss': '1.331', 'grad_norm': '10.94', 'learning_rate': '8.786e-06', 'num_tokens': '4.544e+05', 'mean_token_accuracy': '0.651', 'epoch': '0.1216'}
+{'loss': '1.639', 'grad_norm': '21.88', 'learning_rate': '8.784e-06', 'num_tokens': '4.547e+05', 'mean_token_accuracy': '0.6448', 'epoch': '0.1217'}
+{'loss': '2.027', 'grad_norm': '17.38', 'learning_rate': '8.783e-06', 'num_tokens': '4.552e+05', 'mean_token_accuracy': '0.6089', 'epoch': '0.1219'}
+{'loss': '2.103', 'grad_norm': '10', 'learning_rate': '8.781e-06', 'num_tokens': '4.563e+05', 'mean_token_accuracy': '0.585', 'epoch': '0.1221'}
+{'loss': '1.785', 'grad_norm': '9.188', 'learning_rate': '8.779e-06', 'num_tokens': '4.584e+05', 'mean_token_accuracy': '0.6224', 'epoch': '0.1223'}
+{'loss': '2.064', 'grad_norm': '15.75', 'learning_rate': '8.777e-06', 'num_tokens': '4.591e+05', 'mean_token_accuracy': '0.5807', 'epoch': '0.1224'}
+{'loss': '2.261', 'grad_norm': '16.5', 'learning_rate': '8.776e-06', 'num_tokens': '4.598e+05', 'mean_token_accuracy': '0.5557', 'epoch': '0.1226'}
+{'loss': '1.006', 'grad_norm': '21', 'learning_rate': '8.774e-06', 'num_tokens': '4.6e+05', 'mean_token_accuracy': '0.8028', 'epoch': '0.1228'}
+{'loss': '2.572', 'grad_norm': '22.88', 'learning_rate': '8.772e-06', 'num_tokens': '4.602e+05', 'mean_token_accuracy': '0.5163', 'epoch': '0.123'}
+{'loss': '1.108', 'grad_norm': '9.312', 'learning_rate': '8.77e-06', 'num_tokens': '4.613e+05', 'mean_token_accuracy': '0.7562', 'epoch': '0.1231'}
+{'loss': '1.411', 'grad_norm': '12.81', 'learning_rate': '8.769e-06', 'num_tokens': '4.621e+05', 'mean_token_accuracy': '0.6162', 'epoch': '0.1233'}
+{'loss': '1.215', 'grad_norm': '11.31', 'learning_rate': '8.767e-06', 'num_tokens': '4.628e+05', 'mean_token_accuracy': '0.6703', 'epoch': '0.1235'}
+{'loss': '1.742', 'grad_norm': '25.75', 'learning_rate': '8.765e-06', 'num_tokens': '4.63e+05', 'mean_token_accuracy': '0.5972', 'epoch': '0.1237'}
+{'loss': '1.537', 'grad_norm': '12.75', 'learning_rate': '8.763e-06', 'num_tokens': '4.638e+05', 'mean_token_accuracy': '0.6653', 'epoch': '0.1238'}
+{'loss': '1.626', 'grad_norm': '12.38', 'learning_rate': '8.762e-06', 'num_tokens': '4.646e+05', 'mean_token_accuracy': '0.5781', 'epoch': '0.124'}
+{'loss': '1.371', 'grad_norm': '11.06', 'learning_rate': '8.76e-06', 'num_tokens': '4.655e+05', 'mean_token_accuracy': '0.6352', 'epoch': '0.1242'}
+{'loss': '0.7226', 'grad_norm': '6.125', 'learning_rate': '8.758e-06', 'num_tokens': '4.678e+05', 'mean_token_accuracy': '0.7889', 'epoch': '0.1243'}
+{'loss': '0.7717', 'grad_norm': '7.062', 'learning_rate': '8.757e-06', 'num_tokens': '4.694e+05', 'mean_token_accuracy': '0.7673', 'epoch': '0.1245'}
+{'loss': '1.334', 'grad_norm': '12.19', 'learning_rate': '8.755e-06', 'num_tokens': '4.7e+05', 'mean_token_accuracy': '0.6504', 'epoch': '0.1247'}
+{'loss': '0.9046', 'grad_norm': '28.62', 'learning_rate': '8.753e-06', 'num_tokens': '4.701e+05', 'mean_token_accuracy': '0.8407', 'epoch': '0.1249'}
+
 13%|█▎        | 720/5758 [04:04<27:47,  3.02it/s]
 13%|█▎        | 721/5758 [04:05<27:46,  3.02it/s]
                                                  

 13%|█▎        | 721/5758 [04:05<27:46,  3.02it/s]
 13%|█▎        | 722/5758 [04:05<27:20,  3.07it/s]
                                                  

 13%|█▎        | 722/5758 [04:05<27:20,  3.07it/s]
 13%|█▎        | 723/5758 [04:05<26:35,  3.16it/s]
                                                  

 13%|█▎        | 723/5758 [04:05<26:35,  3.16it/s]
 13%|█▎        | 724/5758 [04:06<29:52,  2.81it/s]
                                                  

 13%|█▎        | 724/5758 [04:06<29:52,  2.81it/s]
 13%|█▎        | 725/5758 [04:06<29:10,  2.88it/s]
                                                  

 13%|█▎        | 725/5758 [04:06<29:10,  2.88it/s]
 13%|█▎        | 726/5758 [04:06<28:21,  2.96it/s]
                                                  

 13%|█▎        | 726/5758 [04:06<28:21,  2.96it/s]
 13%|█▎        | 727/5758 [04:07<28:41,  2.92it/s]
                                                  

 13%|█▎        | 727/5758 [04:07<28:41,  2.92it/s]
 13%|█▎        | 728/5758 [04:07<28:32,  2.94it/s]
                                                  

 13%|█▎        | 728/5758 [04:07<28:32,  2.94it/s]
 13%|█▎        | 729/5758 [04:07<27:25,  3.06it/s]
                                                  

 13%|█▎        | 729/5758 [04:07<27:25,  3.06it/s]
 13%|█▎        | 730/5758 [04:08<26:38,  3.15it/s]
                                                  

 13%|█▎        | 730/5758 [04:08<26:38,  3.15it/s]
 13%|█▎        | 731/5758 [04:08<28:04,  2.98it/s]
                                                  

 13%|█▎        | 731/5758 [04:08<28:04,  2.98it/s]
 13%|█▎        | 732/5758 [04:08<27:56,  3.00it/s]
                                                  

 13%|█▎        | 732/5758 [04:08<27:56,  3.00it/s]
 13%|█▎        | 733/5758 [04:09<29:31,  2.84it/s]
                                                  

 13%|█▎        | 733/5758 [04:09<29:31,  2.84it/s]
 13%|█▎        | 734/5758 [04:09<29:47,  2.81it/s]
                                                  

 13%|█▎        | 734/5758 [04:09<29:47,  2.81it/s]
 13%|█▎        | 735/5758 [04:09<28:48,  2.91it/s]
                                                  

 13%|█▎        | 735/5758 [04:09<28:48,  2.91it/s]
 13%|█▎        | 736/5758 [04:10<27:54,  3.00it/s]
                                                  

 13%|█▎        | 736/5758 [04:10<27:54,  3.00it/s]
 13%|█▎        | 737/5758 [04:10<26:42,  3.13it/s]
                                                  

 13%|█▎        | 737/5758 [04:10<26:42,  3.13it/s]
 13%|█▎        | 738/5758 [04:10<26:42,  3.13it/s]
                                                  

 13%|█▎        | 738/5758 [04:10<26:42,  3.13it/s]
 13%|█▎        | 739/5758 [04:11<26:23,  3.17it/s]
                                                  

 13%|█▎        | 739/5758 [04:11<26:23,  3.17it/s]
 13%|█▎        | 740/5758 [04:11<26:25,  3.17it/s]
                                                  

 13%|█▎        | 740/5758 [04:11<26:25,  3.17it/s]
 13%|█▎        | 741/5758 [04:11<26:29,  3.16it/s]
                                                  

 13%|█▎        | 741/5758 [04:11<26:29,  3.16it/s]
 13%|█▎        | 742/5758 [04:12<25:54,  3.23it/s]
                                                  

 13%|█▎        | 742/5758 [04:12<25:54,  3.23it/s]
 13%|█▎        | 743/5758 [04:12<26:11,  3.19it/s]
                                                  

 13%|█▎        | 743/5758 [04:12<26:11,  3.19it/s]
 13%|█▎        | 744/5758 [04:12<25:41,  3.25it/s]
                                                  

 13%|█▎        | 744/5758 [04:12<25:41,  3.25it/s]
 13%|█▎        | 745/5758 [04:13<26:16,  3.18it/s]
                                                  

 13%|█▎        | 745/5758 [04:13<26:16,  3.18it/s]
 13%|█▎        | 746/5758 [04:13<26:26,  3.16it/s]
                                                  

 13%|█▎        | 746/5758 [04:13<26:26,  3.16it/s]
 13%|█▎        | 747/5758 [04:13<26:42,  3.13it/s]
                                                  

 13%|█▎        | 747/5758 [04:13<26:42,  3.13it/s]
 13%|█▎        | 748/5758 [04:13<25:57,  3.22it/s]
                                                  

 13%|█▎        | 748/5758 [04:13<25:57,  3.22it/s]
 13%|█▎        | 749/5758 [04:14<27:05,  3.08it/s]
                                                  

 13%|█▎        | 749/5758 [04:14<27:05,  3.08it/s]
 13%|█▎        | 750/5758 [04:14<26:21,  3.17it/s]
                                                  

 13%|█▎        | 750/5758 [04:14<26:21,  3.17it/s]
 13%|█▎        | 751/5758 [04:15<29:04,  2.87it/s]
                                                  

 13%|█▎        | 751/5758 [04:15<29:04,  2.87it/s]
 13%|█▎        | 752/5758 [04:15<27:34,  3.03it/s]
                                                  

 13%|█▎        | 752/5758 [04:15<27:34,  3.03it/s]
 13%|█▎        | 753/5758 [04:15<27:08,  3.07it/s]
                                                  

 13%|█▎        | 753/5758 [04:15<27:08,  3.07it/s]
 13%|█▎        | 754/5758 [04:15<27:05,  3.08it/s]
                                                  

 13%|█▎        | 754/5758 [04:15<27:05,  3.08it/s]
 13%|█▎        | 755/5758 [04:16<27:34,  3.02it/s]
                                                  

 13%|█▎        | 755/5758 [04:16<27:34,  3.02it/s]
 13%|█▎        | 756/5758 [04:16<26:46,  3.11it/s]
                                                  

 13%|█▎        | 756/5758 [04:16<26:46,  3.11it/s]
 13%|█▎        | 757/5758 [04:16<27:57,  2.98it/s]
                                                  

 13%|█▎        | 757/5758 [04:16<27:57,  2.98it/s]
 13%|█▎        | 758/5758 [04:17<27:33,  3.02it/s]
                                                  

 13%|█▎        | 758/5758 [04:17<27:33,  3.02it/s]
 13%|█▎        | 759/5758 [04:17<28:22,  2.94it/s]
                                                  

 13%|█▎        | 759/5758 [04:17<28:22,  2.94it/s]
 13%|█▎        | 760/5758 [04:17<27:28,  3.03it/s]
                                                  

 13%|█▎        | 760/5758 [04:17<27:28,  3.03it/s]
 13%|█▎        | 761/5758 [04:18<28:38,  2.91it/s]
                                                  

 13%|█▎        | 761/5758 [04:18<28:38,  2.91it/s]
 13%|█▎        | 762/5758 [04:18<29:42,  2.80it/s]
                                                  

 13%|█▎        | 762/5758 [04:18<29:42,  2.80it/s]
 13%|█▎        | 763/5758 [04:19<28:31,  2.92it/s]
                                                  

 13%|█▎        | 763/5758 [04:19<28:31,  2.92it/s]
 13%|█▎        | 764/5758 [04:19<29:43,  2.80it/s]
                                                  

 13%|█▎        | 764/5758 [04:19<29:43,  2.80it/s]
 13%|█▎        | 765/5758 [04:19<31:25,  2.65it/s]
                                                  

 13%|█▎        | 765/5758 [04:19<31:25,  2.65it/s]
 13%|█▎        | 766/5758 [04:20<29:41,  2.80it/s]
                                                  

 13%|█▎        | 766/5758 [04:20<29:41,  2.80it/s]
 13%|█▎        | 767/5758 [04:20<28:24,  2.93it/s]
                                                  

 13%|█▎        | 767/5758 [04:20<28:24,  2.93it/s]
 13%|█▎        | 768/5758 [04:20<27:29,  3.03it/s]
                                                  

 13%|█▎        | 768/5758 [04:20<27:29,  3.03it/s]
 13%|█▎        | 769/5758 [04:21<27:04,  3.07it/s]
                                                  

 13%|█▎        | 769/5758 [04:21<27:04,  3.07it/s]
 13%|█▎        | 770/5758 [04:21<28:08,  2.95it/s]
                                                  

 13%|█▎        | 770/5758 [04:21<28:08,  2.95it/s]
 13%|█▎        | 771/5758 [04:21<29:03,  2.86it/s]
                                                  

 13%|█▎        | 771/5758 [04:21<29:03,  2.86it/s]
 13%|█▎        | 772/5758 [04:22<27:04,  3.07it/s]
                                                  

 13%|█▎        | 772/5758 [04:22<27:04,  3.07it/s]
 13%|█▎        | 773/5758 [04:22<28:16,  2.94it/s]
                                                  

 13%|█▎        | 773/5758 [04:22<28:16,  2.94it/s]
 13%|█▎        | 774/5758 [04:22<27:54,  2.98it/s]
                                                  

 13%|█▎        | 774/5758 [04:22<27:54,  2.98it/s]
 13%|█▎        | 775/5758 [04:23<27:17,  3.04it/s]
                                                  
{'loss': '1.704', 'grad_norm': '6.656', 'learning_rate': '8.751e-06', 'num_tokens': '4.731e+05', 'mean_token_accuracy': '0.6235', 'epoch': '0.125'}
+{'loss': '1.255', 'grad_norm': '12.06', 'learning_rate': '8.75e-06', 'num_tokens': '4.74e+05', 'mean_token_accuracy': '0.6504', 'epoch': '0.1252'}
+{'loss': '1.303', 'grad_norm': '12.12', 'learning_rate': '8.748e-06', 'num_tokens': '4.747e+05', 'mean_token_accuracy': '0.6466', 'epoch': '0.1254'}
+{'loss': '1.051', 'grad_norm': '12.5', 'learning_rate': '8.746e-06', 'num_tokens': '4.753e+05', 'mean_token_accuracy': '0.694', 'epoch': '0.1256'}
+{'loss': '1.982', 'grad_norm': '7.281', 'learning_rate': '8.744e-06', 'num_tokens': '4.77e+05', 'mean_token_accuracy': '0.591', 'epoch': '0.1257'}
+{'loss': '1.144', 'grad_norm': '10.56', 'learning_rate': '8.743e-06', 'num_tokens': '4.779e+05', 'mean_token_accuracy': '0.6718', 'epoch': '0.1259'}
+{'loss': '1.361', 'grad_norm': '12', 'learning_rate': '8.741e-06', 'num_tokens': '4.786e+05', 'mean_token_accuracy': '0.6337', 'epoch': '0.1261'}
+{'loss': '1.878', 'grad_norm': '14.81', 'learning_rate': '8.739e-06', 'num_tokens': '4.791e+05', 'mean_token_accuracy': '0.6125', 'epoch': '0.1263'}
+{'loss': '1.845', 'grad_norm': '18.62', 'learning_rate': '8.737e-06', 'num_tokens': '4.795e+05', 'mean_token_accuracy': '0.5835', 'epoch': '0.1264'}
+{'loss': '2.041', 'grad_norm': '12', 'learning_rate': '8.736e-06', 'num_tokens': '4.802e+05', 'mean_token_accuracy': '0.5885', 'epoch': '0.1266'}
+{'loss': '1.148', 'grad_norm': '12.38', 'learning_rate': '8.734e-06', 'num_tokens': '4.808e+05', 'mean_token_accuracy': '0.6718', 'epoch': '0.1268'}
+{'loss': '1.639', 'grad_norm': '19.25', 'learning_rate': '8.732e-06', 'num_tokens': '4.811e+05', 'mean_token_accuracy': '0.6557', 'epoch': '0.127'}
+{'loss': '1.406', 'grad_norm': '15.69', 'learning_rate': '8.73e-06', 'num_tokens': '4.815e+05', 'mean_token_accuracy': '0.6894', 'epoch': '0.1271'}
+{'loss': '2.023', 'grad_norm': '31', 'learning_rate': '8.729e-06', 'num_tokens': '4.816e+05', 'mean_token_accuracy': '0.6241', 'epoch': '0.1273'}
+{'loss': '1.202', 'grad_norm': '29.25', 'learning_rate': '8.727e-06', 'num_tokens': '4.817e+05', 'mean_token_accuracy': '0.7949', 'epoch': '0.1275'}
+{'loss': '1.439', 'grad_norm': '22.12', 'learning_rate': '8.725e-06', 'num_tokens': '4.819e+05', 'mean_token_accuracy': '0.7047', 'epoch': '0.1276'}
+{'loss': '1.422', 'grad_norm': '12.5', 'learning_rate': '8.724e-06', 'num_tokens': '4.826e+05', 'mean_token_accuracy': '0.6368', 'epoch': '0.1278'}
+{'loss': '0.916', 'grad_norm': '27.62', 'learning_rate': '8.722e-06', 'num_tokens': '4.828e+05', 'mean_token_accuracy': '0.8278', 'epoch': '0.128'}
+{'loss': '2.106', 'grad_norm': '15.31', 'learning_rate': '8.72e-06', 'num_tokens': '4.832e+05', 'mean_token_accuracy': '0.5711', 'epoch': '0.1282'}
+{'loss': '0.7305', 'grad_norm': '8.75', 'learning_rate': '8.718e-06', 'num_tokens': '4.846e+05', 'mean_token_accuracy': '0.777', 'epoch': '0.1283'}
+{'loss': '1.115', 'grad_norm': '22.5', 'learning_rate': '8.717e-06', 'num_tokens': '4.848e+05', 'mean_token_accuracy': '0.8045', 'epoch': '0.1285'}
+{'loss': '1.779', 'grad_norm': '51.25', 'learning_rate': '8.715e-06', 'num_tokens': '4.849e+05', 'mean_token_accuracy': '0.7059', 'epoch': '0.1287'}
+{'loss': '0.6468', 'grad_norm': '8.562', 'learning_rate': '8.713e-06', 'num_tokens': '4.861e+05', 'mean_token_accuracy': '0.8067', 'epoch': '0.1289'}
+{'loss': '0.9894', 'grad_norm': '7.656', 'learning_rate': '8.711e-06', 'num_tokens': '4.876e+05', 'mean_token_accuracy': '0.7468', 'epoch': '0.129'}
+{'loss': '1.968', 'grad_norm': '9.188', 'learning_rate': '8.71e-06', 'num_tokens': '4.889e+05', 'mean_token_accuracy': '0.604', 'epoch': '0.1292'}
+{'loss': '2.571', 'grad_norm': '26.12', 'learning_rate': '8.708e-06', 'num_tokens': '4.891e+05', 'mean_token_accuracy': '0.4905', 'epoch': '0.1294'}
+{'loss': '1.318', 'grad_norm': '12.44', 'learning_rate': '8.706e-06', 'num_tokens': '4.899e+05', 'mean_token_accuracy': '0.6267', 'epoch': '0.1296'}
+{'loss': '1.38', 'grad_norm': '10.69', 'learning_rate': '8.704e-06', 'num_tokens': '4.907e+05', 'mean_token_accuracy': '0.679', 'epoch': '0.1297'}
+{'loss': '1.328', 'grad_norm': '25', 'learning_rate': '8.703e-06', 'num_tokens': '4.908e+05', 'mean_token_accuracy': '0.7273', 'epoch': '0.1299'}
+{'loss': '2.152', 'grad_norm': '17.88', 'learning_rate': '8.701e-06', 'num_tokens': '4.913e+05', 'mean_token_accuracy': '0.5817', 'epoch': '0.1301'}
+{'loss': '1.244', 'grad_norm': '13.38', 'learning_rate': '8.699e-06', 'num_tokens': '4.92e+05', 'mean_token_accuracy': '0.6422', 'epoch': '0.1303'}
+{'loss': '1.251', 'grad_norm': '12.25', 'learning_rate': '8.697e-06', 'num_tokens': '4.926e+05', 'mean_token_accuracy': '0.6429', 'epoch': '0.1304'}
+{'loss': '1.034', 'grad_norm': '41.25', 'learning_rate': '8.696e-06', 'num_tokens': '4.927e+05', 'mean_token_accuracy': '0.7313', 'epoch': '0.1306'}
+{'loss': '1.977', 'grad_norm': '19.88', 'learning_rate': '8.694e-06', 'num_tokens': '4.931e+05', 'mean_token_accuracy': '0.6083', 'epoch': '0.1308'}
+{'loss': '0.9981', 'grad_norm': '10.75', 'learning_rate': '8.692e-06', 'num_tokens': '4.939e+05', 'mean_token_accuracy': '0.7096', 'epoch': '0.1309'}
+{'loss': '1.979', 'grad_norm': '11.75', 'learning_rate': '8.691e-06', 'num_tokens': '4.948e+05', 'mean_token_accuracy': '0.5811', 'epoch': '0.1311'}
+{'loss': '1.316', 'grad_norm': '13.38', 'learning_rate': '8.689e-06', 'num_tokens': '4.954e+05', 'mean_token_accuracy': '0.627', 'epoch': '0.1313'}
+{'loss': '1.282', 'grad_norm': '14.12', 'learning_rate': '8.687e-06', 'num_tokens': '4.959e+05', 'mean_token_accuracy': '0.639', 'epoch': '0.1315'}
+{'loss': '1.42', 'grad_norm': '12.06', 'learning_rate': '8.685e-06', 'num_tokens': '4.967e+05', 'mean_token_accuracy': '0.6292', 'epoch': '0.1316'}
+{'loss': '1.05', 'grad_norm': '24.62', 'learning_rate': '8.684e-06', 'num_tokens': '4.968e+05', 'mean_token_accuracy': '0.7203', 'epoch': '0.1318'}
+{'loss': '1.601', 'grad_norm': '40.75', 'learning_rate': '8.682e-06', 'num_tokens': '4.969e+05', 'mean_token_accuracy': '0.6579', 'epoch': '0.132'}
+{'loss': '1.048', 'grad_norm': '12.88', 'learning_rate': '8.68e-06', 'num_tokens': '4.974e+05', 'mean_token_accuracy': '0.7023', 'epoch': '0.1322'}
+{'loss': '2.138', 'grad_norm': '13.94', 'learning_rate': '8.678e-06', 'num_tokens': '4.98e+05', 'mean_token_accuracy': '0.5647', 'epoch': '0.1323'}
+{'loss': '0.696', 'grad_norm': '7.531', 'learning_rate': '8.677e-06', 'num_tokens': '4.994e+05', 'mean_token_accuracy': '0.8011', 'epoch': '0.1325'}
+{'loss': '1.38', 'grad_norm': '20.12', 'learning_rate': '8.675e-06', 'num_tokens': '4.997e+05', 'mean_token_accuracy': '0.7356', 'epoch': '0.1327'}
+{'loss': '1.182', 'grad_norm': '5.188', 'learning_rate': '8.673e-06', 'num_tokens': '5.032e+05', 'mean_token_accuracy': '0.7148', 'epoch': '0.1329'}
+{'loss': '1.505', 'grad_norm': '12.81', 'learning_rate': '8.671e-06', 'num_tokens': '5.039e+05', 'mean_token_accuracy': '0.6029', 'epoch': '0.133'}
+{'loss': '2.423', 'grad_norm': '12.88', 'learning_rate': '8.67e-06', 'num_tokens': '5.046e+05', 'mean_token_accuracy': '0.5192', 'epoch': '0.1332'}
+{'loss': '1.136', 'grad_norm': '11.5', 'learning_rate': '8.668e-06', 'num_tokens': '5.052e+05', 'mean_token_accuracy': '0.7072', 'epoch': '0.1334'}
+{'loss': '0.8522', 'grad_norm': '9.688', 'learning_rate': '8.666e-06', 'num_tokens': '5.06e+05', 'mean_token_accuracy': '0.7967', 'epoch': '0.1336'}
+{'loss': '1.922', 'grad_norm': '20.25', 'learning_rate': '8.664e-06', 'num_tokens': '5.062e+05', 'mean_token_accuracy': '0.6593', 'epoch': '0.1337'}
+{'loss': '2.278', 'grad_norm': '13', 'learning_rate': '8.663e-06', 'num_tokens': '5.068e+05', 'mean_token_accuracy': '0.5504', 'epoch': '0.1339'}
+{'loss': '1.216', 'grad_norm': '51.25', 'learning_rate': '8.661e-06', 'num_tokens': '5.069e+05', 'mean_token_accuracy': '0.6863', 'epoch': '0.1341'}
+{'loss': '1.244', 'grad_norm': '6.469', 'learning_rate': '8.659e-06', 'num_tokens': '5.09e+05', 'mean_token_accuracy': '0.7029', 'epoch': '0.1342'}
+{'loss': '2.373', 'grad_norm': '10.62', 'learning_rate': '8.658e-06', 'num_tokens': '5.098e+05', 'mean_token_accuracy': '0.5421', 'epoch': '0.1344'}
+
 13%|█▎        | 775/5758 [04:23<27:17,  3.04it/s]
 13%|█▎        | 776/5758 [04:23<28:13,  2.94it/s]
                                                  

 13%|█▎        | 776/5758 [04:23<28:13,  2.94it/s]
 13%|█▎        | 777/5758 [04:23<27:15,  3.05it/s]
                                                  

 13%|█▎        | 777/5758 [04:23<27:15,  3.05it/s]
 14%|█▎        | 778/5758 [04:24<31:11,  2.66it/s]
                                                  

 14%|█▎        | 778/5758 [04:24<31:11,  2.66it/s]
 14%|█▎        | 779/5758 [04:24<30:16,  2.74it/s]
                                                  

 14%|█▎        | 779/5758 [04:24<30:16,  2.74it/s]
 14%|█▎        | 780/5758 [04:24<30:22,  2.73it/s]
                                                  

 14%|█▎        | 780/5758 [04:24<30:22,  2.73it/s]
 14%|█▎        | 781/5758 [04:25<30:14,  2.74it/s]
                                                  

 14%|█▎        | 781/5758 [04:25<30:14,  2.74it/s]
 14%|█▎        | 782/5758 [04:25<28:55,  2.87it/s]
                                                  

 14%|█▎        | 782/5758 [04:25<28:55,  2.87it/s]
 14%|█▎        | 783/5758 [04:25<28:57,  2.86it/s]
                                                  

 14%|█▎        | 783/5758 [04:25<28:57,  2.86it/s]
 14%|█▎        | 784/5758 [04:26<29:52,  2.78it/s]
                                                  

 14%|█▎        | 784/5758 [04:26<29:52,  2.78it/s]
 14%|█▎        | 785/5758 [04:26<29:49,  2.78it/s]
                                                  

 14%|█▎        | 785/5758 [04:26<29:49,  2.78it/s]
 14%|█▎        | 786/5758 [04:27<28:26,  2.91it/s]
                                                  

 14%|█▎        | 786/5758 [04:27<28:26,  2.91it/s]
 14%|█▎        | 787/5758 [04:27<29:25,  2.82it/s]
                                                  

 14%|█▎        | 787/5758 [04:27<29:25,  2.82it/s]
 14%|█▎        | 788/5758 [04:27<29:45,  2.78it/s]
                                                  

 14%|█▎        | 788/5758 [04:27<29:45,  2.78it/s]
 14%|█▎        | 789/5758 [04:28<28:25,  2.91it/s]
                                                  

 14%|█▎        | 789/5758 [04:28<28:25,  2.91it/s]
 14%|█▎        | 790/5758 [04:28<27:28,  3.01it/s]
                                                  

 14%|█▎        | 790/5758 [04:28<27:28,  3.01it/s]
 14%|█▎        | 791/5758 [04:28<28:09,  2.94it/s]
                                                  

 14%|█▎        | 791/5758 [04:28<28:09,  2.94it/s]
 14%|█▍        | 792/5758 [04:29<28:52,  2.87it/s]
                                                  

 14%|█▍        | 792/5758 [04:29<28:52,  2.87it/s]
 14%|█▍        | 793/5758 [04:29<28:06,  2.94it/s]
                                                  

 14%|█▍        | 793/5758 [04:29<28:06,  2.94it/s]
 14%|█▍        | 794/5758 [04:29<27:29,  3.01it/s]
                                                  

 14%|█▍        | 794/5758 [04:29<27:29,  3.01it/s]
 14%|█▍        | 795/5758 [04:30<27:04,  3.05it/s]
                                                  

 14%|█▍        | 795/5758 [04:30<27:04,  3.05it/s]
 14%|█▍        | 796/5758 [04:30<26:27,  3.13it/s]
                                                  

 14%|█▍        | 796/5758 [04:30<26:27,  3.13it/s]
 14%|█▍        | 797/5758 [04:30<26:25,  3.13it/s]
                                                  

 14%|█▍        | 797/5758 [04:30<26:25,  3.13it/s]
 14%|█▍        | 798/5758 [04:31<27:40,  2.99it/s]
                                                  

 14%|█▍        | 798/5758 [04:31<27:40,  2.99it/s]
 14%|█▍        | 799/5758 [04:31<28:20,  2.92it/s]
                                                  

 14%|█▍        | 799/5758 [04:31<28:20,  2.92it/s]
 14%|█▍        | 800/5758 [04:31<27:28,  3.01it/s]
                                                  

 14%|█▍        | 800/5758 [04:31<27:28,  3.01it/s]
 14%|█▍        | 801/5758 [04:32<28:43,  2.88it/s]
                                                  

 14%|█▍        | 801/5758 [04:32<28:43,  2.88it/s]
 14%|█▍        | 802/5758 [04:32<27:51,  2.96it/s]
                                                  

 14%|█▍        | 802/5758 [04:32<27:51,  2.96it/s]
 14%|█▍        | 803/5758 [04:32<30:12,  2.73it/s]
                                                  

 14%|█▍        | 803/5758 [04:32<30:12,  2.73it/s]
 14%|█▍        | 804/5758 [04:33<29:25,  2.81it/s]
                                                  

 14%|█▍        | 804/5758 [04:33<29:25,  2.81it/s]
 14%|█▍        | 805/5758 [04:33<27:46,  2.97it/s]
                                                  

 14%|█▍        | 805/5758 [04:33<27:46,  2.97it/s]
 14%|���▍        | 806/5758 [04:33<28:29,  2.90it/s]
                                                  

 14%|█▍        | 806/5758 [04:33<28:29,  2.90it/s]
 14%|█▍        | 807/5758 [04:34<27:19,  3.02it/s]
                                                  

 14%|█▍        | 807/5758 [04:34<27:19,  3.02it/s]
 14%|█▍        | 808/5758 [04:34<28:41,  2.88it/s]
                                                  

 14%|█▍        | 808/5758 [04:34<28:41,  2.88it/s]
 14%|█▍        | 809/5758 [04:34<28:41,  2.88it/s]
                                                  

 14%|█▍        | 809/5758 [04:34<28:41,  2.88it/s]
 14%|█▍        | 810/5758 [04:35<29:14,  2.82it/s]
                                                  

 14%|█▍        | 810/5758 [04:35<29:14,  2.82it/s]
 14%|█▍        | 811/5758 [04:35<27:58,  2.95it/s]
                                                  

 14%|█▍        | 811/5758 [04:35<27:58,  2.95it/s]
 14%|█▍        | 812/5758 [04:35<27:25,  3.01it/s]
                                                  

 14%|█▍        | 812/5758 [04:35<27:25,  3.01it/s]
 14%|█▍        | 813/5758 [04:36<27:01,  3.05it/s]
                                                  

 14%|█▍        | 813/5758 [04:36<27:01,  3.05it/s]
 14%|█▍        | 814/5758 [04:36<27:06,  3.04it/s]
                                                  

 14%|█▍        | 814/5758 [04:36<27:06,  3.04it/s]
 14%|█▍        | 815/5758 [04:36<27:58,  2.95it/s]
                                                  

 14%|█▍        | 815/5758 [04:36<27:58,  2.95it/s]
 14%|█▍        | 816/5758 [04:37<27:00,  3.05it/s]
                                                  

 14%|█▍        | 816/5758 [04:37<27:00,  3.05it/s]
 14%|█▍        | 817/5758 [04:37<26:25,  3.12it/s]
                                                  

 14%|█▍        | 817/5758 [04:37<26:25,  3.12it/s]
 14%|█▍        | 818/5758 [04:37<27:30,  2.99it/s]
                                                  

 14%|█▍        | 818/5758 [04:37<27:30,  2.99it/s]
 14%|█▍        | 819/5758 [04:38<25:53,  3.18it/s]
                                                  

 14%|█▍        | 819/5758 [04:38<25:53,  3.18it/s]
 14%|█▍        | 820/5758 [04:38<24:57,  3.30it/s]
                                                  

 14%|█▍        | 820/5758 [04:38<24:57,  3.30it/s]
 14%|█▍        | 821/5758 [04:38<24:49,  3.31it/s]
                                                  

 14%|█▍        | 821/5758 [04:38<24:49,  3.31it/s]
 14%|█▍        | 822/5758 [04:39<25:37,  3.21it/s]
                                                  

 14%|█▍        | 822/5758 [04:39<25:37,  3.21it/s]
 14%|█▍        | 823/5758 [04:39<25:44,  3.19it/s]
                                                  

 14%|█▍        | 823/5758 [04:39<25:44,  3.19it/s]
 14%|█▍        | 824/5758 [04:39<25:35,  3.21it/s]
                                                  

 14%|█▍        | 824/5758 [04:39<25:35,  3.21it/s]
 14%|█▍        | 825/5758 [04:40<26:23,  3.11it/s]
                                                  

 14%|█▍        | 825/5758 [04:40<26:23,  3.11it/s]
 14%|█▍        | 826/5758 [04:40<26:05,  3.15it/s]
                                                  

 14%|█▍        | 826/5758 [04:40<26:05,  3.15it/s]
 14%|█▍        | 827/5758 [04:40<26:26,  3.11it/s]
                                                  

 14%|█▍        | 827/5758 [04:40<26:26,  3.11it/s]
 14%|█▍        | 828/5758 [04:40<26:10,  3.14it/s]
                                                  

 14%|█▍        | 828/5758 [04:40<26:10,  3.14it/s]
 14%|█▍        | 829/5758 [04:41<29:36,  2.77it/s]
                                                  

 14%|█▍        | 829/5758 [04:41<29:36,  2.77it/s]
 14%|█▍        | 830/5758 [04:41<29:25,  2.79it/s]
                                                  
{'loss': '2.404', 'grad_norm': '17.5', 'learning_rate': '8.656e-06', 'num_tokens': '5.102e+05', 'mean_token_accuracy': '0.5616', 'epoch': '0.1346'}
+{'loss': '1.511', 'grad_norm': '28.62', 'learning_rate': '8.654e-06', 'num_tokens': '5.103e+05', 'mean_token_accuracy': '0.6949', 'epoch': '0.1348'}
+{'loss': '1.177', 'grad_norm': '12.12', 'learning_rate': '8.652e-06', 'num_tokens': '5.11e+05', 'mean_token_accuracy': '0.6809', 'epoch': '0.1349'}
+{'loss': '1.907', 'grad_norm': '16.38', 'learning_rate': '8.651e-06', 'num_tokens': '5.115e+05', 'mean_token_accuracy': '0.5494', 'epoch': '0.1351'}
+{'loss': '1.539', 'grad_norm': '12.62', 'learning_rate': '8.649e-06', 'num_tokens': '5.123e+05', 'mean_token_accuracy': '0.6005', 'epoch': '0.1353'}
+{'loss': '1.734', 'grad_norm': '10.06', 'learning_rate': '8.647e-06', 'num_tokens': '5.134e+05', 'mean_token_accuracy': '0.6176', 'epoch': '0.1355'}
+{'loss': '1.871', 'grad_norm': '17.25', 'learning_rate': '8.645e-06', 'num_tokens': '5.139e+05', 'mean_token_accuracy': '0.6252', 'epoch': '0.1356'}
+{'loss': '1.464', 'grad_norm': '25', 'learning_rate': '8.644e-06', 'num_tokens': '5.141e+05', 'mean_token_accuracy': '0.6464', 'epoch': '0.1358'}
+{'loss': '1.383', 'grad_norm': '6.688', 'learning_rate': '8.642e-06', 'num_tokens': '5.16e+05', 'mean_token_accuracy': '0.6785', 'epoch': '0.136'}
+{'loss': '1.485', 'grad_norm': '18.88', 'learning_rate': '8.64e-06', 'num_tokens': '5.163e+05', 'mean_token_accuracy': '0.6849', 'epoch': '0.1362'}
+{'loss': '1.008', 'grad_norm': '32.25', 'learning_rate': '8.638e-06', 'num_tokens': '5.164e+05', 'mean_token_accuracy': '0.7946', 'epoch': '0.1363'}
+{'loss': '1.747', 'grad_norm': '11.62', 'learning_rate': '8.637e-06', 'num_tokens': '5.171e+05', 'mean_token_accuracy': '0.6291', 'epoch': '0.1365'}
+{'loss': '1.155', 'grad_norm': '13', 'learning_rate': '8.635e-06', 'num_tokens': '5.177e+05', 'mean_token_accuracy': '0.7028', 'epoch': '0.1367'}
+{'loss': '1.413', 'grad_norm': '19.12', 'learning_rate': '8.633e-06', 'num_tokens': '5.179e+05', 'mean_token_accuracy': '0.7194', 'epoch': '0.1369'}
+{'loss': '1.213', 'grad_norm': '12.31', 'learning_rate': '8.631e-06', 'num_tokens': '5.186e+05', 'mean_token_accuracy': '0.6681', 'epoch': '0.137'}
+{'loss': '2.362', 'grad_norm': '22.75', 'learning_rate': '8.63e-06', 'num_tokens': '5.19e+05', 'mean_token_accuracy': '0.5549', 'epoch': '0.1372'}
+{'loss': '1.13', 'grad_norm': '13.56', 'learning_rate': '8.628e-06', 'num_tokens': '5.195e+05', 'mean_token_accuracy': '0.7251', 'epoch': '0.1374'}
+{'loss': '2.103', 'grad_norm': '16.25', 'learning_rate': '8.626e-06', 'num_tokens': '5.2e+05', 'mean_token_accuracy': '0.5584', 'epoch': '0.1375'}
+{'loss': '2.149', 'grad_norm': '8.875', 'learning_rate': '8.625e-06', 'num_tokens': '5.216e+05', 'mean_token_accuracy': '0.5467', 'epoch': '0.1377'}
+{'loss': '1.257', 'grad_norm': '11.69', 'learning_rate': '8.623e-06', 'num_tokens': '5.223e+05', 'mean_token_accuracy': '0.6819', 'epoch': '0.1379'}
+{'loss': '1.131', 'grad_norm': '11.81', 'learning_rate': '8.621e-06', 'num_tokens': '5.231e+05', 'mean_token_accuracy': '0.671', 'epoch': '0.1381'}
+{'loss': '1.109', 'grad_norm': '8.5', 'learning_rate': '8.619e-06', 'num_tokens': '5.244e+05', 'mean_token_accuracy': '0.7431', 'epoch': '0.1382'}
+{'loss': '1.483', 'grad_norm': '13.69', 'learning_rate': '8.618e-06', 'num_tokens': '5.248e+05', 'mean_token_accuracy': '0.683', 'epoch': '0.1384'}
+{'loss': '2.456', 'grad_norm': '19.38', 'learning_rate': '8.616e-06', 'num_tokens': '5.254e+05', 'mean_token_accuracy': '0.538', 'epoch': '0.1386'}
+{'loss': '1.303', 'grad_norm': '7.094', 'learning_rate': '8.614e-06', 'num_tokens': '5.274e+05', 'mean_token_accuracy': '0.7024', 'epoch': '0.1388'}
+{'loss': '1.566', 'grad_norm': '14.06', 'learning_rate': '8.612e-06', 'num_tokens': '5.28e+05', 'mean_token_accuracy': '0.5821', 'epoch': '0.1389'}
+{'loss': '1.192', 'grad_norm': '13.5', 'learning_rate': '8.611e-06', 'num_tokens': '5.286e+05', 'mean_token_accuracy': '0.7041', 'epoch': '0.1391'}
+{'loss': '1.26', 'grad_norm': '35.75', 'learning_rate': '8.609e-06', 'num_tokens': '5.287e+05', 'mean_token_accuracy': '0.8026', 'epoch': '0.1393'}
+{'loss': '0.7263', 'grad_norm': '20', 'learning_rate': '8.607e-06', 'num_tokens': '5.289e+05', 'mean_token_accuracy': '0.8118', 'epoch': '0.1395'}
+{'loss': '1.264', 'grad_norm': '11.12', 'learning_rate': '8.605e-06', 'num_tokens': '5.297e+05', 'mean_token_accuracy': '0.7312', 'epoch': '0.1396'}
+{'loss': '1.052', 'grad_norm': '13', 'learning_rate': '8.604e-06', 'num_tokens': '5.303e+05', 'mean_token_accuracy': '0.7033', 'epoch': '0.1398'}
+{'loss': '1.156', 'grad_norm': '14.62', 'learning_rate': '8.602e-06', 'num_tokens': '5.308e+05', 'mean_token_accuracy': '0.7294', 'epoch': '0.14'}
+{'loss': '1.085', 'grad_norm': '28.75', 'learning_rate': '8.6e-06', 'num_tokens': '5.31e+05', 'mean_token_accuracy': '0.7707', 'epoch': '0.1402'}
+{'loss': '0.7174', 'grad_norm': '7.969', 'learning_rate': '8.598e-06', 'num_tokens': '5.321e+05', 'mean_token_accuracy': '0.7974', 'epoch': '0.1403'}
+{'loss': '1.536', 'grad_norm': '29.88', 'learning_rate': '8.597e-06', 'num_tokens': '5.322e+05', 'mean_token_accuracy': '0.7333', 'epoch': '0.1405'}
+{'loss': '1.106', 'grad_norm': '9.375', 'learning_rate': '8.595e-06', 'num_tokens': '5.333e+05', 'mean_token_accuracy': '0.7295', 'epoch': '0.1407'}
+{'loss': '1.197', 'grad_norm': '12.19', 'learning_rate': '8.593e-06', 'num_tokens': '5.34e+05', 'mean_token_accuracy': '0.7001', 'epoch': '0.1408'}
+{'loss': '1.826', 'grad_norm': '42.75', 'learning_rate': '8.592e-06', 'num_tokens': '5.341e+05', 'mean_token_accuracy': '0.6', 'epoch': '0.141'}
+{'loss': '1.234', 'grad_norm': '24.62', 'learning_rate': '8.59e-06', 'num_tokens': '5.343e+05', 'mean_token_accuracy': '0.7596', 'epoch': '0.1412'}
+{'loss': '1.72', 'grad_norm': '11.81', 'learning_rate': '8.588e-06', 'num_tokens': '5.351e+05', 'mean_token_accuracy': '0.6584', 'epoch': '0.1414'}
+{'loss': '1.744', 'grad_norm': '19', 'learning_rate': '8.586e-06', 'num_tokens': '5.356e+05', 'mean_token_accuracy': '0.5864', 'epoch': '0.1415'}
+{'loss': '1.098', 'grad_norm': '7.844', 'learning_rate': '8.585e-06', 'num_tokens': '5.382e+05', 'mean_token_accuracy': '0.7578', 'epoch': '0.1417'}
+{'loss': '2.07', 'grad_norm': '10.12', 'learning_rate': '8.583e-06', 'num_tokens': '5.395e+05', 'mean_token_accuracy': '0.5842', 'epoch': '0.1419'}
+{'loss': '1.715', 'grad_norm': '32', 'learning_rate': '8.581e-06', 'num_tokens': '5.396e+05', 'mean_token_accuracy': '0.6496', 'epoch': '0.1421'}
+{'loss': '1.734', 'grad_norm': '48.75', 'learning_rate': '8.579e-06', 'num_tokens': '5.397e+05', 'mean_token_accuracy': '0.7885', 'epoch': '0.1422'}
+{'loss': '0.8648', 'grad_norm': '34', 'learning_rate': '8.578e-06', 'num_tokens': '5.397e+05', 'mean_token_accuracy': '0.8621', 'epoch': '0.1424'}
+{'loss': '1.723', 'grad_norm': '12.81', 'learning_rate': '8.576e-06', 'num_tokens': '5.403e+05', 'mean_token_accuracy': '0.6493', 'epoch': '0.1426'}
+{'loss': '2.203', 'grad_norm': '22.38', 'learning_rate': '8.574e-06', 'num_tokens': '5.408e+05', 'mean_token_accuracy': '0.5718', 'epoch': '0.1428'}
+{'loss': '0.9267', 'grad_norm': '32.5', 'learning_rate': '8.572e-06', 'num_tokens': '5.408e+05', 'mean_token_accuracy': '0.7195', 'epoch': '0.1429'}
+{'loss': '1.57', 'grad_norm': '17.88', 'learning_rate': '8.571e-06', 'num_tokens': '5.412e+05', 'mean_token_accuracy': '0.6596', 'epoch': '0.1431'}
+{'loss': '1.18', 'grad_norm': '11.56', 'learning_rate': '8.569e-06', 'num_tokens': '5.421e+05', 'mean_token_accuracy': '0.6655', 'epoch': '0.1433'}
+{'loss': '1.557', 'grad_norm': '17.5', 'learning_rate': '8.567e-06', 'num_tokens': '5.424e+05', 'mean_token_accuracy': '0.6755', 'epoch': '0.1435'}
+{'loss': '1.649', 'grad_norm': '13.06', 'learning_rate': '8.565e-06', 'num_tokens': '5.432e+05', 'mean_token_accuracy': '0.6238', 'epoch': '0.1436'}
+{'loss': '1.36', 'grad_norm': '16.75', 'learning_rate': '8.564e-06', 'num_tokens': '5.436e+05', 'mean_token_accuracy': '0.7039', 'epoch': '0.1438'}
+{'loss': '2.493', 'grad_norm': '13.75', 'learning_rate': '8.562e-06', 'num_tokens': '5.44e+05', 'mean_token_accuracy': '0.509', 'epoch': '0.144'}
+
 14%|█▍        | 830/5758 [04:41<29:25,  2.79it/s]
 14%|█▍        | 831/5758 [04:42<27:57,  2.94it/s]
                                                  

 14%|█▍        | 831/5758 [04:42<27:57,  2.94it/s]
 14%|█▍        | 832/5758 [04:42<29:15,  2.81it/s]
                                                  

 14%|█▍        | 832/5758 [04:42<29:15,  2.81it/s]
 14%|█▍        | 833/5758 [04:42<27:39,  2.97it/s]
                                                  

 14%|█▍        | 833/5758 [04:42<27:39,  2.97it/s]
 14%|█▍        | 834/5758 [04:43<27:28,  2.99it/s]
                                                  

 14%|█▍        | 834/5758 [04:43<27:28,  2.99it/s]
 15%|█▍        | 835/5758 [04:43<27:35,  2.97it/s]
                                                  

 15%|█▍        | 835/5758 [04:43<27:35,  2.97it/s]
 15%|█▍        | 836/5758 [04:43<28:02,  2.92it/s]
                                                  

 15%|█▍        | 836/5758 [04:43<28:02,  2.92it/s]
 15%|█▍        | 837/5758 [04:44<28:27,  2.88it/s]
                                                  

 15%|█▍        | 837/5758 [04:44<28:27,  2.88it/s]
 15%|█▍        | 838/5758 [04:44<27:49,  2.95it/s]
                                                  

 15%|█▍        | 838/5758 [04:44<27:49,  2.95it/s]
 15%|█▍        | 839/5758 [04:44<27:24,  2.99it/s]
                                                  

 15%|█▍        | 839/5758 [04:44<27:24,  2.99it/s]
 15%|█▍        | 840/5758 [04:45<29:00,  2.83it/s]
                                                  

 15%|█▍        | 840/5758 [04:45<29:00,  2.83it/s]
 15%|█▍        | 841/5758 [04:45<29:18,  2.80it/s]
                                                  

 15%|█▍        | 841/5758 [04:45<29:18,  2.80it/s]
 15%|█▍        | 842/5758 [04:45<29:12,  2.81it/s]
                                                  

 15%|█▍        | 842/5758 [04:45<29:12,  2.81it/s]
 15%|█▍        | 843/5758 [04:46<29:11,  2.81it/s]
                                                  

 15%|█▍        | 843/5758 [04:46<29:11,  2.81it/s]
 15%|█▍        | 844/5758 [04:46<28:19,  2.89it/s]
                                                  

 15%|█▍        | 844/5758 [04:46<28:19,  2.89it/s]
 15%|█▍        | 845/5758 [04:46<27:11,  3.01it/s]
                                                  

 15%|█▍        | 845/5758 [04:46<27:11,  3.01it/s]
 15%|█▍        | 846/5758 [04:47<26:51,  3.05it/s]
                                                  

 15%|█▍        | 846/5758 [04:47<26:51,  3.05it/s]
 15%|█▍        | 847/5758 [04:47<27:53,  2.94it/s]
                                                  

 15%|█▍        | 847/5758 [04:47<27:53,  2.94it/s]
 15%|█▍        | 848/5758 [04:47<26:40,  3.07it/s]
                                                  

 15%|█▍        | 848/5758 [04:47<26:40,  3.07it/s]
 15%|█▍        | 849/5758 [04:48<25:56,  3.15it/s]
                                                  

 15%|█▍        | 849/5758 [04:48<25:56,  3.15it/s]
 15%|█▍        | 850/5758 [04:48<25:44,  3.18it/s]
                                                  

 15%|█▍        | 850/5758 [04:48<25:44,  3.18it/s]
 15%|█▍        | 851/5758 [04:48<26:36,  3.07it/s]
                                                  

 15%|█▍        | 851/5758 [04:48<26:36,  3.07it/s]
 15%|█▍        | 852/5758 [04:49<26:11,  3.12it/s]
                                                  

 15%|█▍        | 852/5758 [04:49<26:11,  3.12it/s]
 15%|█▍        | 853/5758 [04:49<26:23,  3.10it/s]
                                                  

 15%|█▍        | 853/5758 [04:49<26:23,  3.10it/s]
 15%|█▍        | 854/5758 [04:49<25:46,  3.17it/s]
                                                  

 15%|█▍        | 854/5758 [04:49<25:46,  3.17it/s]
 15%|█▍        | 855/5758 [04:50<25:28,  3.21it/s]
                                                  

 15%|█▍        | 855/5758 [04:50<25:28,  3.21it/s]
 15%|█▍        | 856/5758 [04:50<25:00,  3.27it/s]
                                                  

 15%|█▍        | 856/5758 [04:50<25:00,  3.27it/s]
 15%|█▍        | 857/5758 [04:50<27:32,  2.97it/s]
                                                  

 15%|█▍        | 857/5758 [04:50<27:32,  2.97it/s]
 15%|█▍        | 858/5758 [04:51<26:36,  3.07it/s]
                                                  

 15%|█▍        | 858/5758 [04:51<26:36,  3.07it/s]
 15%|█▍        | 859/5758 [04:51<27:24,  2.98it/s]
                                                  

 15%|█▍        | 859/5758 [04:51<27:24,  2.98it/s]
 15%|█▍        | 860/5758 [04:51<26:59,  3.02it/s]
                                                  

 15%|█▍        | 860/5758 [04:51<26:59,  3.02it/s]
 15%|█▍        | 861/5758 [04:52<26:59,  3.02it/s]
                                                  

 15%|█▍        | 861/5758 [04:52<26:59,  3.02it/s]
 15%|█▍        | 862/5758 [04:52<26:07,  3.12it/s]
                                                  

 15%|█▍        | 862/5758 [04:52<26:07,  3.12it/s]
 15%|█▍        | 863/5758 [04:52<26:59,  3.02it/s]
                                                  

 15%|█▍        | 863/5758 [04:52<26:59,  3.02it/s]
 15%|█▌        | 864/5758 [04:53<25:57,  3.14it/s]
                                                  

 15%|█▌        | 864/5758 [04:53<25:57,  3.14it/s]
 15%|█▌        | 865/5758 [04:53<25:38,  3.18it/s]
                                                  

 15%|█▌        | 865/5758 [04:53<25:38,  3.18it/s]
 15%|█▌        | 866/5758 [04:53<26:53,  3.03it/s]
                                                  

 15%|█▌        | 866/5758 [04:53<26:53,  3.03it/s]
 15%|█▌        | 867/5758 [04:53<26:13,  3.11it/s]
                                                  

 15%|█▌        | 867/5758 [04:53<26:13,  3.11it/s]
 15%|█▌        | 868/5758 [04:54<26:29,  3.08it/s]
                                                  

 15%|█▌        | 868/5758 [04:54<26:29,  3.08it/s]
 15%|█▌        | 869/5758 [04:54<25:37,  3.18it/s]
                                                  

 15%|█▌        | 869/5758 [04:54<25:37,  3.18it/s]
 15%|█▌        | 870/5758 [04:54<26:18,  3.10it/s]
                                                  

 15%|█▌        | 870/5758 [04:54<26:18,  3.10it/s]
 15%|█▌        | 871/5758 [04:55<25:51,  3.15it/s]
                                                  

 15%|█▌        | 871/5758 [04:55<25:51,  3.15it/s]
 15%|█▌        | 872/5758 [04:55<26:00,  3.13it/s]
                                                  

 15%|█▌        | 872/5758 [04:55<26:00,  3.13it/s]
 15%|█▌        | 873/5758 [04:55<27:01,  3.01it/s]
                                                  

 15%|█▌        | 873/5758 [04:55<27:01,  3.01it/s]
 15%|█▌        | 874/5758 [04:56<26:39,  3.05it/s]
                                                  

 15%|█▌        | 874/5758 [04:56<26:39,  3.05it/s]
 15%|█▌        | 875/5758 [04:56<27:03,  3.01it/s]
                                                  

 15%|█▌        | 875/5758 [04:56<27:03,  3.01it/s]
 15%|█▌        | 876/5758 [04:56<26:16,  3.10it/s]
                                                  

 15%|█▌        | 876/5758 [04:56<26:16,  3.10it/s]
 15%|█▌        | 877/5758 [04:57<27:24,  2.97it/s]
                                                  

 15%|█▌        | 877/5758 [04:57<27:24,  2.97it/s]
 15%|█▌        | 878/5758 [04:57<28:32,  2.85it/s]
                                                  

 15%|█▌        | 878/5758 [04:57<28:32,  2.85it/s]
 15%|█▌        | 879/5758 [04:57<28:21,  2.87it/s]
                                                  

 15%|█▌        | 879/5758 [04:57<28:21,  2.87it/s]
 15%|█▌        | 880/5758 [04:58<27:09,  2.99it/s]
                                                  

 15%|█▌        | 880/5758 [04:58<27:09,  2.99it/s]
 15%|█▌        | 881/5758 [04:58<28:15,  2.88it/s]
                                                  

 15%|█▌        | 881/5758 [04:58<28:15,  2.88it/s]
 15%|█▌        | 882/5758 [04:59<28:02,  2.90it/s]
                                                  

 15%|█▌        | 882/5758 [04:59<28:02,  2.90it/s]
 15%|█▌        | 883/5758 [04:59<27:05,  3.00it/s]
                                                  

 15%|█▌        | 883/5758 [04:59<27:05,  3.00it/s]
 15%|█▌        | 884/5758 [04:59<26:38,  3.05it/s]
                                                  

 15%|█▌        | 884/5758 [04:59<26:38,  3.05it/s]
 15%|█▌        | 885/5758 [04:59<26:20,  3.08it/s]
                                                  
{'loss': '2.072', 'grad_norm': '16.88', 'learning_rate': '8.56e-06', 'num_tokens': '5.445e+05', 'mean_token_accuracy': '0.5848', 'epoch': '0.1441'}
+{'loss': '1.154', 'grad_norm': '13.94', 'learning_rate': '8.559e-06', 'num_tokens': '5.451e+05', 'mean_token_accuracy': '0.6597', 'epoch': '0.1443'}
+{'loss': '1.09', 'grad_norm': '13.12', 'learning_rate': '8.557e-06', 'num_tokens': '5.457e+05', 'mean_token_accuracy': '0.6971', 'epoch': '0.1445'}
+{'loss': '1.893', 'grad_norm': '12.81', 'learning_rate': '8.555e-06', 'num_tokens': '5.463e+05', 'mean_token_accuracy': '0.6254', 'epoch': '0.1447'}
+{'loss': '2.106', 'grad_norm': '10.5', 'learning_rate': '8.553e-06', 'num_tokens': '5.471e+05', 'mean_token_accuracy': '0.5896', 'epoch': '0.1448'}
+{'loss': '1.663', 'grad_norm': '15.62', 'learning_rate': '8.552e-06', 'num_tokens': '5.476e+05', 'mean_token_accuracy': '0.6319', 'epoch': '0.145'}
+{'loss': '1.045', 'grad_norm': '18', 'learning_rate': '8.55e-06', 'num_tokens': '5.478e+05', 'mean_token_accuracy': '0.7511', 'epoch': '0.1452'}
+{'loss': '0.6447', 'grad_norm': '6.25', 'learning_rate': '8.548e-06', 'num_tokens': '5.498e+05', 'mean_token_accuracy': '0.8149', 'epoch': '0.1454'}
+{'loss': '1.272', 'grad_norm': '7.656', 'learning_rate': '8.546e-06', 'num_tokens': '5.512e+05', 'mean_token_accuracy': '0.6861', 'epoch': '0.1455'}
+{'loss': '1.216', 'grad_norm': '11.75', 'learning_rate': '8.545e-06', 'num_tokens': '5.52e+05', 'mean_token_accuracy': '0.6662', 'epoch': '0.1457'}
+{'loss': '1.565', 'grad_norm': '6.25', 'learning_rate': '8.543e-06', 'num_tokens': '5.551e+05', 'mean_token_accuracy': '0.6495', 'epoch': '0.1459'}
+{'loss': '0.7322', 'grad_norm': '28.62', 'learning_rate': '8.541e-06', 'num_tokens': '5.552e+05', 'mean_token_accuracy': '0.8291', 'epoch': '0.1461'}
+{'loss': '1.617', 'grad_norm': '18', 'learning_rate': '8.539e-06', 'num_tokens': '5.554e+05', 'mean_token_accuracy': '0.6849', 'epoch': '0.1462'}
+{'loss': '1.273', 'grad_norm': '6.812', 'learning_rate': '8.538e-06', 'num_tokens': '5.573e+05', 'mean_token_accuracy': '0.6847', 'epoch': '0.1464'}
+{'loss': '0.8859', 'grad_norm': '16.38', 'learning_rate': '8.536e-06', 'num_tokens': '5.577e+05', 'mean_token_accuracy': '0.7609', 'epoch': '0.1466'}
+{'loss': '2.001', 'grad_norm': '17', 'learning_rate': '8.534e-06', 'num_tokens': '5.584e+05', 'mean_token_accuracy': '0.6146', 'epoch': '0.1468'}
+{'loss': '1.666', 'grad_norm': '26.38', 'learning_rate': '8.532e-06', 'num_tokens': '5.585e+05', 'mean_token_accuracy': '0.6364', 'epoch': '0.1469'}
+{'loss': '1.525', 'grad_norm': '10.12', 'learning_rate': '8.531e-06', 'num_tokens': '5.596e+05', 'mean_token_accuracy': '0.659', 'epoch': '0.1471'}
+{'loss': '1.135', 'grad_norm': '12.88', 'learning_rate': '8.529e-06', 'num_tokens': '5.602e+05', 'mean_token_accuracy': '0.6895', 'epoch': '0.1473'}
+{'loss': '1.839', 'grad_norm': '18.88', 'learning_rate': '8.527e-06', 'num_tokens': '5.605e+05', 'mean_token_accuracy': '0.6316', 'epoch': '0.1474'}
+{'loss': '1.942', 'grad_norm': '28.5', 'learning_rate': '8.526e-06', 'num_tokens': '5.607e+05', 'mean_token_accuracy': '0.6171', 'epoch': '0.1476'}
+{'loss': '2.172', 'grad_norm': '13.69', 'learning_rate': '8.524e-06', 'num_tokens': '5.611e+05', 'mean_token_accuracy': '0.5772', 'epoch': '0.1478'}
+{'loss': '1.409', 'grad_norm': '12.75', 'learning_rate': '8.522e-06', 'num_tokens': '5.618e+05', 'mean_token_accuracy': '0.6035', 'epoch': '0.148'}
+{'loss': '1.305', 'grad_norm': '12.19', 'learning_rate': '8.52e-06', 'num_tokens': '5.627e+05', 'mean_token_accuracy': '0.6623', 'epoch': '0.1481'}
+{'loss': '1.413', 'grad_norm': '12.94', 'learning_rate': '8.519e-06', 'num_tokens': '5.633e+05', 'mean_token_accuracy': '0.6159', 'epoch': '0.1483'}
+{'loss': '1.193', 'grad_norm': '12.25', 'learning_rate': '8.517e-06', 'num_tokens': '5.64e+05', 'mean_token_accuracy': '0.656', 'epoch': '0.1485'}
+{'loss': '2.159', 'grad_norm': '17.88', 'learning_rate': '8.515e-06', 'num_tokens': '5.646e+05', 'mean_token_accuracy': '0.5534', 'epoch': '0.1487'}
+{'loss': '1.311', 'grad_norm': '14.38', 'learning_rate': '8.513e-06', 'num_tokens': '5.652e+05', 'mean_token_accuracy': '0.6516', 'epoch': '0.1488'}
+{'loss': '1.121', 'grad_norm': '16.12', 'learning_rate': '8.512e-06', 'num_tokens': '5.658e+05', 'mean_token_accuracy': '0.731', 'epoch': '0.149'}
+{'loss': '2.254', 'grad_norm': '8.062', 'learning_rate': '8.51e-06', 'num_tokens': '5.678e+05', 'mean_token_accuracy': '0.5437', 'epoch': '0.1492'}
+{'loss': '1.946', 'grad_norm': '15.62', 'learning_rate': '8.508e-06', 'num_tokens': '5.682e+05', 'mean_token_accuracy': '0.6104', 'epoch': '0.1494'}
+{'loss': '1.439', 'grad_norm': '12.88', 'learning_rate': '8.506e-06', 'num_tokens': '5.691e+05', 'mean_token_accuracy': '0.6532', 'epoch': '0.1495'}
+{'loss': '0.716', 'grad_norm': '8', 'learning_rate': '8.505e-06', 'num_tokens': '5.704e+05', 'mean_token_accuracy': '0.7966', 'epoch': '0.1497'}
+{'loss': '1.684', 'grad_norm': '21.75', 'learning_rate': '8.503e-06', 'num_tokens': '5.706e+05', 'mean_token_accuracy': '0.6504', 'epoch': '0.1499'}
+{'loss': '2.53', 'grad_norm': '20.5', 'learning_rate': '8.501e-06', 'num_tokens': '5.712e+05', 'mean_token_accuracy': '0.5315', 'epoch': '0.1501'}
+{'loss': '1.693', 'grad_norm': '24.38', 'learning_rate': '8.499e-06', 'num_tokens': '5.714e+05', 'mean_token_accuracy': '0.6461', 'epoch': '0.1502'}
+{'loss': '0.798', 'grad_norm': '10.06', 'learning_rate': '8.498e-06', 'num_tokens': '5.725e+05', 'mean_token_accuracy': '0.7744', 'epoch': '0.1504'}
+{'loss': '1.499', 'grad_norm': '17.12', 'learning_rate': '8.496e-06', 'num_tokens': '5.728e+05', 'mean_token_accuracy': '0.6686', 'epoch': '0.1506'}
+{'loss': '1.279', 'grad_norm': '7.156', 'learning_rate': '8.494e-06', 'num_tokens': '5.746e+05', 'mean_token_accuracy': '0.6855', 'epoch': '0.1507'}
+{'loss': '1.533', 'grad_norm': '10.44', 'learning_rate': '8.493e-06', 'num_tokens': '5.758e+05', 'mean_token_accuracy': '0.6559', 'epoch': '0.1509'}
+{'loss': '1.066', 'grad_norm': '12.06', 'learning_rate': '8.491e-06', 'num_tokens': '5.763e+05', 'mean_token_accuracy': '0.7624', 'epoch': '0.1511'}
+{'loss': '1.616', 'grad_norm': '17.75', 'learning_rate': '8.489e-06', 'num_tokens': '5.766e+05', 'mean_token_accuracy': '0.6881', 'epoch': '0.1513'}
+{'loss': '1.256', 'grad_norm': '22.25', 'learning_rate': '8.487e-06', 'num_tokens': '5.768e+05', 'mean_token_accuracy': '0.6976', 'epoch': '0.1514'}
+{'loss': '1.273', 'grad_norm': '11.5', 'learning_rate': '8.486e-06', 'num_tokens': '5.778e+05', 'mean_token_accuracy': '0.6598', 'epoch': '0.1516'}
+{'loss': '2.098', 'grad_norm': '11.19', 'learning_rate': '8.484e-06', 'num_tokens': '5.786e+05', 'mean_token_accuracy': '0.5885', 'epoch': '0.1518'}
+{'loss': '2.126', 'grad_norm': '18.12', 'learning_rate': '8.482e-06', 'num_tokens': '5.791e+05', 'mean_token_accuracy': '0.5941', 'epoch': '0.152'}
+{'loss': '1.747', 'grad_norm': '16.38', 'learning_rate': '8.48e-06', 'num_tokens': '5.794e+05', 'mean_token_accuracy': '0.6448', 'epoch': '0.1521'}
+{'loss': '1.096', 'grad_norm': '13.31', 'learning_rate': '8.479e-06', 'num_tokens': '5.799e+05', 'mean_token_accuracy': '0.7097', 'epoch': '0.1523'}
+{'loss': '2.026', 'grad_norm': '16.38', 'learning_rate': '8.477e-06', 'num_tokens': '5.805e+05', 'mean_token_accuracy': '0.6021', 'epoch': '0.1525'}
+{'loss': '1.338', 'grad_norm': '11.38', 'learning_rate': '8.475e-06', 'num_tokens': '5.815e+05', 'mean_token_accuracy': '0.6218', 'epoch': '0.1527'}
+{'loss': '1.187', 'grad_norm': '13.06', 'learning_rate': '8.473e-06', 'num_tokens': '5.821e+05', 'mean_token_accuracy': '0.6755', 'epoch': '0.1528'}
+{'loss': '2.087', 'grad_norm': '16.75', 'learning_rate': '8.472e-06', 'num_tokens': '5.827e+05', 'mean_token_accuracy': '0.5899', 'epoch': '0.153'}
+{'loss': '1.166', 'grad_norm': '10.75', 'learning_rate': '8.47e-06', 'num_tokens': '5.836e+05', 'mean_token_accuracy': '0.6834', 'epoch': '0.1532'}
+{'loss': '1.274', 'grad_norm': '12.81', 'learning_rate': '8.468e-06', 'num_tokens': '5.843e+05', 'mean_token_accuracy': '0.6483', 'epoch': '0.1534'}
+{'loss': '1.227', 'grad_norm': '11.31', 'learning_rate': '8.466e-06', 'num_tokens': '5.851e+05', 'mean_token_accuracy': '0.6542', 'epoch': '0.1535'}
+
 15%|█▌        | 885/5758 [04:59<26:20,  3.08it/s]
 15%|█▌        | 886/5758 [05:00<29:39,  2.74it/s]
                                                  

 15%|█▌        | 886/5758 [05:00<29:39,  2.74it/s]
 15%|█▌        | 887/5758 [05:00<29:49,  2.72it/s]
                                                  

 15%|█▌        | 887/5758 [05:00<29:49,  2.72it/s]
 15%|█▌        | 888/5758 [05:01<28:28,  2.85it/s]
                                                  

 15%|█▌        | 888/5758 [05:01<28:28,  2.85it/s]
 15%|█▌        | 889/5758 [05:01<27:27,  2.95it/s]
                                                  

 15%|█▌        | 889/5758 [05:01<27:27,  2.95it/s]
 15%|█▌        | 890/5758 [05:01<26:38,  3.04it/s]
                                                  

 15%|█▌        | 890/5758 [05:01<26:38,  3.04it/s]
 15%|█▌        | 891/5758 [05:02<28:26,  2.85it/s]
                                                  

 15%|█▌        | 891/5758 [05:02<28:26,  2.85it/s]
 15%|█▌        | 892/5758 [05:02<26:41,  3.04it/s]
                                                  

 15%|█▌        | 892/5758 [05:02<26:41,  3.04it/s]
 16%|█▌        | 893/5758 [05:02<26:23,  3.07it/s]
                                                  

 16%|█▌        | 893/5758 [05:02<26:23,  3.07it/s]
 16%|█▌        | 894/5758 [05:02<24:53,  3.26it/s]
                                                  

 16%|█▌        | 894/5758 [05:02<24:53,  3.26it/s]
 16%|█▌        | 895/5758 [05:03<24:54,  3.25it/s]
                                                  

 16%|█▌        | 895/5758 [05:03<24:54,  3.25it/s]
 16%|█▌        | 896/5758 [05:03<24:42,  3.28it/s]
                                                  

 16%|█▌        | 896/5758 [05:03<24:42,  3.28it/s]
 16%|█▌        | 897/5758 [05:03<26:10,  3.10it/s]
                                                  

 16%|█▌        | 897/5758 [05:03<26:10,  3.10it/s]
 16%|█▌        | 898/5758 [05:04<27:06,  2.99it/s]
                                                  

 16%|█▌        | 898/5758 [05:04<27:06,  2.99it/s]
 16%|█▌        | 899/5758 [05:04<27:21,  2.96it/s]
                                                  

 16%|█▌        | 899/5758 [05:04<27:21,  2.96it/s]
 16%|█▌        | 900/5758 [05:05<27:46,  2.91it/s]
                                                  

 16%|█▌        | 900/5758 [05:05<27:46,  2.91it/s]
 16%|█▌        | 901/5758 [05:05<27:03,  2.99it/s]
                                                  

 16%|█▌        | 901/5758 [05:05<27:03,  2.99it/s]
 16%|█▌        | 902/5758 [05:05<26:03,  3.11it/s]
                                                  

 16%|█▌        | 902/5758 [05:05<26:03,  3.11it/s]
 16%|█▌        | 903/5758 [05:05<26:06,  3.10it/s]
                                                  

 16%|█▌        | 903/5758 [05:05<26:06,  3.10it/s]
 16%|█▌        | 904/5758 [05:06<24:48,  3.26it/s]
                                                  

 16%|█▌        | 904/5758 [05:06<24:48,  3.26it/s]
 16%|█▌        | 905/5758 [05:06<26:39,  3.03it/s]
                                                  

 16%|█▌        | 905/5758 [05:06<26:39,  3.03it/s]
 16%|█▌        | 906/5758 [05:06<25:45,  3.14it/s]
                                                  

 16%|█▌        | 906/5758 [05:06<25:45,  3.14it/s]
 16%|█▌        | 907/5758 [05:07<27:08,  2.98it/s]
                                                  

 16%|█▌        | 907/5758 [05:07<27:08,  2.98it/s]
 16%|█▌        | 908/5758 [05:07<27:17,  2.96it/s]
                                                  

 16%|█▌        | 908/5758 [05:07<27:17,  2.96it/s]
 16%|█▌        | 909/5758 [05:07<28:06,  2.87it/s]
                                                  

 16%|█▌        | 909/5758 [05:07<28:06,  2.87it/s]
 16%|█▌        | 910/5758 [05:08<26:44,  3.02it/s]
                                                  

 16%|█▌        | 910/5758 [05:08<26:44,  3.02it/s]
 16%|█▌        | 911/5758 [05:08<29:02,  2.78it/s]
                                                  

 16%|█▌        | 911/5758 [05:08<29:02,  2.78it/s]
 16%|█▌        | 912/5758 [05:09<28:13,  2.86it/s]
                                                  

 16%|█▌        | 912/5758 [05:09<28:13,  2.86it/s]
 16%|█▌        | 913/5758 [05:09<28:55,  2.79it/s]
                                                  

 16%|█▌        | 913/5758 [05:09<28:55,  2.79it/s]
 16%|█▌        | 914/5758 [05:09<28:27,  2.84it/s]
                                                  

 16%|█▌        | 914/5758 [05:09<28:27,  2.84it/s]
 16%|█▌        | 915/5758 [05:10<26:58,  2.99it/s]
                                                  

 16%|█▌        | 915/5758 [05:10<26:58,  2.99it/s]
 16%|█▌        | 916/5758 [05:10<26:00,  3.10it/s]
                                                  

 16%|█▌        | 916/5758 [05:10<26:00,  3.10it/s]
 16%|█▌        | 917/5758 [05:10<25:23,  3.18it/s]
                                                  

 16%|█▌        | 917/5758 [05:10<25:23,  3.18it/s]
 16%|█▌        | 918/5758 [05:10<24:49,  3.25it/s]
                                                  

 16%|█▌        | 918/5758 [05:10<24:49,  3.25it/s]
 16%|█▌        | 919/5758 [05:11<25:37,  3.15it/s]
                                                  

 16%|█▌        | 919/5758 [05:11<25:37,  3.15it/s]
 16%|█▌        | 920/5758 [05:11<25:57,  3.11it/s]
                                                  

 16%|█▌        | 920/5758 [05:11<25:57,  3.11it/s]
 16%|█▌        | 921/5758 [05:11<25:03,  3.22it/s]
                                                  

 16%|█▌        | 921/5758 [05:11<25:03,  3.22it/s]
 16%|█▌        | 922/5758 [05:12<25:03,  3.22it/s]
                                                  

 16%|█▌        | 922/5758 [05:12<25:03,  3.22it/s]
 16%|█▌        | 923/5758 [05:12<24:40,  3.26it/s]
                                                  

 16%|█▌        | 923/5758 [05:12<24:40,  3.26it/s]
 16%|█▌        | 924/5758 [05:12<25:45,  3.13it/s]
                                                  

 16%|█▌        | 924/5758 [05:12<25:45,  3.13it/s]
 16%|█▌        | 925/5758 [05:13<25:24,  3.17it/s]
                                                  

 16%|█▌        | 925/5758 [05:13<25:24,  3.17it/s]
 16%|█▌        | 926/5758 [05:13<25:22,  3.17it/s]
                                                  

 16%|█▌        | 926/5758 [05:13<25:22,  3.17it/s]
 16%|█▌        | 927/5758 [05:13<25:13,  3.19it/s]
                                                  

 16%|█▌        | 927/5758 [05:13<25:13,  3.19it/s]
 16%|█▌        | 928/5758 [05:14<27:16,  2.95it/s]
                                                  

 16%|█▌        | 928/5758 [05:14<27:16,  2.95it/s]
 16%|█▌        | 929/5758 [05:14<27:44,  2.90it/s]
                                                  

 16%|█▌        | 929/5758 [05:14<27:44,  2.90it/s]
 16%|█▌        | 930/5758 [05:14<27:12,  2.96it/s]
                                                  

 16%|█▌        | 930/5758 [05:14<27:12,  2.96it/s]
 16%|█▌        | 931/5758 [05:15<26:42,  3.01it/s]
                                                  

 16%|█▌        | 931/5758 [05:15<26:42,  3.01it/s]
 16%|█▌        | 932/5758 [05:15<25:50,  3.11it/s]
                                                  

 16%|█▌        | 932/5758 [05:15<25:50,  3.11it/s]
 16%|█▌        | 933/5758 [05:15<25:55,  3.10it/s]
                                                  

 16%|█▌        | 933/5758 [05:15<25:55,  3.10it/s]
 16%|█▌        | 934/5758 [05:16<25:26,  3.16it/s]
                                                  

 16%|█▌        | 934/5758 [05:16<25:26,  3.16it/s]
 16%|█▌        | 935/5758 [05:16<25:31,  3.15it/s]
                                                  

 16%|█▌        | 935/5758 [05:16<25:31,  3.15it/s]
 16%|█▋        | 936/5758 [05:16<27:53,  2.88it/s]
                                                  

 16%|█▋        | 936/5758 [05:16<27:53,  2.88it/s]
 16%|█▋        | 937/5758 [05:17<33:38,  2.39it/s]
                                                  

 16%|█▋        | 937/5758 [05:17<33:38,  2.39it/s]
 16%|█▋        | 938/5758 [05:17<32:24,  2.48it/s]
                                                  

 16%|█▋        | 938/5758 [05:17<32:24,  2.48it/s]
 16%|█▋        | 939/5758 [05:18<32:18,  2.49it/s]
                                                  

 16%|█▋        | 939/5758 [05:18<32:18,  2.49it/s]
 16%|█▋        | 940/5758 [05:18<30:45,  2.61it/s]
                                                  
{'loss': '1.589', 'grad_norm': '12.69', 'learning_rate': '8.465e-06', 'num_tokens': '5.858e+05', 'mean_token_accuracy': '0.5918', 'epoch': '0.1537'}
+{'loss': '1.35', 'grad_norm': '21', 'learning_rate': '8.463e-06', 'num_tokens': '5.86e+05', 'mean_token_accuracy': '0.7176', 'epoch': '0.1539'}
+{'loss': '0.8681', 'grad_norm': '12', 'learning_rate': '8.461e-06', 'num_tokens': '5.866e+05', 'mean_token_accuracy': '0.7763', 'epoch': '0.154'}
+{'loss': '1.066', 'grad_norm': '13.19', 'learning_rate': '8.46e-06', 'num_tokens': '5.873e+05', 'mean_token_accuracy': '0.6821', 'epoch': '0.1542'}
+{'loss': '1.221', 'grad_norm': '13.12', 'learning_rate': '8.458e-06', 'num_tokens': '5.88e+05', 'mean_token_accuracy': '0.6703', 'epoch': '0.1544'}
+{'loss': '1.421', 'grad_norm': '17.12', 'learning_rate': '8.456e-06', 'num_tokens': '5.884e+05', 'mean_token_accuracy': '0.6792', 'epoch': '0.1546'}
+{'loss': '0.8768', 'grad_norm': '25.12', 'learning_rate': '8.454e-06', 'num_tokens': '5.885e+05', 'mean_token_accuracy': '0.8239', 'epoch': '0.1547'}
+{'loss': '0.9049', 'grad_norm': '46.25', 'learning_rate': '8.453e-06', 'num_tokens': '5.886e+05', 'mean_token_accuracy': '0.7705', 'epoch': '0.1549'}
+{'loss': '1.787', 'grad_norm': '10.06', 'learning_rate': '8.451e-06', 'num_tokens': '5.893e+05', 'mean_token_accuracy': '0.6182', 'epoch': '0.1551'}
+{'loss': '0.9982', 'grad_norm': '44.5', 'learning_rate': '8.449e-06', 'num_tokens': '5.894e+05', 'mean_token_accuracy': '0.8077', 'epoch': '0.1553'}
+{'loss': '0.93', 'grad_norm': '14.56', 'learning_rate': '8.447e-06', 'num_tokens': '5.897e+05', 'mean_token_accuracy': '0.8141', 'epoch': '0.1554'}
+{'loss': '1.309', 'grad_norm': '12.19', 'learning_rate': '8.446e-06', 'num_tokens': '5.904e+05', 'mean_token_accuracy': '0.6561', 'epoch': '0.1556'}
+{'loss': '2.04', 'grad_norm': '18.88', 'learning_rate': '8.444e-06', 'num_tokens': '5.909e+05', 'mean_token_accuracy': '0.5781', 'epoch': '0.1558'}
+{'loss': '2.114', 'grad_norm': '15.94', 'learning_rate': '8.442e-06', 'num_tokens': '5.914e+05', 'mean_token_accuracy': '0.5926', 'epoch': '0.156'}
+{'loss': '2.191', 'grad_norm': '16.62', 'learning_rate': '8.44e-06', 'num_tokens': '5.919e+05', 'mean_token_accuracy': '0.548', 'epoch': '0.1561'}
+{'loss': '1.297', 'grad_norm': '19.25', 'learning_rate': '8.439e-06', 'num_tokens': '5.921e+05', 'mean_token_accuracy': '0.7213', 'epoch': '0.1563'}
+{'loss': '2.28', 'grad_norm': '20', 'learning_rate': '8.437e-06', 'num_tokens': '5.925e+05', 'mean_token_accuracy': '0.5341', 'epoch': '0.1565'}
+{'loss': '1.452', 'grad_norm': '12.81', 'learning_rate': '8.435e-06', 'num_tokens': '5.931e+05', 'mean_token_accuracy': '0.6213', 'epoch': '0.1567'}
+{'loss': '0.9877', 'grad_norm': '11.56', 'learning_rate': '8.433e-06', 'num_tokens': '5.939e+05', 'mean_token_accuracy': '0.696', 'epoch': '0.1568'}
+{'loss': '1.345', 'grad_norm': '46.5', 'learning_rate': '8.432e-06', 'num_tokens': '5.939e+05', 'mean_token_accuracy': '0.6792', 'epoch': '0.157'}
+{'loss': '2.117', 'grad_norm': '16.5', 'learning_rate': '8.43e-06', 'num_tokens': '5.945e+05', 'mean_token_accuracy': '0.5656', 'epoch': '0.1572'}
+{'loss': '0.9293', 'grad_norm': '16', 'learning_rate': '8.428e-06', 'num_tokens': '5.948e+05', 'mean_token_accuracy': '0.7815', 'epoch': '0.1573'}
+{'loss': '1.276', 'grad_norm': '29.88', 'learning_rate': '8.427e-06', 'num_tokens': '5.949e+05', 'mean_token_accuracy': '0.7016', 'epoch': '0.1575'}
+{'loss': '1.667', 'grad_norm': '17.12', 'learning_rate': '8.425e-06', 'num_tokens': '5.954e+05', 'mean_token_accuracy': '0.6356', 'epoch': '0.1577'}
+{'loss': '1.106', 'grad_norm': '14.31', 'learning_rate': '8.423e-06', 'num_tokens': '5.959e+05', 'mean_token_accuracy': '0.6856', 'epoch': '0.1579'}
+{'loss': '0.6754', 'grad_norm': '8.562', 'learning_rate': '8.421e-06', 'num_tokens': '5.972e+05', 'mean_token_accuracy': '0.7811', 'epoch': '0.158'}
+{'loss': '2.197', 'grad_norm': '11.25', 'learning_rate': '8.42e-06', 'num_tokens': '5.979e+05', 'mean_token_accuracy': '0.5673', 'epoch': '0.1582'}
+{'loss': '1.34', 'grad_norm': '12.19', 'learning_rate': '8.418e-06', 'num_tokens': '5.987e+05', 'mean_token_accuracy': '0.6724', 'epoch': '0.1584'}
+{'loss': '1.399', 'grad_norm': '14.06', 'learning_rate': '8.416e-06', 'num_tokens': '5.993e+05', 'mean_token_accuracy': '0.611', 'epoch': '0.1586'}
+{'loss': '1.983', 'grad_norm': '20.75', 'learning_rate': '8.414e-06', 'num_tokens': '5.998e+05', 'mean_token_accuracy': '0.614', 'epoch': '0.1587'}
+{'loss': '1.065', 'grad_norm': '12.38', 'learning_rate': '8.413e-06', 'num_tokens': '6.004e+05', 'mean_token_accuracy': '0.7048', 'epoch': '0.1589'}
+{'loss': '1.346', 'grad_norm': '13.88', 'learning_rate': '8.411e-06', 'num_tokens': '6.01e+05', 'mean_token_accuracy': '0.6285', 'epoch': '0.1591'}
+{'loss': '1.039', 'grad_norm': '11.19', 'learning_rate': '8.409e-06', 'num_tokens': '6.016e+05', 'mean_token_accuracy': '0.7624', 'epoch': '0.1593'}
+{'loss': '1.328', 'grad_norm': '14.31', 'learning_rate': '8.407e-06', 'num_tokens': '6.022e+05', 'mean_token_accuracy': '0.6557', 'epoch': '0.1594'}
+{'loss': '2.142', 'grad_norm': '19.62', 'learning_rate': '8.406e-06', 'num_tokens': '6.027e+05', 'mean_token_accuracy': '0.5952', 'epoch': '0.1596'}
+{'loss': '1.295', 'grad_norm': '11.56', 'learning_rate': '8.404e-06', 'num_tokens': '6.036e+05', 'mean_token_accuracy': '0.6906', 'epoch': '0.1598'}
+{'loss': '1.128', 'grad_norm': '13.12', 'learning_rate': '8.402e-06', 'num_tokens': '6.042e+05', 'mean_token_accuracy': '0.6901', 'epoch': '0.16'}
+{'loss': '1.878', 'grad_norm': '37.75', 'learning_rate': '8.4e-06', 'num_tokens': '6.043e+05', 'mean_token_accuracy': '0.5904', 'epoch': '0.1601'}
+{'loss': '1.94', 'grad_norm': '16.75', 'learning_rate': '8.399e-06', 'num_tokens': '6.049e+05', 'mean_token_accuracy': '0.6063', 'epoch': '0.1603'}
+{'loss': '1.972', 'grad_norm': '10.19', 'learning_rate': '8.397e-06', 'num_tokens': '6.058e+05', 'mean_token_accuracy': '0.6223', 'epoch': '0.1605'}
+{'loss': '1.482', 'grad_norm': '14.44', 'learning_rate': '8.395e-06', 'num_tokens': '6.065e+05', 'mean_token_accuracy': '0.6549', 'epoch': '0.1606'}
+{'loss': '1.609', 'grad_norm': '24.62', 'learning_rate': '8.394e-06', 'num_tokens': '6.067e+05', 'mean_token_accuracy': '0.6685', 'epoch': '0.1608'}
+{'loss': '0.9033', 'grad_norm': '21.5', 'learning_rate': '8.392e-06', 'num_tokens': '6.068e+05', 'mean_token_accuracy': '0.8111', 'epoch': '0.161'}
+{'loss': '0.9902', 'grad_norm': '31.88', 'learning_rate': '8.39e-06', 'num_tokens': '6.07e+05', 'mean_token_accuracy': '0.8', 'epoch': '0.1612'}
+{'loss': '1.539', 'grad_norm': '40.5', 'learning_rate': '8.388e-06', 'num_tokens': '6.071e+05', 'mean_token_accuracy': '0.6514', 'epoch': '0.1613'}
+{'loss': '2.421', 'grad_norm': '25.38', 'learning_rate': '8.387e-06', 'num_tokens': '6.073e+05', 'mean_token_accuracy': '0.5153', 'epoch': '0.1615'}
+{'loss': '1.256', 'grad_norm': '14', 'learning_rate': '8.385e-06', 'num_tokens': '6.077e+05', 'mean_token_accuracy': '0.7389', 'epoch': '0.1617'}
+{'loss': '1.682', 'grad_norm': '10.94', 'learning_rate': '8.383e-06', 'num_tokens': '6.083e+05', 'mean_token_accuracy': '0.6533', 'epoch': '0.1619'}
+{'loss': '1.212', 'grad_norm': '11.88', 'learning_rate': '8.381e-06', 'num_tokens': '6.091e+05', 'mean_token_accuracy': '0.6601', 'epoch': '0.162'}
+{'loss': '1.322', 'grad_norm': '17.5', 'learning_rate': '8.38e-06', 'num_tokens': '6.094e+05', 'mean_token_accuracy': '0.7556', 'epoch': '0.1622'}
+{'loss': '0.7909', 'grad_norm': '12.06', 'learning_rate': '8.378e-06', 'num_tokens': '6.101e+05', 'mean_token_accuracy': '0.8', 'epoch': '0.1624'}
+{'loss': '2.263', 'grad_norm': '7.594', 'learning_rate': '8.376e-06', 'num_tokens': '6.135e+05', 'mean_token_accuracy': '0.5206', 'epoch': '0.1626'}
+{'loss': '0.9722', 'grad_norm': '15.06', 'learning_rate': '8.374e-06', 'num_tokens': '6.138e+05', 'mean_token_accuracy': '0.7926', 'epoch': '0.1627'}
+{'loss': '2.462', 'grad_norm': '42.75', 'learning_rate': '8.373e-06', 'num_tokens': '6.139e+05', 'mean_token_accuracy': '0.5641', 'epoch': '0.1629'}
+{'loss': '2.176', 'grad_norm': '15.31', 'learning_rate': '8.371e-06', 'num_tokens': '6.145e+05', 'mean_token_accuracy': '0.5164', 'epoch': '0.1631'}
+
 16%|█▋        | 940/5758 [05:18<30:45,  2.61it/s]
 16%|█▋        | 941/5758 [05:18<29:04,  2.76it/s]
                                                  

 16%|█▋        | 941/5758 [05:18<29:04,  2.76it/s]
 16%|█▋        | 942/5758 [05:19<28:25,  2.82it/s]
                                                  

 16%|█▋        | 942/5758 [05:19<28:25,  2.82it/s]
 16%|█▋        | 943/5758 [05:19<28:16,  2.84it/s]
                                                  

 16%|█▋        | 943/5758 [05:19<28:16,  2.84it/s]
 16%|█▋        | 944/5758 [05:19<28:43,  2.79it/s]
                                                  

 16%|█▋        | 944/5758 [05:19<28:43,  2.79it/s]
 16%|█▋        | 945/5758 [05:20<28:03,  2.86it/s]
                                                  

 16%|█▋        | 945/5758 [05:20<28:03,  2.86it/s]
 16%|█▋        | 946/5758 [05:20<27:53,  2.88it/s]
                                                  

 16%|█▋        | 946/5758 [05:20<27:53,  2.88it/s]
 16%|█▋        | 947/5758 [05:20<28:17,  2.83it/s]
                                                  

 16%|█▋        | 947/5758 [05:20<28:17,  2.83it/s]
 16%|█▋        | 948/5758 [05:21<26:58,  2.97it/s]
                                                  

 16%|█▋        | 948/5758 [05:21<26:58,  2.97it/s]
 16%|█▋        | 949/5758 [05:21<25:58,  3.09it/s]
                                                  

 16%|█▋        | 949/5758 [05:21<25:58,  3.09it/s]
 16%|█▋        | 950/5758 [05:21<26:39,  3.01it/s]
                                                  

 16%|█▋        | 950/5758 [05:21<26:39,  3.01it/s]
 17%|█▋        | 951/5758 [05:22<27:22,  2.93it/s]
                                                  

 17%|█▋        | 951/5758 [05:22<27:22,  2.93it/s]
 17%|█▋        | 952/5758 [05:22<27:00,  2.97it/s]
                                                  

 17%|█▋        | 952/5758 [05:22<27:00,  2.97it/s]
 17%|█▋        | 953/5758 [05:22<26:09,  3.06it/s]
                                                  

 17%|█▋        | 953/5758 [05:22<26:09,  3.06it/s]
 17%|█▋        | 954/5758 [05:23<25:26,  3.15it/s]
                                                  

 17%|█▋        | 954/5758 [05:23<25:26,  3.15it/s]
 17%|█▋        | 955/5758 [05:23<25:12,  3.18it/s]
                                                  

 17%|█▋        | 955/5758 [05:23<25:12,  3.18it/s]
 17%|█▋        | 956/5758 [05:23<24:56,  3.21it/s]
                                                  

 17%|█▋        | 956/5758 [05:23<24:56,  3.21it/s]
 17%|█▋        | 957/5758 [05:24<25:16,  3.17it/s]
                                                  

 17%|█▋        | 957/5758 [05:24<25:16,  3.17it/s]
 17%|█▋        | 958/5758 [05:24<24:53,  3.21it/s]
                                                  

 17%|█▋        | 958/5758 [05:24<24:53,  3.21it/s]
 17%|█▋        | 959/5758 [05:24<26:52,  2.98it/s]
                                                  

 17%|█▋        | 959/5758 [05:24<26:52,  2.98it/s]
 17%|█▋        | 960/5758 [05:25<26:53,  2.97it/s]
                                                  

 17%|█▋        | 960/5758 [05:25<26:53,  2.97it/s]
 17%|█▋        | 961/5758 [05:25<26:28,  3.02it/s]
                                                  

 17%|█▋        | 961/5758 [05:25<26:28,  3.02it/s]
 17%|█▋        | 962/5758 [05:25<26:31,  3.01it/s]
                                                  

 17%|█▋        | 962/5758 [05:25<26:31,  3.01it/s]
 17%|█▋        | 963/5758 [05:26<26:09,  3.06it/s]
                                                  

 17%|█▋        | 963/5758 [05:26<26:09,  3.06it/s]
 17%|█▋        | 964/5758 [05:26<29:06,  2.74it/s]
                                                  

 17%|█▋        | 964/5758 [05:26<29:06,  2.74it/s]
 17%|█▋        | 965/5758 [05:26<28:48,  2.77it/s]
                                                  

 17%|█▋        | 965/5758 [05:26<28:48,  2.77it/s]
 17%|█▋        | 966/5758 [05:27<27:57,  2.86it/s]
                                                  

 17%|█▋        | 966/5758 [05:27<27:57,  2.86it/s]
 17%|█▋        | 967/5758 [05:27<26:41,  2.99it/s]
                                                  

 17%|█▋        | 967/5758 [05:27<26:41,  2.99it/s]
 17%|█▋        | 968/5758 [05:27<26:25,  3.02it/s]
                                                  

 17%|█▋        | 968/5758 [05:27<26:25,  3.02it/s]
 17%|█▋        | 969/5758 [05:28<25:51,  3.09it/s]
                                                  

 17%|█▋        | 969/5758 [05:28<25:51,  3.09it/s]
 17%|█▋        | 970/5758 [05:28<25:58,  3.07it/s]
                                                  

 17%|█▋        | 970/5758 [05:28<25:58,  3.07it/s]
 17%|█▋        | 971/5758 [05:28<26:36,  3.00it/s]
                                                  

 17%|█▋        | 971/5758 [05:28<26:36,  3.00it/s]
 17%|█▋        | 972/5758 [05:29<27:20,  2.92it/s]
                                                  

 17%|█▋        | 972/5758 [05:29<27:20,  2.92it/s]
 17%|█▋        | 973/5758 [05:29<26:39,  2.99it/s]
                                                  

 17%|█▋        | 973/5758 [05:29<26:39,  2.99it/s]
 17%|█▋        | 974/5758 [05:29<27:16,  2.92it/s]
                                                  

 17%|█▋        | 974/5758 [05:29<27:16,  2.92it/s]
 17%|█▋        | 975/5758 [05:30<26:54,  2.96it/s]
                                                  

 17%|█▋        | 975/5758 [05:30<26:54,  2.96it/s]
 17%|█▋        | 976/5758 [05:30<26:43,  2.98it/s]
                                                  

 17%|█▋        | 976/5758 [05:30<26:43,  2.98it/s]
 17%|█▋        | 977/5758 [05:30<27:55,  2.85it/s]
                                                  

 17%|█▋        | 977/5758 [05:30<27:55,  2.85it/s]
 17%|█▋        | 978/5758 [05:31<27:23,  2.91it/s]
                                                  

 17%|█▋        | 978/5758 [05:31<27:23,  2.91it/s]
 17%|█▋        | 979/5758 [05:31<26:41,  2.98it/s]
                                                  

 17%|█▋        | 979/5758 [05:31<26:41,  2.98it/s]
 17%|█▋        | 980/5758 [05:31<27:27,  2.90it/s]
                                                  

 17%|█▋        | 980/5758 [05:31<27:27,  2.90it/s]
 17%|█▋        | 981/5758 [05:32<27:13,  2.92it/s]
                                                  

 17%|█▋        | 981/5758 [05:32<27:13,  2.92it/s]
 17%|█▋        | 982/5758 [05:32<27:26,  2.90it/s]
                                                  

 17%|█▋        | 982/5758 [05:32<27:26,  2.90it/s]
 17%|█▋        | 983/5758 [05:32<27:02,  2.94it/s]
                                                  

 17%|█▋        | 983/5758 [05:32<27:02,  2.94it/s]
 17%|█▋        | 984/5758 [05:33<26:00,  3.06it/s]
                                                  

 17%|█▋        | 984/5758 [05:33<26:00,  3.06it/s]
 17%|█▋        | 985/5758 [05:33<25:36,  3.11it/s]
                                                  

 17%|█▋        | 985/5758 [05:33<25:36,  3.11it/s]
 17%|█▋        | 986/5758 [05:33<25:57,  3.06it/s]
                                                  

 17%|█▋        | 986/5758 [05:33<25:57,  3.06it/s]
 17%|█▋        | 987/5758 [05:34<27:04,  2.94it/s]
                                                  

 17%|█▋        | 987/5758 [05:34<27:04,  2.94it/s]
 17%|█▋        | 988/5758 [05:34<27:42,  2.87it/s]
                                                  

 17%|█▋        | 988/5758 [05:34<27:42,  2.87it/s]
 17%|█▋        | 989/5758 [05:34<27:41,  2.87it/s]
                                                  

 17%|█▋        | 989/5758 [05:34<27:41,  2.87it/s]
 17%|█▋        | 990/5758 [05:35<26:35,  2.99it/s]
                                                  

 17%|█▋        | 990/5758 [05:35<26:35,  2.99it/s]
 17%|█▋        | 991/5758 [05:35<26:24,  3.01it/s]
                                                  

 17%|█▋        | 991/5758 [05:35<26:24,  3.01it/s]
 17%|█▋        | 992/5758 [05:35<28:16,  2.81it/s]
                                                  

 17%|█▋        | 992/5758 [05:36<28:16,  2.81it/s]
 17%|█▋        | 993/5758 [05:36<29:43,  2.67it/s]
                                                  

 17%|█▋        | 993/5758 [05:36<29:43,  2.67it/s]
 17%|█▋        | 994/5758 [05:36<29:36,  2.68it/s]
                                                  

 17%|█▋        | 994/5758 [05:36<29:36,  2.68it/s]
 17%|█▋        | 995/5758 [05:37<28:17,  2.81it/s]
                                                  
{'loss': '1.855', 'grad_norm': '10.25', 'learning_rate': '8.369e-06', 'num_tokens': '6.153e+05', 'mean_token_accuracy': '0.6059', 'epoch': '0.1633'}
+{'loss': '0.9869', 'grad_norm': '12.69', 'learning_rate': '8.367e-06', 'num_tokens': '6.16e+05', 'mean_token_accuracy': '0.7185', 'epoch': '0.1634'}
+{'loss': '2.029', 'grad_norm': '11.25', 'learning_rate': '8.366e-06', 'num_tokens': '6.168e+05', 'mean_token_accuracy': '0.5641', 'epoch': '0.1636'}
+{'loss': '0.9946', 'grad_norm': '20', 'learning_rate': '8.364e-06', 'num_tokens': '6.17e+05', 'mean_token_accuracy': '0.7778', 'epoch': '0.1638'}
+{'loss': '1.547', 'grad_norm': '12.31', 'learning_rate': '8.362e-06', 'num_tokens': '6.18e+05', 'mean_token_accuracy': '0.6091', 'epoch': '0.1639'}
+{'loss': '1.15', 'grad_norm': '11.69', 'learning_rate': '8.361e-06', 'num_tokens': '6.188e+05', 'mean_token_accuracy': '0.6879', 'epoch': '0.1641'}
+{'loss': '1.209', 'grad_norm': '10.44', 'learning_rate': '8.359e-06', 'num_tokens': '6.196e+05', 'mean_token_accuracy': '0.6844', 'epoch': '0.1643'}
+{'loss': '1.446', 'grad_norm': '35.25', 'learning_rate': '8.357e-06', 'num_tokens': '6.197e+05', 'mean_token_accuracy': '0.6887', 'epoch': '0.1645'}
+{'loss': '1.409', 'grad_norm': '18.25', 'learning_rate': '8.355e-06', 'num_tokens': '6.2e+05', 'mean_token_accuracy': '0.6921', 'epoch': '0.1646'}
+{'loss': '0.8986', 'grad_norm': '35.25', 'learning_rate': '8.354e-06', 'num_tokens': '6.201e+05', 'mean_token_accuracy': '0.8095', 'epoch': '0.1648'}
+{'loss': '1.563', 'grad_norm': '10.69', 'learning_rate': '8.352e-06', 'num_tokens': '6.21e+05', 'mean_token_accuracy': '0.6473', 'epoch': '0.165'}
+{'loss': '0.8362', 'grad_norm': '28.38', 'learning_rate': '8.35e-06', 'num_tokens': '6.211e+05', 'mean_token_accuracy': '0.8091', 'epoch': '0.1652'}
+{'loss': '0.6568', 'grad_norm': '7.406', 'learning_rate': '8.348e-06', 'num_tokens': '6.226e+05', 'mean_token_accuracy': '0.8074', 'epoch': '0.1653'}
+{'loss': '1.236', 'grad_norm': '13.19', 'learning_rate': '8.347e-06', 'num_tokens': '6.232e+05', 'mean_token_accuracy': '0.6924', 'epoch': '0.1655'}
+{'loss': '2.19', 'grad_norm': '15.5', 'learning_rate': '8.345e-06', 'num_tokens': '6.238e+05', 'mean_token_accuracy': '0.5824', 'epoch': '0.1657'}
+{'loss': '1.236', 'grad_norm': '37', 'learning_rate': '8.343e-06', 'num_tokens': '6.239e+05', 'mean_token_accuracy': '0.7632', 'epoch': '0.1659'}
+{'loss': '1.325', 'grad_norm': '12.88', 'learning_rate': '8.341e-06', 'num_tokens': '6.246e+05', 'mean_token_accuracy': '0.6462', 'epoch': '0.166'}
+{'loss': '1.252', 'grad_norm': '8.875', 'learning_rate': '8.34e-06', 'num_tokens': '6.26e+05', 'mean_token_accuracy': '0.6991', 'epoch': '0.1662'}
+{'loss': '1.185', 'grad_norm': '14', 'learning_rate': '8.338e-06', 'num_tokens': '6.266e+05', 'mean_token_accuracy': '0.6667', 'epoch': '0.1664'}
+{'loss': '2.046', 'grad_norm': '13.5', 'learning_rate': '8.336e-06', 'num_tokens': '6.272e+05', 'mean_token_accuracy': '0.5486', 'epoch': '0.1666'}
+{'loss': '0.5127', 'grad_norm': '24.88', 'learning_rate': '8.334e-06', 'num_tokens': '6.273e+05', 'mean_token_accuracy': '0.8681', 'epoch': '0.1667'}
+{'loss': '1.272', 'grad_norm': '11.19', 'learning_rate': '8.333e-06', 'num_tokens': '6.281e+05', 'mean_token_accuracy': '0.6636', 'epoch': '0.1669'}
+{'loss': '0.8053', 'grad_norm': '28.12', 'learning_rate': '8.331e-06', 'num_tokens': '6.282e+05', 'mean_token_accuracy': '0.8387', 'epoch': '0.1671'}
+{'loss': '0.6555', 'grad_norm': '11.88', 'learning_rate': '8.329e-06', 'num_tokens': '6.289e+05', 'mean_token_accuracy': '0.8431', 'epoch': '0.1672'}
+{'loss': '1.384', 'grad_norm': '12.94', 'learning_rate': '8.328e-06', 'num_tokens': '6.295e+05', 'mean_token_accuracy': '0.6343', 'epoch': '0.1674'}
+{'loss': '2.513', 'grad_norm': '20.5', 'learning_rate': '8.326e-06', 'num_tokens': '6.3e+05', 'mean_token_accuracy': '0.5352', 'epoch': '0.1676'}
+{'loss': '1.041', 'grad_norm': '29.5', 'learning_rate': '8.324e-06', 'num_tokens': '6.301e+05', 'mean_token_accuracy': '0.7356', 'epoch': '0.1678'}
+{'loss': '0.9197', 'grad_norm': '19.38', 'learning_rate': '8.322e-06', 'num_tokens': '6.304e+05', 'mean_token_accuracy': '0.8073', 'epoch': '0.1679'}
+{'loss': '1.638', 'grad_norm': '12.81', 'learning_rate': '8.321e-06', 'num_tokens': '6.312e+05', 'mean_token_accuracy': '0.5659', 'epoch': '0.1681'}
+{'loss': '1.961', 'grad_norm': '11.81', 'learning_rate': '8.319e-06', 'num_tokens': '6.319e+05', 'mean_token_accuracy': '0.6048', 'epoch': '0.1683'}
+{'loss': '1.406', 'grad_norm': '13.5', 'learning_rate': '8.317e-06', 'num_tokens': '6.326e+05', 'mean_token_accuracy': '0.6104', 'epoch': '0.1685'}
+{'loss': '1.943', 'grad_norm': '17.75', 'learning_rate': '8.315e-06', 'num_tokens': '6.331e+05', 'mean_token_accuracy': '0.5853', 'epoch': '0.1686'}
+{'loss': '0.9088', 'grad_norm': '19', 'learning_rate': '8.314e-06', 'num_tokens': '6.333e+05', 'mean_token_accuracy': '0.7863', 'epoch': '0.1688'}
+{'loss': '1.436', 'grad_norm': '13.25', 'learning_rate': '8.312e-06', 'num_tokens': '6.34e+05', 'mean_token_accuracy': '0.6304', 'epoch': '0.169'}
+{'loss': '1.28', 'grad_norm': '11.31', 'learning_rate': '8.31e-06', 'num_tokens': '6.35e+05', 'mean_token_accuracy': '0.6452', 'epoch': '0.1692'}
+{'loss': '1.16', 'grad_norm': '11.81', 'learning_rate': '8.308e-06', 'num_tokens': '6.358e+05', 'mean_token_accuracy': '0.6806', 'epoch': '0.1693'}
+{'loss': '1.468', 'grad_norm': '13.88', 'learning_rate': '8.307e-06', 'num_tokens': '6.366e+05', 'mean_token_accuracy': '0.594', 'epoch': '0.1695'}
+{'loss': '1.231', 'grad_norm': '15', 'learning_rate': '8.305e-06', 'num_tokens': '6.371e+05', 'mean_token_accuracy': '0.754', 'epoch': '0.1697'}
+{'loss': '1.32', 'grad_norm': '13.31', 'learning_rate': '8.303e-06', 'num_tokens': '6.379e+05', 'mean_token_accuracy': '0.6242', 'epoch': '0.1699'}
+{'loss': '0.7724', 'grad_norm': '7.938', 'learning_rate': '8.301e-06', 'num_tokens': '6.393e+05', 'mean_token_accuracy': '0.7689', 'epoch': '0.17'}
+{'loss': '0.6657', 'grad_norm': '6', 'learning_rate': '8.3e-06', 'num_tokens': '6.414e+05', 'mean_token_accuracy': '0.8011', 'epoch': '0.1702'}
+{'loss': '1.259', 'grad_norm': '11.5', 'learning_rate': '8.298e-06', 'num_tokens': '6.422e+05', 'mean_token_accuracy': '0.6549', 'epoch': '0.1704'}
+{'loss': '1.669', 'grad_norm': '21', 'learning_rate': '8.296e-06', 'num_tokens': '6.424e+05', 'mean_token_accuracy': '0.6711', 'epoch': '0.1705'}
+{'loss': '1.397', 'grad_norm': '12.25', 'learning_rate': '8.295e-06', 'num_tokens': '6.431e+05', 'mean_token_accuracy': '0.6391', 'epoch': '0.1707'}
+{'loss': '1.083', 'grad_norm': '13.69', 'learning_rate': '8.293e-06', 'num_tokens': '6.438e+05', 'mean_token_accuracy': '0.6769', 'epoch': '0.1709'}
+{'loss': '2.28', 'grad_norm': '27.62', 'learning_rate': '8.291e-06', 'num_tokens': '6.439e+05', 'mean_token_accuracy': '0.5543', 'epoch': '0.1711'}
+{'loss': '1.832', 'grad_norm': '14.75', 'learning_rate': '8.289e-06', 'num_tokens': '6.444e+05', 'mean_token_accuracy': '0.6103', 'epoch': '0.1712'}
+{'loss': '2.506', 'grad_norm': '16', 'learning_rate': '8.288e-06', 'num_tokens': '6.449e+05', 'mean_token_accuracy': '0.4861', 'epoch': '0.1714'}
+{'loss': '1.747', 'grad_norm': '11.62', 'learning_rate': '8.286e-06', 'num_tokens': '6.458e+05', 'mean_token_accuracy': '0.5395', 'epoch': '0.1716'}
+{'loss': '1.576', 'grad_norm': '18', 'learning_rate': '8.284e-06', 'num_tokens': '6.463e+05', 'mean_token_accuracy': '0.6429', 'epoch': '0.1718'}
+{'loss': '2.065', 'grad_norm': '12.25', 'learning_rate': '8.282e-06', 'num_tokens': '6.469e+05', 'mean_token_accuracy': '0.5607', 'epoch': '0.1719'}
+{'loss': '1.306', 'grad_norm': '10.94', 'learning_rate': '8.281e-06', 'num_tokens': '6.477e+05', 'mean_token_accuracy': '0.6481', 'epoch': '0.1721'}
+{'loss': '2.015', 'grad_norm': '28.88', 'learning_rate': '8.279e-06', 'num_tokens': '6.479e+05', 'mean_token_accuracy': '0.6149', 'epoch': '0.1723'}
+{'loss': '1.39', 'grad_norm': '12.19', 'learning_rate': '8.277e-06', 'num_tokens': '6.486e+05', 'mean_token_accuracy': '0.6113', 'epoch': '0.1725'}
+{'loss': '1.07', 'grad_norm': '12.56', 'learning_rate': '8.275e-06', 'num_tokens': '6.491e+05', 'mean_token_accuracy': '0.713', 'epoch': '0.1726'}
+
 17%|█▋        | 995/5758 [05:37<28:17,  2.81it/s]
 17%|█▋        | 996/5758 [05:37<27:19,  2.90it/s]
                                                  

 17%|█▋        | 996/5758 [05:37<27:19,  2.90it/s]
 17%|█▋        | 997/5758 [05:37<26:04,  3.04it/s]
                                                  

 17%|█▋        | 997/5758 [05:37<26:04,  3.04it/s]
 17%|█▋        | 998/5758 [05:38<25:37,  3.10it/s]
                                                  

 17%|█▋        | 998/5758 [05:38<25:37,  3.10it/s]
 17%|█▋        | 999/5758 [05:38<25:02,  3.17it/s]
                                                  

 17%|█▋        | 999/5758 [05:38<25:02,  3.17it/s]
 17%|█▋        | 1000/5758 [05:38<24:24,  3.25it/s]
                                                   

 17%|█▋        | 1000/5758 [05:38<24:24,  3.25it/s]{'loss': '1.456', 'grad_norm': '11.62', 'learning_rate': '8.274e-06', 'num_tokens': '6.499e+05', 'mean_token_accuracy': '0.6858', 'epoch': '0.1728'}
+{'loss': '1.28', 'grad_norm': '11.56', 'learning_rate': '8.272e-06', 'num_tokens': '6.507e+05', 'mean_token_accuracy': '0.6624', 'epoch': '0.173'}
+{'loss': '2.052', 'grad_norm': '16.5', 'learning_rate': '8.27e-06', 'num_tokens': '6.513e+05', 'mean_token_accuracy': '0.5884', 'epoch': '0.1732'}
+{'loss': '1.227', 'grad_norm': '11.56', 'learning_rate': '8.268e-06', 'num_tokens': '6.52e+05', 'mean_token_accuracy': '0.663', 'epoch': '0.1733'}
+{'loss': '1.197', 'grad_norm': '13.31', 'learning_rate': '8.267e-06', 'num_tokens': '6.526e+05', 'mean_token_accuracy': '0.6894', 'epoch': '0.1735'}
+{'loss': '1.226', 'grad_norm': '14', 'learning_rate': '8.265e-06', 'num_tokens': '6.533e+05', 'mean_token_accuracy': '0.6585', 'epoch': '0.1737'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.38it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.38it/s]
+
 17%|█▋        | 1001/5758 [05:40<1:06:46,  1.19it/s]
                                                     

 17%|█▋        | 1001/5758 [05:40<1:06:46,  1.19it/s]
 17%|█▋        | 1002/5758 [05:41<55:36,  1.43it/s]  
                                                   

 17%|█▋        | 1002/5758 [05:41<55:36,  1.43it/s]
 17%|█▋        | 1003/5758 [05:41<47:18,  1.68it/s]
                                                   

 17%|█▋        | 1003/5758 [05:41<47:18,  1.68it/s]
 17%|█▋        | 1004/5758 [05:41<40:18,  1.97it/s]
                                                   

 17%|█▋        | 1004/5758 [05:41<40:18,  1.97it/s]
 17%|█▋        | 1005/5758 [05:42<36:40,  2.16it/s]
                                                   

 17%|█▋        | 1005/5758 [05:42<36:40,  2.16it/s]
 17%|█▋        | 1006/5758 [05:42<32:59,  2.40it/s]
                                                   

 17%|█▋        | 1006/5758 [05:42<32:59,  2.40it/s]
 17%|█▋        | 1007/5758 [05:42<31:58,  2.48it/s]
                                                   

 17%|█▋        | 1007/5758 [05:42<31:58,  2.48it/s]
 18%|█▊        | 1008/5758 [05:43<30:45,  2.57it/s]
                                                   

 18%|█▊        | 1008/5758 [05:43<30:45,  2.57it/s]
 18%|█▊        | 1009/5758 [05:43<28:26,  2.78it/s]
                                                   

 18%|█▊        | 1009/5758 [05:43<28:26,  2.78it/s]
 18%|█▊        | 1010/5758 [05:43<26:50,  2.95it/s]
                                                   

 18%|█▊        | 1010/5758 [05:43<26:50,  2.95it/s]
 18%|█▊        | 1011/5758 [05:44<27:48,  2.85it/s]
                                                   

 18%|█▊        | 1011/5758 [05:44<27:48,  2.85it/s]
 18%|█▊        | 1012/5758 [05:44<29:04,  2.72it/s]
                                                   

 18%|█▊        | 1012/5758 [05:44<29:04,  2.72it/s]
 18%|█▊        | 1013/5758 [05:44<27:27,  2.88it/s]
                                                   

 18%|█▊        | 1013/5758 [05:44<27:27,  2.88it/s]
 18%|█▊        | 1014/5758 [05:45<27:37,  2.86it/s]
                                                   

 18%|█▊        | 1014/5758 [05:45<27:37,  2.86it/s]
 18%|█▊        | 1015/5758 [05:45<27:04,  2.92it/s]
                                                   

 18%|█▊        | 1015/5758 [05:45<27:04,  2.92it/s]
 18%|█▊        | 1016/5758 [05:45<27:46,  2.85it/s]
                                                   

 18%|█▊        | 1016/5758 [05:45<27:46,  2.85it/s]
 18%|█▊        | 1017/5758 [05:46<31:17,  2.52it/s]
                                                   

 18%|█▊        | 1017/5758 [05:46<31:17,  2.52it/s]
 18%|█▊        | 1018/5758 [05:46<31:01,  2.55it/s]
                                                   

 18%|█▊        | 1018/5758 [05:46<31:01,  2.55it/s]
 18%|█▊        | 1019/5758 [05:47<29:54,  2.64it/s]
                                                   

 18%|█▊        | 1019/5758 [05:47<29:54,  2.64it/s]
 18%|█▊        | 1020/5758 [05:47<30:19,  2.60it/s]
                                                   

 18%|█▊        | 1020/5758 [05:47<30:19,  2.60it/s]
 18%|█▊        | 1021/5758 [05:47<29:13,  2.70it/s]
                                                   

 18%|█▊        | 1021/5758 [05:47<29:13,  2.70it/s]
 18%|█▊        | 1022/5758 [05:48<28:19,  2.79it/s]
                                                   

 18%|█▊        | 1022/5758 [05:48<28:19,  2.79it/s]
 18%|█▊        | 1023/5758 [05:48<27:17,  2.89it/s]
                                                   

 18%|█▊        | 1023/5758 [05:48<27:17,  2.89it/s]
 18%|█▊        | 1024/5758 [05:48<26:43,  2.95it/s]
                                                   

 18%|█▊        | 1024/5758 [05:48<26:43,  2.95it/s]
 18%|█▊        | 1025/5758 [05:49<25:42,  3.07it/s]
                                                   

 18%|█▊        | 1025/5758 [05:49<25:42,  3.07it/s]
 18%|█▊        | 1026/5758 [05:49<26:08,  3.02it/s]
                                                   

 18%|█▊        | 1026/5758 [05:49<26:08,  3.02it/s]
 18%|█▊        | 1027/5758 [05:49<27:03,  2.91it/s]
                                                   

 18%|█▊        | 1027/5758 [05:49<27:03,  2.91it/s]
 18%|█▊        | 1028/5758 [05:50<26:39,  2.96it/s]
                                                   

 18%|█▊        | 1028/5758 [05:50<26:39,  2.96it/s]
 18%|█▊        | 1029/5758 [05:50<26:12,  3.01it/s]
                                                   

 18%|█▊        | 1029/5758 [05:50<26:12,  3.01it/s]
 18%|█▊        | 1030/5758 [05:50<25:01,  3.15it/s]
                                                   

 18%|█▊        | 1030/5758 [05:50<25:01,  3.15it/s]
 18%|█▊        | 1031/5758 [05:51<24:39,  3.20it/s]
                                                   

 18%|█▊        | 1031/5758 [05:51<24:39,  3.20it/s]
 18%|█▊        | 1032/5758 [05:51<25:07,  3.13it/s]
                                                   

 18%|█▊        | 1032/5758 [05:51<25:07,  3.13it/s]
 18%|█▊        | 1033/5758 [05:51<24:59,  3.15it/s]
                                                   

 18%|█▊        | 1033/5758 [05:51<24:59,  3.15it/s]
 18%|█▊        | 1034/5758 [05:51<24:51,  3.17it/s]
                                                   

 18%|█▊        | 1034/5758 [05:51<24:51,  3.17it/s]
 18%|█▊        | 1035/5758 [05:52<24:21,  3.23it/s]
                                                   

 18%|█▊        | 1035/5758 [05:52<24:21,  3.23it/s]
 18%|█▊        | 1036/5758 [05:52<24:52,  3.16it/s]
                                                   

 18%|█▊        | 1036/5758 [05:52<24:52,  3.16it/s]
 18%|█▊        | 1037/5758 [05:52<26:31,  2.97it/s]
                                                   

 18%|█▊        | 1037/5758 [05:52<26:31,  2.97it/s]
 18%|█▊        | 1038/5758 [05:53<26:48,  2.93it/s]
                                                   

 18%|█▊        | 1038/5758 [05:53<26:48,  2.93it/s]
 18%|█▊        | 1039/5758 [05:53<25:42,  3.06it/s]
                                                   

 18%|█▊        | 1039/5758 [05:53<25:42,  3.06it/s]
 18%|█▊        | 1040/5758 [05:53<25:18,  3.11it/s]
                                                   

 18%|█▊        | 1040/5758 [05:53<25:18,  3.11it/s]
 18%|█▊        | 1041/5758 [05:54<25:48,  3.05it/s]
                                                   

 18%|█▊        | 1041/5758 [05:54<25:48,  3.05it/s]
 18%|█▊        | 1042/5758 [05:54<25:54,  3.03it/s]
                                                   

 18%|█▊        | 1042/5758 [05:54<25:54,  3.03it/s]
 18%|█▊        | 1043/5758 [05:55<28:12,  2.79it/s]
                                                   

 18%|█▊        | 1043/5758 [05:55<28:12,  2.79it/s]
 18%|█▊        | 1044/5758 [05:55<26:51,  2.92it/s]
                                                   

 18%|█▊        | 1044/5758 [05:55<26:51,  2.92it/s]
 18%|█▊        | 1045/5758 [05:55<26:43,  2.94it/s]
                                                   

 18%|█▊        | 1045/5758 [05:55<26:43,  2.94it/s]
 18%|█▊        | 1046/5758 [05:56<26:22,  2.98it/s]
                                                   

 18%|█▊        | 1046/5758 [05:56<26:22,  2.98it/s]
 18%|█▊        | 1047/5758 [05:56<25:32,  3.07it/s]
                                                   

 18%|█▊        | 1047/5758 [05:56<25:32,  3.07it/s]
 18%|█▊        | 1048/5758 [05:56<25:29,  3.08it/s]
                                                   

 18%|█▊        | 1048/5758 [05:56<25:29,  3.08it/s]
 18%|█▊        | 1049/5758 [05:57<27:02,  2.90it/s]
                                                   

 18%|█▊        | 1049/5758 [05:57<27:02,  2.90it/s]
 18%|█▊        | 1050/5758 [05:57<27:35,  2.84it/s]
                                                   

 18%|█▊        | 1050/5758 [05:57<27:35,  2.84it/s]
 18%|█▊        | 1051/5758 [05:57<26:16,  2.99it/s]
                                                   

 18%|█▊        | 1051/5758 [05:57<26:16,  2.99it/s]
 18%|█▊        | 1052/5758 [05:58<26:03,  3.01it/s]
                                                   

 18%|█▊        | 1052/5758 [05:58<26:03,  3.01it/s]
 18%|█▊        | 1053/5758 [05:58<26:14,  2.99it/s]
                                                   

 18%|█▊        | 1053/5758 [05:58<26:14,  2.99it/s]
 18%|█▊        | 1054/5758 [05:58<25:10,  3.11it/s]
                                                   

 18%|█▊        | 1054/5758 [05:58<25:10,  3.11it/s]
 18%|█▊        | 1055/5758 [05:58<24:49,  3.16it/s]
                                                   

 18%|█▊        | 1055/5758 [05:58<24:49,  3.16it/s]
 18%|█▊        | 1056/5758 [05:59<24:16,  3.23it/s]
                                                   
{'loss': '2.253', 'grad_norm': '10.44', 'learning_rate': '8.263e-06', 'num_tokens': '6.541e+05', 'mean_token_accuracy': '0.5466', 'epoch': '0.1738'}
+{'loss': '0.919', 'grad_norm': '17.25', 'learning_rate': '8.262e-06', 'num_tokens': '6.543e+05', 'mean_token_accuracy': '0.8178', 'epoch': '0.174'}
+{'loss': '2.444', 'grad_norm': '20.12', 'learning_rate': '8.26e-06', 'num_tokens': '6.548e+05', 'mean_token_accuracy': '0.5196', 'epoch': '0.1742'}
+{'loss': '1.092', 'grad_norm': '27.38', 'learning_rate': '8.258e-06', 'num_tokens': '6.55e+05', 'mean_token_accuracy': '0.7486', 'epoch': '0.1744'}
+{'loss': '0.9647', 'grad_norm': '12.38', 'learning_rate': '8.256e-06', 'num_tokens': '6.555e+05', 'mean_token_accuracy': '0.7827', 'epoch': '0.1745'}
+{'loss': '0.9602', 'grad_norm': '11.06', 'learning_rate': '8.255e-06', 'num_tokens': '6.562e+05', 'mean_token_accuracy': '0.7346', 'epoch': '0.1747'}
+{'loss': '1.762', 'grad_norm': '17.62', 'learning_rate': '8.253e-06', 'num_tokens': '6.568e+05', 'mean_token_accuracy': '0.6261', 'epoch': '0.1749'}
+{'loss': '1.256', 'grad_norm': '9.875', 'learning_rate': '8.251e-06', 'num_tokens': '6.578e+05', 'mean_token_accuracy': '0.7105', 'epoch': '0.1751'}
+{'loss': '1.937', 'grad_norm': '22.88', 'learning_rate': '8.249e-06', 'num_tokens': '6.579e+05', 'mean_token_accuracy': '0.6258', 'epoch': '0.1752'}
+{'loss': '1.259', 'grad_norm': '13.88', 'learning_rate': '8.248e-06', 'num_tokens': '6.585e+05', 'mean_token_accuracy': '0.6677', 'epoch': '0.1754'}
+{'loss': '2.103', 'grad_norm': '36.75', 'learning_rate': '8.246e-06', 'num_tokens': '6.587e+05', 'mean_token_accuracy': '0.5859', 'epoch': '0.1756'}
+{'loss': '1.374', 'grad_norm': '24', 'learning_rate': '8.244e-06', 'num_tokens': '6.588e+05', 'mean_token_accuracy': '0.6783', 'epoch': '0.1758'}
+{'loss': '1.424', 'grad_norm': '28.75', 'learning_rate': '8.242e-06', 'num_tokens': '6.59e+05', 'mean_token_accuracy': '0.7239', 'epoch': '0.1759'}
+{'loss': '0.5302', 'grad_norm': '6.062', 'learning_rate': '8.241e-06', 'num_tokens': '6.61e+05', 'mean_token_accuracy': '0.8317', 'epoch': '0.1761'}
+{'loss': '1.43', 'grad_norm': '17.12', 'learning_rate': '8.239e-06', 'num_tokens': '6.614e+05', 'mean_token_accuracy': '0.6373', 'epoch': '0.1763'}
+{'loss': '1.352', 'grad_norm': '29.12', 'learning_rate': '8.237e-06', 'num_tokens': '6.615e+05', 'mean_token_accuracy': '0.7377', 'epoch': '0.1765'}
+{'loss': '0.7618', 'grad_norm': '8.625', 'learning_rate': '8.235e-06', 'num_tokens': '6.627e+05', 'mean_token_accuracy': '0.7692', 'epoch': '0.1766'}
+{'loss': '2.017', 'grad_norm': '18.62', 'learning_rate': '8.234e-06', 'num_tokens': '6.632e+05', 'mean_token_accuracy': '0.5761', 'epoch': '0.1768'}
+{'loss': '1.448', 'grad_norm': '21.62', 'learning_rate': '8.232e-06', 'num_tokens': '6.635e+05', 'mean_token_accuracy': '0.6681', 'epoch': '0.177'}
+{'loss': '0.9888', 'grad_norm': '33.25', 'learning_rate': '8.23e-06', 'num_tokens': '6.636e+05', 'mean_token_accuracy': '0.7687', 'epoch': '0.1771'}
+{'loss': '0.5621', 'grad_norm': '6.906', 'learning_rate': '8.229e-06', 'num_tokens': '6.65e+05', 'mean_token_accuracy': '0.8374', 'epoch': '0.1773'}
+{'loss': '0.6701', 'grad_norm': '7.844', 'learning_rate': '8.227e-06', 'num_tokens': '6.664e+05', 'mean_token_accuracy': '0.815', 'epoch': '0.1775'}
+{'loss': '0.961', 'grad_norm': '20', 'learning_rate': '8.225e-06', 'num_tokens': '6.666e+05', 'mean_token_accuracy': '0.816', 'epoch': '0.1777'}
+{'loss': '1.177', 'grad_norm': '9.312', 'learning_rate': '8.223e-06', 'num_tokens': '6.679e+05', 'mean_token_accuracy': '0.6878', 'epoch': '0.1778'}
+{'loss': '1.902', 'grad_norm': '43.25', 'learning_rate': '8.222e-06', 'num_tokens': '6.68e+05', 'mean_token_accuracy': '0.5821', 'epoch': '0.178'}
+{'loss': '1.772', 'grad_norm': '15.88', 'learning_rate': '8.22e-06', 'num_tokens': '6.684e+05', 'mean_token_accuracy': '0.6261', 'epoch': '0.1782'}
+{'loss': '1.452', 'grad_norm': '5.531', 'learning_rate': '8.218e-06', 'num_tokens': '6.725e+05', 'mean_token_accuracy': '0.6601', 'epoch': '0.1784'}
+{'loss': '2.137', 'grad_norm': '21.38', 'learning_rate': '8.216e-06', 'num_tokens': '6.729e+05', 'mean_token_accuracy': '0.5494', 'epoch': '0.1785'}
+{'loss': '0.8082', 'grad_norm': '14.25', 'learning_rate': '8.215e-06', 'num_tokens': '6.733e+05', 'mean_token_accuracy': '0.803', 'epoch': '0.1787'}
+{'loss': '1.421', 'grad_norm': '29.75', 'learning_rate': '8.213e-06', 'num_tokens': '6.735e+05', 'mean_token_accuracy': '0.6887', 'epoch': '0.1789'}
+{'loss': '1.054', 'grad_norm': '11.81', 'learning_rate': '8.211e-06', 'num_tokens': '6.742e+05', 'mean_token_accuracy': '0.7203', 'epoch': '0.1791'}
+{'loss': '1.061', 'grad_norm': '10.75', 'learning_rate': '8.209e-06', 'num_tokens': '6.75e+05', 'mean_token_accuracy': '0.7053', 'epoch': '0.1792'}
+{'loss': '0.8746', 'grad_norm': '9.625', 'learning_rate': '8.208e-06', 'num_tokens': '6.758e+05', 'mean_token_accuracy': '0.7754', 'epoch': '0.1794'}
+{'loss': '1.513', 'grad_norm': '13.31', 'learning_rate': '8.206e-06', 'num_tokens': '6.766e+05', 'mean_token_accuracy': '0.7028', 'epoch': '0.1796'}
+{'loss': '0.8793', 'grad_norm': '18.88', 'learning_rate': '8.204e-06', 'num_tokens': '6.767e+05', 'mean_token_accuracy': '0.8221', 'epoch': '0.1797'}
+{'loss': '1.459', 'grad_norm': '11.81', 'learning_rate': '8.203e-06', 'num_tokens': '6.776e+05', 'mean_token_accuracy': '0.623', 'epoch': '0.1799'}
+{'loss': '2.171', 'grad_norm': '20.25', 'learning_rate': '8.201e-06', 'num_tokens': '6.782e+05', 'mean_token_accuracy': '0.5746', 'epoch': '0.1801'}
+{'loss': '0.9286', 'grad_norm': '16.25', 'learning_rate': '8.199e-06', 'num_tokens': '6.784e+05', 'mean_token_accuracy': '0.7839', 'epoch': '0.1803'}
+{'loss': '1.591', 'grad_norm': '9.438', 'learning_rate': '8.197e-06', 'num_tokens': '6.797e+05', 'mean_token_accuracy': '0.6491', 'epoch': '0.1804'}
+{'loss': '0.6412', 'grad_norm': '7.281', 'learning_rate': '8.196e-06', 'num_tokens': '6.811e+05', 'mean_token_accuracy': '0.8126', 'epoch': '0.1806'}
+{'loss': '1.321', 'grad_norm': '11.5', 'learning_rate': '8.194e-06', 'num_tokens': '6.821e+05', 'mean_token_accuracy': '0.6556', 'epoch': '0.1808'}
+{'loss': '1.82', 'grad_norm': '14.25', 'learning_rate': '8.192e-06', 'num_tokens': '6.825e+05', 'mean_token_accuracy': '0.6224', 'epoch': '0.181'}
+{'loss': '1.207', 'grad_norm': '8.5', 'learning_rate': '8.19e-06', 'num_tokens': '6.839e+05', 'mean_token_accuracy': '0.725', 'epoch': '0.1811'}
+{'loss': '1.267', 'grad_norm': '15', 'learning_rate': '8.189e-06', 'num_tokens': '6.846e+05', 'mean_token_accuracy': '0.676', 'epoch': '0.1813'}
+{'loss': '1.6', 'grad_norm': '11.56', 'learning_rate': '8.187e-06', 'num_tokens': '6.855e+05', 'mean_token_accuracy': '0.5861', 'epoch': '0.1815'}
+{'loss': '1.289', 'grad_norm': '11.88', 'learning_rate': '8.185e-06', 'num_tokens': '6.863e+05', 'mean_token_accuracy': '0.7055', 'epoch': '0.1817'}
+{'loss': '1.778', 'grad_norm': '11.38', 'learning_rate': '8.183e-06', 'num_tokens': '6.87e+05', 'mean_token_accuracy': '0.6026', 'epoch': '0.1818'}
+{'loss': '1.47', 'grad_norm': '12.25', 'learning_rate': '8.182e-06', 'num_tokens': '6.878e+05', 'mean_token_accuracy': '0.6169', 'epoch': '0.182'}
+{'loss': '0.8096', 'grad_norm': '8.812', 'learning_rate': '8.18e-06', 'num_tokens': '6.89e+05', 'mean_token_accuracy': '0.7508', 'epoch': '0.1822'}
+{'loss': '1.734', 'grad_norm': '26.88', 'learning_rate': '8.178e-06', 'num_tokens': '6.893e+05', 'mean_token_accuracy': '0.6543', 'epoch': '0.1824'}
+{'loss': '1.409', 'grad_norm': '13.25', 'learning_rate': '8.176e-06', 'num_tokens': '6.9e+05', 'mean_token_accuracy': '0.6335', 'epoch': '0.1825'}
+{'loss': '1.935', 'grad_norm': '15.56', 'learning_rate': '8.175e-06', 'num_tokens': '6.904e+05', 'mean_token_accuracy': '0.586', 'epoch': '0.1827'}
+{'loss': '1.676', 'grad_norm': '21.38', 'learning_rate': '8.173e-06', 'num_tokens': '6.906e+05', 'mean_token_accuracy': '0.636', 'epoch': '0.1829'}
+{'loss': '1.514', 'grad_norm': '14.62', 'learning_rate': '8.171e-06', 'num_tokens': '6.912e+05', 'mean_token_accuracy': '0.6068', 'epoch': '0.183'}
+{'loss': '0.659', 'grad_norm': '21', 'learning_rate': '8.17e-06', 'num_tokens': '6.914e+05', 'mean_token_accuracy': '0.8757', 'epoch': '0.1832'}
+
 18%|█▊        | 1056/5758 [05:59<24:16,  3.23it/s]
 18%|█▊        | 1057/5758 [05:59<23:51,  3.28it/s]
                                                   

 18%|█▊        | 1057/5758 [05:59<23:51,  3.28it/s]
 18%|█▊        | 1058/5758 [05:59<24:40,  3.18it/s]
                                                   

 18%|█▊        | 1058/5758 [05:59<24:40,  3.18it/s]
 18%|█▊        | 1059/5758 [06:00<26:13,  2.99it/s]
                                                   

 18%|█▊        | 1059/5758 [06:00<26:13,  2.99it/s]
 18%|█▊        | 1060/5758 [06:00<27:28,  2.85it/s]
                                                   

 18%|█▊        | 1060/5758 [06:00<27:28,  2.85it/s]
 18%|█▊        | 1061/5758 [06:00<26:23,  2.97it/s]
                                                   

 18%|█▊        | 1061/5758 [06:00<26:23,  2.97it/s]
 18%|█▊        | 1062/5758 [06:01<26:23,  2.97it/s]
                                                   

 18%|█▊        | 1062/5758 [06:01<26:23,  2.97it/s]
 18%|█▊        | 1063/5758 [06:01<25:28,  3.07it/s]
                                                   

 18%|█▊        | 1063/5758 [06:01<25:28,  3.07it/s]
 18%|█▊        | 1064/5758 [06:01<26:44,  2.93it/s]
                                                   

 18%|█▊        | 1064/5758 [06:01<26:44,  2.93it/s]
 18%|█▊        | 1065/5758 [06:02<27:04,  2.89it/s]
                                                   

 18%|█▊        | 1065/5758 [06:02<27:04,  2.89it/s]
 19%|█▊        | 1066/5758 [06:02<25:28,  3.07it/s]
                                                   

 19%|█▊        | 1066/5758 [06:02<25:28,  3.07it/s]
 19%|█▊        | 1067/5758 [06:02<24:58,  3.13it/s]
                                                   

 19%|█▊        | 1067/5758 [06:02<24:58,  3.13it/s]
 19%|█▊        | 1068/5758 [06:03<24:47,  3.15it/s]
                                                   

 19%|█▊        | 1068/5758 [06:03<24:47,  3.15it/s]
 19%|█▊        | 1069/5758 [06:03<27:18,  2.86it/s]
                                                   

 19%|█▊        | 1069/5758 [06:03<27:18,  2.86it/s]
 19%|█▊        | 1070/5758 [06:03<26:40,  2.93it/s]
                                                   

 19%|█▊        | 1070/5758 [06:03<26:40,  2.93it/s]
 19%|█▊        | 1071/5758 [06:04<25:38,  3.05it/s]
                                                   

 19%|█▊        | 1071/5758 [06:04<25:38,  3.05it/s]
 19%|█▊        | 1072/5758 [06:04<25:09,  3.10it/s]
                                                   

 19%|█▊        | 1072/5758 [06:04<25:09,  3.10it/s]
 19%|█▊        | 1073/5758 [06:04<24:40,  3.16it/s]
                                                   

 19%|█▊        | 1073/5758 [06:04<24:40,  3.16it/s]
 19%|█▊        | 1074/5758 [06:05<24:20,  3.21it/s]
                                                   

 19%|█▊        | 1074/5758 [06:05<24:20,  3.21it/s]
 19%|█▊        | 1075/5758 [06:05<25:53,  3.02it/s]
                                                   

 19%|█▊        | 1075/5758 [06:05<25:53,  3.02it/s]
 19%|█▊        | 1076/5758 [06:05<25:31,  3.06it/s]
                                                   

 19%|█▊        | 1076/5758 [06:05<25:31,  3.06it/s]
 19%|█▊        | 1077/5758 [06:06<25:57,  3.01it/s]
                                                   

 19%|█▊        | 1077/5758 [06:06<25:57,  3.01it/s]
 19%|█▊        | 1078/5758 [06:06<25:20,  3.08it/s]
                                                   

 19%|█▊        | 1078/5758 [06:06<25:20,  3.08it/s]
 19%|█▊        | 1079/5758 [06:06<25:37,  3.04it/s]
                                                   

 19%|█▊        | 1079/5758 [06:06<25:37,  3.04it/s]
 19%|█▉        | 1080/5758 [06:07<25:50,  3.02it/s]
                                                   

 19%|█▉        | 1080/5758 [06:07<25:50,  3.02it/s]
 19%|█▉        | 1081/5758 [06:07<25:02,  3.11it/s]
                                                   

 19%|█▉        | 1081/5758 [06:07<25:02,  3.11it/s]
 19%|█▉        | 1082/5758 [06:07<24:27,  3.19it/s]
                                                   

 19%|█▉        | 1082/5758 [06:07<24:27,  3.19it/s]
 19%|█▉        | 1083/5758 [06:08<25:54,  3.01it/s]
                                                   

 19%|█▉        | 1083/5758 [06:08<25:54,  3.01it/s]
 19%|█▉        | 1084/5758 [06:08<25:03,  3.11it/s]
                                                   

 19%|█▉        | 1084/5758 [06:08<25:03,  3.11it/s]
 19%|█▉        | 1085/5758 [06:08<25:00,  3.11it/s]
                                                   

 19%|█▉        | 1085/5758 [06:08<25:00,  3.11it/s]
 19%|█▉        | 1086/5758 [06:09<26:24,  2.95it/s]
                                                   

 19%|█▉        | 1086/5758 [06:09<26:24,  2.95it/s]
 19%|█▉        | 1087/5758 [06:09<27:11,  2.86it/s]
                                                   

 19%|█▉        | 1087/5758 [06:09<27:11,  2.86it/s]
 19%|█▉        | 1088/5758 [06:09<25:57,  3.00it/s]
                                                   

 19%|█▉        | 1088/5758 [06:09<25:57,  3.00it/s]
 19%|█▉        | 1089/5758 [06:10<25:04,  3.10it/s]
                                                   

 19%|█▉        | 1089/5758 [06:10<25:04,  3.10it/s]
 19%|█▉        | 1090/5758 [06:10<25:00,  3.11it/s]
                                                   

 19%|█▉        | 1090/5758 [06:10<25:00,  3.11it/s]
 19%|█▉        | 1091/5758 [06:10<26:33,  2.93it/s]
                                                   

 19%|█▉        | 1091/5758 [06:10<26:33,  2.93it/s]
 19%|█▉        | 1092/5758 [06:11<26:55,  2.89it/s]
                                                   

 19%|█▉        | 1092/5758 [06:11<26:55,  2.89it/s]
 19%|█▉        | 1093/5758 [06:11<26:59,  2.88it/s]
                                                   

 19%|█▉        | 1093/5758 [06:11<26:59,  2.88it/s]
 19%|█▉        | 1094/5758 [06:11<26:05,  2.98it/s]
                                                   

 19%|█▉        | 1094/5758 [06:11<26:05,  2.98it/s]
 19%|█▉        | 1095/5758 [06:12<28:44,  2.70it/s]
                                                   

 19%|█▉        | 1095/5758 [06:12<28:44,  2.70it/s]
 19%|█▉        | 1096/5758 [06:12<27:58,  2.78it/s]
                                                   

 19%|█▉        | 1096/5758 [06:12<27:58,  2.78it/s]
 19%|█▉        | 1097/5758 [06:13<28:45,  2.70it/s]
                                                   

 19%|█▉        | 1097/5758 [06:13<28:45,  2.70it/s]
 19%|█▉        | 1098/5758 [06:13<26:57,  2.88it/s]
                                                   

 19%|█▉        | 1098/5758 [06:13<26:57,  2.88it/s]
 19%|█▉        | 1099/5758 [06:13<27:17,  2.84it/s]
                                                   

 19%|█▉        | 1099/5758 [06:13<27:17,  2.84it/s]
 19%|█▉        | 1100/5758 [06:13<25:59,  2.99it/s]
                                                   

 19%|█▉        | 1100/5758 [06:13<25:59,  2.99it/s]
 19%|█▉        | 1101/5758 [06:14<26:58,  2.88it/s]
                                                   

 19%|█▉        | 1101/5758 [06:14<26:58,  2.88it/s]
 19%|█▉        | 1102/5758 [06:14<26:13,  2.96it/s]
                                                   

 19%|█▉        | 1102/5758 [06:14<26:13,  2.96it/s]
 19%|█▉        | 1103/5758 [06:14<25:37,  3.03it/s]
                                                   

 19%|█▉        | 1103/5758 [06:14<25:37,  3.03it/s]
 19%|█▉        | 1104/5758 [06:15<26:00,  2.98it/s]
                                                   

 19%|█▉        | 1104/5758 [06:15<26:00,  2.98it/s]
 19%|█▉        | 1105/5758 [06:15<26:28,  2.93it/s]
                                                   

 19%|█▉        | 1105/5758 [06:15<26:28,  2.93it/s]
 19%|█▉        | 1106/5758 [06:16<26:21,  2.94it/s]
                                                   

 19%|█▉        | 1106/5758 [06:16<26:21,  2.94it/s]
 19%|█▉        | 1107/5758 [06:16<25:43,  3.01it/s]
                                                   

 19%|█▉        | 1107/5758 [06:16<25:43,  3.01it/s]
 19%|█▉        | 1108/5758 [06:16<24:54,  3.11it/s]
                                                   

 19%|█▉        | 1108/5758 [06:16<24:54,  3.11it/s]
 19%|█▉        | 1109/5758 [06:17<26:13,  2.95it/s]
                                                   

 19%|█▉        | 1109/5758 [06:17<26:13,  2.95it/s]
 19%|█▉        | 1110/5758 [06:17<25:41,  3.02it/s]
                                                   

 19%|█▉        | 1110/5758 [06:17<25:41,  3.02it/s]
 19%|█▉        | 1111/5758 [06:17<25:19,  3.06it/s]
                                                   
{'loss': '1.228', 'grad_norm': '13.06', 'learning_rate': '8.168e-06', 'num_tokens': '6.92e+05', 'mean_token_accuracy': '0.6497', 'epoch': '0.1834'}
+{'loss': '1.403', 'grad_norm': '13.12', 'learning_rate': '8.166e-06', 'num_tokens': '6.926e+05', 'mean_token_accuracy': '0.6516', 'epoch': '0.1836'}
+{'loss': '1.386', 'grad_norm': '33.5', 'learning_rate': '8.164e-06', 'num_tokens': '6.927e+05', 'mean_token_accuracy': '0.7059', 'epoch': '0.1837'}
+{'loss': '1.346', 'grad_norm': '8.75', 'learning_rate': '8.163e-06', 'num_tokens': '6.939e+05', 'mean_token_accuracy': '0.6888', 'epoch': '0.1839'}
+{'loss': '0.969', 'grad_norm': '12.44', 'learning_rate': '8.161e-06', 'num_tokens': '6.945e+05', 'mean_token_accuracy': '0.7247', 'epoch': '0.1841'}
+{'loss': '1.387', 'grad_norm': '13.19', 'learning_rate': '8.159e-06', 'num_tokens': '6.952e+05', 'mean_token_accuracy': '0.6416', 'epoch': '0.1843'}
+{'loss': '2.13', 'grad_norm': '26.12', 'learning_rate': '8.157e-06', 'num_tokens': '6.954e+05', 'mean_token_accuracy': '0.5963', 'epoch': '0.1844'}
+{'loss': '2.13', 'grad_norm': '16.5', 'learning_rate': '8.156e-06', 'num_tokens': '6.957e+05', 'mean_token_accuracy': '0.5723', 'epoch': '0.1846'}
+{'loss': '0.7004', 'grad_norm': '10', 'learning_rate': '8.154e-06', 'num_tokens': '6.969e+05', 'mean_token_accuracy': '0.7914', 'epoch': '0.1848'}
+{'loss': '0.9554', 'grad_norm': '31.88', 'learning_rate': '8.152e-06', 'num_tokens': '6.97e+05', 'mean_token_accuracy': '0.7928', 'epoch': '0.185'}
+{'loss': '0.8578', 'grad_norm': '42.5', 'learning_rate': '8.15e-06', 'num_tokens': '6.97e+05', 'mean_token_accuracy': '0.7797', 'epoch': '0.1851'}
+{'loss': '1.076', 'grad_norm': '26.88', 'learning_rate': '8.149e-06', 'num_tokens': '6.972e+05', 'mean_token_accuracy': '0.7738', 'epoch': '0.1853'}
+{'loss': '1.273', 'grad_norm': '12.69', 'learning_rate': '8.147e-06', 'num_tokens': '6.979e+05', 'mean_token_accuracy': '0.6537', 'epoch': '0.1855'}
+{'loss': '1.691', 'grad_norm': '16.5', 'learning_rate': '8.145e-06', 'num_tokens': '6.983e+05', 'mean_token_accuracy': '0.6735', 'epoch': '0.1857'}
+{'loss': '0.8172', 'grad_norm': '23.5', 'learning_rate': '8.143e-06', 'num_tokens': '6.985e+05', 'mean_token_accuracy': '0.8235', 'epoch': '0.1858'}
+{'loss': '1.125', 'grad_norm': '16.88', 'learning_rate': '8.142e-06', 'num_tokens': '6.988e+05', 'mean_token_accuracy': '0.7327', 'epoch': '0.186'}
+{'loss': '1.237', 'grad_norm': '14.62', 'learning_rate': '8.14e-06', 'num_tokens': '6.995e+05', 'mean_token_accuracy': '0.6497', 'epoch': '0.1862'}
+{'loss': '1.142', 'grad_norm': '13.06', 'learning_rate': '8.138e-06', 'num_tokens': '7.002e+05', 'mean_token_accuracy': '0.6933', 'epoch': '0.1863'}
+{'loss': '0.8604', 'grad_norm': '21.75', 'learning_rate': '8.137e-06', 'num_tokens': '7.003e+05', 'mean_token_accuracy': '0.8204', 'epoch': '0.1865'}
+{'loss': '1.15', 'grad_norm': '14.06', 'learning_rate': '8.135e-06', 'num_tokens': '7.009e+05', 'mean_token_accuracy': '0.6582', 'epoch': '0.1867'}
+{'loss': '1.157', 'grad_norm': '11.88', 'learning_rate': '8.133e-06', 'num_tokens': '7.016e+05', 'mean_token_accuracy': '0.6895', 'epoch': '0.1869'}
+{'loss': '1.643', 'grad_norm': '14.06', 'learning_rate': '8.131e-06', 'num_tokens': '7.021e+05', 'mean_token_accuracy': '0.6961', 'epoch': '0.187'}
+{'loss': '0.8113', 'grad_norm': '24.75', 'learning_rate': '8.13e-06', 'num_tokens': '7.023e+05', 'mean_token_accuracy': '0.8529', 'epoch': '0.1872'}
+{'loss': '1.1', 'grad_norm': '10.5', 'learning_rate': '8.128e-06', 'num_tokens': '7.031e+05', 'mean_token_accuracy': '0.7323', 'epoch': '0.1874'}
+{'loss': '1.589', 'grad_norm': '14.69', 'learning_rate': '8.126e-06', 'num_tokens': '7.036e+05', 'mean_token_accuracy': '0.6881', 'epoch': '0.1876'}
+{'loss': '1.232', 'grad_norm': '13.44', 'learning_rate': '8.124e-06', 'num_tokens': '7.042e+05', 'mean_token_accuracy': '0.6957', 'epoch': '0.1877'}
+{'loss': '1.135', 'grad_norm': '12.62', 'learning_rate': '8.123e-06', 'num_tokens': '7.049e+05', 'mean_token_accuracy': '0.6813', 'epoch': '0.1879'}
+{'loss': '1.276', 'grad_norm': '6.781', 'learning_rate': '8.121e-06', 'num_tokens': '7.071e+05', 'mean_token_accuracy': '0.6955', 'epoch': '0.1881'}
+{'loss': '1.31', 'grad_norm': '8.812', 'learning_rate': '8.119e-06', 'num_tokens': '7.083e+05', 'mean_token_accuracy': '0.6981', 'epoch': '0.1883'}
+{'loss': '1.891', 'grad_norm': '38.5', 'learning_rate': '8.117e-06', 'num_tokens': '7.084e+05', 'mean_token_accuracy': '0.6092', 'epoch': '0.1884'}
+{'loss': '1.461', 'grad_norm': '16.38', 'learning_rate': '8.116e-06', 'num_tokens': '7.087e+05', 'mean_token_accuracy': '0.6748', 'epoch': '0.1886'}
+{'loss': '1.233', 'grad_norm': '10.06', 'learning_rate': '8.114e-06', 'num_tokens': '7.098e+05', 'mean_token_accuracy': '0.6534', 'epoch': '0.1888'}
+{'loss': '1.113', 'grad_norm': '12.56', 'learning_rate': '8.112e-06', 'num_tokens': '7.104e+05', 'mean_token_accuracy': '0.7011', 'epoch': '0.189'}
+{'loss': '1.27', 'grad_norm': '25.38', 'learning_rate': '8.11e-06', 'num_tokens': '7.106e+05', 'mean_token_accuracy': '0.6975', 'epoch': '0.1891'}
+{'loss': '0.8801', 'grad_norm': '14.06', 'learning_rate': '8.109e-06', 'num_tokens': '7.11e+05', 'mean_token_accuracy': '0.8041', 'epoch': '0.1893'}
+{'loss': '1.594', 'grad_norm': '28.25', 'learning_rate': '8.107e-06', 'num_tokens': '7.111e+05', 'mean_token_accuracy': '0.6567', 'epoch': '0.1895'}
+{'loss': '1.078', 'grad_norm': '10.31', 'learning_rate': '8.105e-06', 'num_tokens': '7.121e+05', 'mean_token_accuracy': '0.6913', 'epoch': '0.1896'}
+{'loss': '0.7171', 'grad_norm': '8.25', 'learning_rate': '8.104e-06', 'num_tokens': '7.14e+05', 'mean_token_accuracy': '0.7813', 'epoch': '0.1898'}
+{'loss': '1.323', 'grad_norm': '11.94', 'learning_rate': '8.102e-06', 'num_tokens': '7.148e+05', 'mean_token_accuracy': '0.6577', 'epoch': '0.19'}
+{'loss': '2.049', 'grad_norm': '11.19', 'learning_rate': '8.1e-06', 'num_tokens': '7.156e+05', 'mean_token_accuracy': '0.5843', 'epoch': '0.1902'}
+{'loss': '1.539', 'grad_norm': '6.188', 'learning_rate': '8.098e-06', 'num_tokens': '7.191e+05', 'mean_token_accuracy': '0.6583', 'epoch': '0.1903'}
+{'loss': '0.6056', 'grad_norm': '7.281', 'learning_rate': '8.097e-06', 'num_tokens': '7.206e+05', 'mean_token_accuracy': '0.8117', 'epoch': '0.1905'}
+{'loss': '2.165', 'grad_norm': '17.62', 'learning_rate': '8.095e-06', 'num_tokens': '7.212e+05', 'mean_token_accuracy': '0.5624', 'epoch': '0.1907'}
+{'loss': '1.467', 'grad_norm': '33.25', 'learning_rate': '8.093e-06', 'num_tokens': '7.213e+05', 'mean_token_accuracy': '0.6991', 'epoch': '0.1909'}
+{'loss': '1.577', 'grad_norm': '10.44', 'learning_rate': '8.091e-06', 'num_tokens': '7.226e+05', 'mean_token_accuracy': '0.6469', 'epoch': '0.191'}
+{'loss': '1.814', 'grad_norm': '13.12', 'learning_rate': '8.09e-06', 'num_tokens': '7.231e+05', 'mean_token_accuracy': '0.6552', 'epoch': '0.1912'}
+{'loss': '0.8366', 'grad_norm': '18.75', 'learning_rate': '8.088e-06', 'num_tokens': '7.233e+05', 'mean_token_accuracy': '0.8075', 'epoch': '0.1914'}
+{'loss': '1.363', 'grad_norm': '17.12', 'learning_rate': '8.086e-06', 'num_tokens': '7.236e+05', 'mean_token_accuracy': '0.6771', 'epoch': '0.1916'}
+{'loss': '1.934', 'grad_norm': '15.75', 'learning_rate': '8.084e-06', 'num_tokens': '7.241e+05', 'mean_token_accuracy': '0.5556', 'epoch': '0.1917'}
+{'loss': '2.029', 'grad_norm': '16', 'learning_rate': '8.083e-06', 'num_tokens': '7.246e+05', 'mean_token_accuracy': '0.5953', 'epoch': '0.1919'}
+{'loss': '1.081', 'grad_norm': '11.38', 'learning_rate': '8.081e-06', 'num_tokens': '7.254e+05', 'mean_token_accuracy': '0.6964', 'epoch': '0.1921'}
+{'loss': '1.485', 'grad_norm': '29.75', 'learning_rate': '8.079e-06', 'num_tokens': '7.256e+05', 'mean_token_accuracy': '0.6648', 'epoch': '0.1923'}
+{'loss': '2.189', 'grad_norm': '8.375', 'learning_rate': '8.077e-06', 'num_tokens': '7.269e+05', 'mean_token_accuracy': '0.5368', 'epoch': '0.1924'}
+{'loss': '1.138', 'grad_norm': '13.75', 'learning_rate': '8.076e-06', 'num_tokens': '7.274e+05', 'mean_token_accuracy': '0.6527', 'epoch': '0.1926'}
+{'loss': '1.268', 'grad_norm': '11.88', 'learning_rate': '8.074e-06', 'num_tokens': '7.281e+05', 'mean_token_accuracy': '0.6676', 'epoch': '0.1928'}
+
 19%|█▉        | 1111/5758 [06:17<25:19,  3.06it/s]
 19%|█▉        | 1112/5758 [06:18<26:35,  2.91it/s]
                                                   

 19%|█▉        | 1112/5758 [06:18<26:35,  2.91it/s]
 19%|█▉        | 1113/5758 [06:18<25:50,  3.00it/s]
                                                   

 19%|█▉        | 1113/5758 [06:18<25:50,  3.00it/s]
 19%|█▉        | 1114/5758 [06:18<26:17,  2.94it/s]
                                                   

 19%|█▉        | 1114/5758 [06:18<26:17,  2.94it/s]
 19%|█▉        | 1115/5758 [06:19<26:53,  2.88it/s]
                                                   

 19%|█▉        | 1115/5758 [06:19<26:53,  2.88it/s]
 19%|█▉        | 1116/5758 [06:19<27:34,  2.81it/s]
                                                   

 19%|█▉        | 1116/5758 [06:19<27:34,  2.81it/s]
 19%|█▉        | 1117/5758 [06:19<29:41,  2.61it/s]
                                                   

 19%|█▉        | 1117/5758 [06:19<29:41,  2.61it/s]
 19%|█▉        | 1118/5758 [06:20<29:24,  2.63it/s]
                                                   

 19%|█▉        | 1118/5758 [06:20<29:24,  2.63it/s]
 19%|█▉        | 1119/5758 [06:20<29:14,  2.64it/s]
                                                   

 19%|█▉        | 1119/5758 [06:20<29:14,  2.64it/s]
 19%|█▉        | 1120/5758 [06:21<29:14,  2.64it/s]
                                                   

 19%|█▉        | 1120/5758 [06:21<29:14,  2.64it/s]
 19%|█▉        | 1121/5758 [06:21<28:23,  2.72it/s]
                                                   

 19%|█▉        | 1121/5758 [06:21<28:23,  2.72it/s]
 19%|█▉        | 1122/5758 [06:21<31:35,  2.45it/s]
                                                   

 19%|█▉        | 1122/5758 [06:21<31:35,  2.45it/s]
 20%|█▉        | 1123/5758 [06:22<31:36,  2.44it/s]
                                                   

 20%|█▉        | 1123/5758 [06:22<31:36,  2.44it/s]
 20%|█▉        | 1124/5758 [06:22<31:15,  2.47it/s]
                                                   

 20%|█▉        | 1124/5758 [06:22<31:15,  2.47it/s]
 20%|█▉        | 1125/5758 [06:23<32:39,  2.36it/s]
                                                   

 20%|█▉        | 1125/5758 [06:23<32:39,  2.36it/s]
 20%|█▉        | 1126/5758 [06:23<31:01,  2.49it/s]
                                                   

 20%|█▉        | 1126/5758 [06:23<31:01,  2.49it/s]
 20%|█▉        | 1127/5758 [06:23<32:16,  2.39it/s]
                                                   

 20%|█▉        | 1127/5758 [06:23<32:16,  2.39it/s]
 20%|█▉        | 1128/5758 [06:24<33:27,  2.31it/s]
                                                   

 20%|█▉        | 1128/5758 [06:24<33:27,  2.31it/s]
 20%|█▉        | 1129/5758 [06:24<31:38,  2.44it/s]
                                                   

 20%|█▉        | 1129/5758 [06:24<31:38,  2.44it/s]
 20%|█▉        | 1130/5758 [06:25<30:57,  2.49it/s]
                                                   

 20%|█▉        | 1130/5758 [06:25<30:57,  2.49it/s]
 20%|█▉        | 1131/5758 [06:25<30:05,  2.56it/s]
                                                   

 20%|█▉        | 1131/5758 [06:25<30:05,  2.56it/s]
 20%|█▉        | 1132/5758 [06:25<29:39,  2.60it/s]
                                                   

 20%|█▉        | 1132/5758 [06:25<29:39,  2.60it/s]
 20%|█▉        | 1133/5758 [06:26<29:29,  2.61it/s]
                                                   

 20%|█▉        | 1133/5758 [06:26<29:29,  2.61it/s]
 20%|█▉        | 1134/5758 [06:26<29:17,  2.63it/s]
                                                   

 20%|█▉        | 1134/5758 [06:26<29:17,  2.63it/s]
 20%|█▉        | 1135/5758 [06:27<30:08,  2.56it/s]
                                                   

 20%|█▉        | 1135/5758 [06:27<30:08,  2.56it/s]
 20%|█▉        | 1136/5758 [06:27<30:02,  2.56it/s]
                                                   

 20%|█▉        | 1136/5758 [06:27<30:02,  2.56it/s]
 20%|█▉        | 1137/5758 [06:27<29:20,  2.62it/s]
                                                   

 20%|█▉        | 1137/5758 [06:27<29:20,  2.62it/s]
 20%|█▉        | 1138/5758 [06:28<29:03,  2.65it/s]
                                                   

 20%|█▉        | 1138/5758 [06:28<29:03,  2.65it/s]
 20%|█▉        | 1139/5758 [06:28<29:58,  2.57it/s]
                                                   

 20%|█▉        | 1139/5758 [06:28<29:58,  2.57it/s]
 20%|█▉        | 1140/5758 [06:28<30:12,  2.55it/s]
                                                   

 20%|█▉        | 1140/5758 [06:28<30:12,  2.55it/s]
 20%|█▉        | 1141/5758 [06:29<32:00,  2.40it/s]
                                                   

 20%|█▉        | 1141/5758 [06:29<32:00,  2.40it/s]
 20%|█▉        | 1142/5758 [06:29<32:38,  2.36it/s]
                                                   

 20%|█▉        | 1142/5758 [06:29<32:38,  2.36it/s]
 20%|█▉        | 1143/5758 [06:30<32:01,  2.40it/s]
                                                   

 20%|█▉        | 1143/5758 [06:30<32:01,  2.40it/s]
 20%|█▉        | 1144/5758 [06:30<30:56,  2.49it/s]
                                                   

 20%|█▉        | 1144/5758 [06:30<30:56,  2.49it/s]
 20%|█▉        | 1145/5758 [06:31<30:38,  2.51it/s]
                                                   

 20%|█▉        | 1145/5758 [06:31<30:38,  2.51it/s]
 20%|█▉        | 1146/5758 [06:31<32:31,  2.36it/s]
                                                   

 20%|█▉        | 1146/5758 [06:31<32:31,  2.36it/s]
 20%|█▉        | 1147/5758 [06:31<32:17,  2.38it/s]
                                                   

 20%|█▉        | 1147/5758 [06:31<32:17,  2.38it/s]
 20%|█▉        | 1148/5758 [06:32<31:20,  2.45it/s]
                                                   

 20%|█▉        | 1148/5758 [06:32<31:20,  2.45it/s]
 20%|█▉        | 1149/5758 [06:32<34:26,  2.23it/s]
                                                   

 20%|█▉        | 1149/5758 [06:32<34:26,  2.23it/s]
 20%|█▉        | 1150/5758 [06:33<33:59,  2.26it/s]
                                                   

 20%|█▉        | 1150/5758 [06:33<33:59,  2.26it/s]
 20%|█▉        | 1151/5758 [06:33<32:01,  2.40it/s]
                                                   

 20%|█▉        | 1151/5758 [06:33<32:01,  2.40it/s]
 20%|██        | 1152/5758 [06:34<32:59,  2.33it/s]
                                                   

 20%|██        | 1152/5758 [06:34<32:59,  2.33it/s]
 20%|██        | 1153/5758 [06:34<31:07,  2.47it/s]
                                                   

 20%|██        | 1153/5758 [06:34<31:07,  2.47it/s]
 20%|██        | 1154/5758 [06:34<32:18,  2.37it/s]
                                                   

 20%|██        | 1154/5758 [06:34<32:18,  2.37it/s]
 20%|██        | 1155/5758 [06:35<31:30,  2.44it/s]
                                                   

 20%|██        | 1155/5758 [06:35<31:30,  2.44it/s]
 20%|██        | 1156/5758 [06:35<32:55,  2.33it/s]
                                                   

 20%|██        | 1156/5758 [06:35<32:55,  2.33it/s]
 20%|██        | 1157/5758 [06:36<31:47,  2.41it/s]
                                                   

 20%|██        | 1157/5758 [06:36<31:47,  2.41it/s]
 20%|██        | 1158/5758 [06:36<31:00,  2.47it/s]
                                                   

 20%|██        | 1158/5758 [06:36<31:00,  2.47it/s]
 20%|██        | 1159/5758 [06:36<29:40,  2.58it/s]
                                                   

 20%|██        | 1159/5758 [06:36<29:40,  2.58it/s]
 20%|██        | 1160/5758 [06:37<31:01,  2.47it/s]
                                                   

 20%|██        | 1160/5758 [06:37<31:01,  2.47it/s]
 20%|██        | 1161/5758 [06:37<30:06,  2.55it/s]
                                                   

 20%|██        | 1161/5758 [06:37<30:06,  2.55it/s]
 20%|██        | 1162/5758 [06:38<30:07,  2.54it/s]
                                                   

 20%|██        | 1162/5758 [06:38<30:07,  2.54it/s]
 20%|██        | 1163/5758 [06:38<29:13,  2.62it/s]
                                                   

 20%|██        | 1163/5758 [06:38<29:13,  2.62it/s]
 20%|██        | 1164/5758 [06:38<28:19,  2.70it/s]
                                                   

 20%|██        | 1164/5758 [06:38<28:19,  2.70it/s]
 20%|██        | 1165/5758 [06:39<28:10,  2.72it/s]
                                                   

 20%|██        | 1165/5758 [06:39<28:10,  2.72it/s]
 20%|██        | 1166/5758 [06:39<30:14,  2.53it/s]
                                                   
{'loss': '0.6943', 'grad_norm': '7.375', 'learning_rate': '8.072e-06', 'num_tokens': '7.296e+05', 'mean_token_accuracy': '0.8174', 'epoch': '0.1929'}
+{'loss': '1.285', 'grad_norm': '16.75', 'learning_rate': '8.071e-06', 'num_tokens': '7.299e+05', 'mean_token_accuracy': '0.7462', 'epoch': '0.1931'}
+{'loss': '0.6143', 'grad_norm': '7.656', 'learning_rate': '8.069e-06', 'num_tokens': '7.312e+05', 'mean_token_accuracy': '0.8153', 'epoch': '0.1933'}
+{'loss': '2.11', 'grad_norm': '11.44', 'learning_rate': '8.067e-06', 'num_tokens': '7.32e+05', 'mean_token_accuracy': '0.5889', 'epoch': '0.1935'}
+{'loss': '0.9689', 'grad_norm': '11.62', 'learning_rate': '8.065e-06', 'num_tokens': '7.327e+05', 'mean_token_accuracy': '0.6996', 'epoch': '0.1936'}
+{'loss': '0.9241', 'grad_norm': '28.12', 'learning_rate': '8.064e-06', 'num_tokens': '7.327e+05', 'mean_token_accuracy': '0.7529', 'epoch': '0.1938'}
+{'loss': '1.33', 'grad_norm': '21', 'learning_rate': '8.062e-06', 'num_tokens': '7.333e+05', 'mean_token_accuracy': '0.6395', 'epoch': '0.194'}
+{'loss': '0.8958', 'grad_norm': '23.25', 'learning_rate': '8.06e-06', 'num_tokens': '7.334e+05', 'mean_token_accuracy': '0.767', 'epoch': '0.1942'}
+{'loss': '1.498', 'grad_norm': '12.19', 'learning_rate': '8.058e-06', 'num_tokens': '7.342e+05', 'mean_token_accuracy': '0.6051', 'epoch': '0.1943'}
+{'loss': '2.025', 'grad_norm': '11.75', 'learning_rate': '8.057e-06', 'num_tokens': '7.349e+05', 'mean_token_accuracy': '0.5878', 'epoch': '0.1945'}
+{'loss': '1.067', 'grad_norm': '24.88', 'learning_rate': '8.055e-06', 'num_tokens': '7.351e+05', 'mean_token_accuracy': '0.8289', 'epoch': '0.1947'}
+{'loss': '1.636', 'grad_norm': '11.12', 'learning_rate': '8.053e-06', 'num_tokens': '7.358e+05', 'mean_token_accuracy': '0.6355', 'epoch': '0.1949'}
+{'loss': '0.8352', 'grad_norm': '30', 'learning_rate': '8.051e-06', 'num_tokens': '7.359e+05', 'mean_token_accuracy': '0.7778', 'epoch': '0.195'}
+{'loss': '1.084', 'grad_norm': '7.938', 'learning_rate': '8.05e-06', 'num_tokens': '7.376e+05', 'mean_token_accuracy': '0.7419', 'epoch': '0.1952'}
+{'loss': '1.906', 'grad_norm': '20.75', 'learning_rate': '8.048e-06', 'num_tokens': '7.379e+05', 'mean_token_accuracy': '0.6458', 'epoch': '0.1954'}
+{'loss': '2.076', 'grad_norm': '15.75', 'learning_rate': '8.046e-06', 'num_tokens': '7.385e+05', 'mean_token_accuracy': '0.5735', 'epoch': '0.1956'}
+{'loss': '1.266', 'grad_norm': '19.5', 'learning_rate': '8.044e-06', 'num_tokens': '7.388e+05', 'mean_token_accuracy': '0.6852', 'epoch': '0.1957'}
+{'loss': '1.824', 'grad_norm': '12.19', 'learning_rate': '8.043e-06', 'num_tokens': '7.393e+05', 'mean_token_accuracy': '0.5925', 'epoch': '0.1959'}
+{'loss': '0.503', 'grad_norm': '6.906', 'learning_rate': '8.041e-06', 'num_tokens': '7.406e+05', 'mean_token_accuracy': '0.8501', 'epoch': '0.1961'}
+{'loss': '1.259', 'grad_norm': '12.12', 'learning_rate': '8.039e-06', 'num_tokens': '7.414e+05', 'mean_token_accuracy': '0.6439', 'epoch': '0.1962'}
+{'loss': '0.6813', 'grad_norm': '14.19', 'learning_rate': '8.038e-06', 'num_tokens': '7.417e+05', 'mean_token_accuracy': '0.8423', 'epoch': '0.1964'}
+{'loss': '1.558', 'grad_norm': '11.31', 'learning_rate': '8.036e-06', 'num_tokens': '7.425e+05', 'mean_token_accuracy': '0.6946', 'epoch': '0.1966'}
+{'loss': '0.6416', 'grad_norm': '33.25', 'learning_rate': '8.034e-06', 'num_tokens': '7.425e+05', 'mean_token_accuracy': '0.881', 'epoch': '0.1968'}
+{'loss': '2.094', 'grad_norm': '18.88', 'learning_rate': '8.032e-06', 'num_tokens': '7.429e+05', 'mean_token_accuracy': '0.5819', 'epoch': '0.1969'}
+{'loss': '1.607', 'grad_norm': '11.81', 'learning_rate': '8.031e-06', 'num_tokens': '7.438e+05', 'mean_token_accuracy': '0.6435', 'epoch': '0.1971'}
+{'loss': '1.512', 'grad_norm': '12.44', 'learning_rate': '8.029e-06', 'num_tokens': '7.446e+05', 'mean_token_accuracy': '0.6536', 'epoch': '0.1973'}
+{'loss': '1.223', 'grad_norm': '14.19', 'learning_rate': '8.027e-06', 'num_tokens': '7.453e+05', 'mean_token_accuracy': '0.6626', 'epoch': '0.1975'}
+{'loss': '0.7422', 'grad_norm': '15.56', 'learning_rate': '8.025e-06', 'num_tokens': '7.456e+05', 'mean_token_accuracy': '0.8162', 'epoch': '0.1976'}
+{'loss': '2.034', 'grad_norm': '14', 'learning_rate': '8.024e-06', 'num_tokens': '7.461e+05', 'mean_token_accuracy': '0.5753', 'epoch': '0.1978'}
+{'loss': '0.6695', 'grad_norm': '7.125', 'learning_rate': '8.022e-06', 'num_tokens': '7.477e+05', 'mean_token_accuracy': '0.7931', 'epoch': '0.198'}
+{'loss': '1.101', 'grad_norm': '12.81', 'learning_rate': '8.02e-06', 'num_tokens': '7.482e+05', 'mean_token_accuracy': '0.6814', 'epoch': '0.1982'}
+{'loss': '0.7106', 'grad_norm': '6.75', 'learning_rate': '8.018e-06', 'num_tokens': '7.503e+05', 'mean_token_accuracy': '0.7833', 'epoch': '0.1983'}
+{'loss': '1.945', 'grad_norm': '29.25', 'learning_rate': '8.017e-06', 'num_tokens': '7.505e+05', 'mean_token_accuracy': '0.5941', 'epoch': '0.1985'}
+{'loss': '1.169', 'grad_norm': '16.88', 'learning_rate': '8.015e-06', 'num_tokens': '7.508e+05', 'mean_token_accuracy': '0.745', 'epoch': '0.1987'}
+{'loss': '1.895', 'grad_norm': '43', 'learning_rate': '8.013e-06', 'num_tokens': '7.509e+05', 'mean_token_accuracy': '0.6484', 'epoch': '0.1989'}
+{'loss': '0.8939', 'grad_norm': '13.94', 'learning_rate': '8.011e-06', 'num_tokens': '7.512e+05', 'mean_token_accuracy': '0.8076', 'epoch': '0.199'}
+{'loss': '1.138', 'grad_norm': '14.81', 'learning_rate': '8.01e-06', 'num_tokens': '7.517e+05', 'mean_token_accuracy': '0.7309', 'epoch': '0.1992'}
+{'loss': '3.344', 'grad_norm': '21.12', 'learning_rate': '8.008e-06', 'num_tokens': '7.524e+05', 'mean_token_accuracy': '0.3854', 'epoch': '0.1994'}
+{'loss': '1.953', 'grad_norm': '10.12', 'learning_rate': '8.006e-06', 'num_tokens': '7.533e+05', 'mean_token_accuracy': '0.6022', 'epoch': '0.1995'}
+{'loss': '1.302', 'grad_norm': '18.12', 'learning_rate': '8.005e-06', 'num_tokens': '7.535e+05', 'mean_token_accuracy': '0.7254', 'epoch': '0.1997'}
+{'loss': '0.9016', 'grad_norm': '12.94', 'learning_rate': '8.003e-06', 'num_tokens': '7.542e+05', 'mean_token_accuracy': '0.7322', 'epoch': '0.1999'}
+{'loss': '1.981', 'grad_norm': '28.5', 'learning_rate': '8.001e-06', 'num_tokens': '7.543e+05', 'mean_token_accuracy': '0.5935', 'epoch': '0.2001'}
+{'loss': '1.367', 'grad_norm': '7', 'learning_rate': '7.999e-06', 'num_tokens': '7.569e+05', 'mean_token_accuracy': '0.6761', 'epoch': '0.2002'}
+{'loss': '1.79', 'grad_norm': '19.12', 'learning_rate': '7.998e-06', 'num_tokens': '7.571e+05', 'mean_token_accuracy': '0.6554', 'epoch': '0.2004'}
+{'loss': '0.6141', 'grad_norm': '17.75', 'learning_rate': '7.996e-06', 'num_tokens': '7.573e+05', 'mean_token_accuracy': '0.8673', 'epoch': '0.2006'}
+{'loss': '1.802', 'grad_norm': '11.62', 'learning_rate': '7.994e-06', 'num_tokens': '7.579e+05', 'mean_token_accuracy': '0.6353', 'epoch': '0.2008'}
+{'loss': '1.161', 'grad_norm': '11.5', 'learning_rate': '7.992e-06', 'num_tokens': '7.587e+05', 'mean_token_accuracy': '0.6727', 'epoch': '0.2009'}
+{'loss': '1.193', 'grad_norm': '11.69', 'learning_rate': '7.991e-06', 'num_tokens': '7.594e+05', 'mean_token_accuracy': '0.6768', 'epoch': '0.2011'}
+{'loss': '1.608', 'grad_norm': '9', 'learning_rate': '7.989e-06', 'num_tokens': '7.607e+05', 'mean_token_accuracy': '0.6312', 'epoch': '0.2013'}
+{'loss': '1.061', 'grad_norm': '10.12', 'learning_rate': '7.987e-06', 'num_tokens': '7.617e+05', 'mean_token_accuracy': '0.6982', 'epoch': '0.2015'}
+{'loss': '1.569', 'grad_norm': '26.75', 'learning_rate': '7.985e-06', 'num_tokens': '7.619e+05', 'mean_token_accuracy': '0.7326', 'epoch': '0.2016'}
+{'loss': '1.386', 'grad_norm': '12.38', 'learning_rate': '7.984e-06', 'num_tokens': '7.627e+05', 'mean_token_accuracy': '0.6185', 'epoch': '0.2018'}
+{'loss': '2.475', 'grad_norm': '11.69', 'learning_rate': '7.982e-06', 'num_tokens': '7.64e+05', 'mean_token_accuracy': '0.5008', 'epoch': '0.202'}
+{'loss': '1.774', 'grad_norm': '25.62', 'learning_rate': '7.98e-06', 'num_tokens': '7.642e+05', 'mean_token_accuracy': '0.6242', 'epoch': '0.2022'}
+{'loss': '0.7625', 'grad_norm': '9.188', 'learning_rate': '7.978e-06', 'num_tokens': '7.655e+05', 'mean_token_accuracy': '0.7726', 'epoch': '0.2023'}
+
 20%|██        | 1166/5758 [06:39<30:14,  2.53it/s]
 20%|██        | 1167/5758 [06:40<32:46,  2.34it/s]
                                                   

 20%|██        | 1167/5758 [06:40<32:46,  2.34it/s]
 20%|��█        | 1168/5758 [06:40<31:01,  2.47it/s]
                                                   

 20%|██        | 1168/5758 [06:40<31:01,  2.47it/s]
 20%|██        | 1169/5758 [06:40<30:07,  2.54it/s]
                                                   

 20%|██        | 1169/5758 [06:40<30:07,  2.54it/s]
 20%|██        | 1170/5758 [06:41<30:38,  2.50it/s]
                                                   

 20%|██        | 1170/5758 [06:41<30:38,  2.50it/s]
 20%|██        | 1171/5758 [06:41<29:30,  2.59it/s]
                                                   

 20%|██        | 1171/5758 [06:41<29:30,  2.59it/s]
 20%|██        | 1172/5758 [06:41<28:43,  2.66it/s]
                                                   

 20%|██        | 1172/5758 [06:41<28:43,  2.66it/s]
 20%|██        | 1173/5758 [06:42<30:41,  2.49it/s]
                                                   

 20%|██        | 1173/5758 [06:42<30:41,  2.49it/s]
 20%|██        | 1174/5758 [06:42<32:34,  2.35it/s]
                                                   

 20%|██        | 1174/5758 [06:42<32:34,  2.35it/s]
 20%|██        | 1175/5758 [06:43<31:18,  2.44it/s]
                                                   

 20%|██        | 1175/5758 [06:43<31:18,  2.44it/s]
 20%|██        | 1176/5758 [06:43<29:42,  2.57it/s]
                                                   

 20%|██        | 1176/5758 [06:43<29:42,  2.57it/s]
 20%|██        | 1177/5758 [06:44<30:00,  2.54it/s]
                                                   

 20%|██        | 1177/5758 [06:44<30:00,  2.54it/s]
 20%|██        | 1178/5758 [06:44<30:07,  2.53it/s]
                                                   

 20%|██        | 1178/5758 [06:44<30:07,  2.53it/s]
 20%|██        | 1179/5758 [06:44<29:48,  2.56it/s]
                                                   

 20%|██        | 1179/5758 [06:44<29:48,  2.56it/s]
 20%|██        | 1180/5758 [06:45<30:36,  2.49it/s]
                                                   

 20%|██        | 1180/5758 [06:45<30:36,  2.49it/s]
 21%|██        | 1181/5758 [06:45<30:48,  2.48it/s]
                                                   

 21%|██        | 1181/5758 [06:45<30:48,  2.48it/s]
 21%|██        | 1182/5758 [06:46<31:27,  2.42it/s]
                                                   

 21%|██        | 1182/5758 [06:46<31:27,  2.42it/s]
 21%|██        | 1183/5758 [06:46<31:33,  2.42it/s]
                                                   

 21%|██        | 1183/5758 [06:46<31:33,  2.42it/s]
 21%|██        | 1184/5758 [06:46<30:44,  2.48it/s]
                                                   

 21%|██        | 1184/5758 [06:46<30:44,  2.48it/s]
 21%|██        | 1185/5758 [06:47<29:51,  2.55it/s]
                                                   

 21%|██        | 1185/5758 [06:47<29:51,  2.55it/s]
 21%|██        | 1186/5758 [06:47<29:09,  2.61it/s]
                                                   

 21%|██        | 1186/5758 [06:47<29:09,  2.61it/s]
 21%|██        | 1187/5758 [06:48<31:32,  2.42it/s]
                                                   

 21%|██        | 1187/5758 [06:48<31:32,  2.42it/s]
 21%|██        | 1188/5758 [06:48<30:26,  2.50it/s]
                                                   

 21%|██        | 1188/5758 [06:48<30:26,  2.50it/s]
 21%|██        | 1189/5758 [06:48<30:16,  2.51it/s]
                                                   

 21%|██        | 1189/5758 [06:48<30:16,  2.51it/s]
 21%|██        | 1190/5758 [06:49<30:03,  2.53it/s]
                                                   

 21%|██        | 1190/5758 [06:49<30:03,  2.53it/s]
 21%|██        | 1191/5758 [06:49<28:52,  2.64it/s]
                                                   

 21%|██        | 1191/5758 [06:49<28:52,  2.64it/s]
 21%|██        | 1192/5758 [06:49<28:56,  2.63it/s]
                                                   

 21%|██        | 1192/5758 [06:49<28:56,  2.63it/s]
 21%|██        | 1193/5758 [06:50<29:05,  2.61it/s]
                                                   

 21%|██        | 1193/5758 [06:50<29:05,  2.61it/s]
 21%|██        | 1194/5758 [06:50<28:06,  2.71it/s]
                                                   

 21%|██        | 1194/5758 [06:50<28:06,  2.71it/s]
 21%|██        | 1195/5758 [06:51<27:31,  2.76it/s]
                                                   

 21%|██        | 1195/5758 [06:51<27:31,  2.76it/s]
 21%|██        | 1196/5758 [06:51<27:55,  2.72it/s]
                                                   

 21%|██        | 1196/5758 [06:51<27:55,  2.72it/s]
 21%|██        | 1197/5758 [06:51<29:29,  2.58it/s]
                                                   

 21%|██        | 1197/5758 [06:51<29:29,  2.58it/s]
 21%|██        | 1198/5758 [06:52<27:58,  2.72it/s]
                                                   

 21%|██        | 1198/5758 [06:52<27:58,  2.72it/s]
 21%|██        | 1199/5758 [06:52<27:10,  2.80it/s]
                                                   

 21%|██        | 1199/5758 [06:52<27:10,  2.80it/s]
 21%|██        | 1200/5758 [06:52<27:55,  2.72it/s]
                                                   

 21%|██        | 1200/5758 [06:52<27:55,  2.72it/s]
 21%|██        | 1201/5758 [06:53<26:37,  2.85it/s]
                                                   

 21%|██        | 1201/5758 [06:53<26:37,  2.85it/s]
 21%|██        | 1202/5758 [06:53<26:32,  2.86it/s]
                                                   

 21%|██        | 1202/5758 [06:53<26:32,  2.86it/s]
 21%|██        | 1203/5758 [06:53<25:08,  3.02it/s]
                                                   

 21%|██        | 1203/5758 [06:53<25:08,  3.02it/s]
 21%|██        | 1204/5758 [06:54<24:20,  3.12it/s]
                                                   

 21%|██        | 1204/5758 [06:54<24:20,  3.12it/s]
 21%|██        | 1205/5758 [06:54<24:34,  3.09it/s]
                                                   

 21%|██        | 1205/5758 [06:54<24:34,  3.09it/s]
 21%|██        | 1206/5758 [06:54<24:01,  3.16it/s]
                                                   

 21%|██        | 1206/5758 [06:54<24:01,  3.16it/s]
 21%|██        | 1207/5758 [06:55<23:41,  3.20it/s]
                                                   

 21%|██        | 1207/5758 [06:55<23:41,  3.20it/s]
 21%|██        | 1208/5758 [06:55<23:52,  3.18it/s]
                                                   

 21%|██        | 1208/5758 [06:55<23:52,  3.18it/s]
 21%|██        | 1209/5758 [06:55<24:47,  3.06it/s]
                                                   

 21%|██        | 1209/5758 [06:55<24:47,  3.06it/s]
 21%|██        | 1210/5758 [06:56<25:02,  3.03it/s]
                                                   

 21%|██        | 1210/5758 [06:56<25:02,  3.03it/s]
 21%|██        | 1211/5758 [06:56<26:13,  2.89it/s]
                                                   

 21%|██        | 1211/5758 [06:56<26:13,  2.89it/s]
 21%|██        | 1212/5758 [06:56<25:24,  2.98it/s]
                                                   

 21%|██        | 1212/5758 [06:56<25:24,  2.98it/s]
 21%|██        | 1213/5758 [06:57<25:19,  2.99it/s]
                                                   

 21%|██        | 1213/5758 [06:57<25:19,  2.99it/s]
 21%|██        | 1214/5758 [06:57<24:28,  3.09it/s]
                                                   

 21%|██        | 1214/5758 [06:57<24:28,  3.09it/s]
 21%|██        | 1215/5758 [06:57<23:42,  3.19it/s]
                                                   

 21%|██        | 1215/5758 [06:57<23:42,  3.19it/s]
 21%|██        | 1216/5758 [06:57<23:34,  3.21it/s]
                                                   

 21%|██        | 1216/5758 [06:57<23:34,  3.21it/s]
 21%|██        | 1217/5758 [06:58<24:25,  3.10it/s]
                                                   

 21%|██        | 1217/5758 [06:58<24:25,  3.10it/s]
 21%|██        | 1218/5758 [06:58<24:30,  3.09it/s]
                                                   

 21%|██        | 1218/5758 [06:58<24:30,  3.09it/s]
 21%|██        | 1219/5758 [06:58<24:37,  3.07it/s]
                                                   

 21%|██        | 1219/5758 [06:58<24:37,  3.07it/s]
 21%|██        | 1220/5758 [06:59<23:59,  3.15it/s]
                                                   

 21%|██        | 1220/5758 [06:59<23:59,  3.15it/s]
 21%|██        | 1221/5758 [06:59<23:30,  3.22it/s]
                                                   
{'loss': '2.176', 'grad_norm': '9', 'learning_rate': '7.977e-06', 'num_tokens': '7.666e+05', 'mean_token_accuracy': '0.5617', 'epoch': '0.2025'}
+{'loss': '1.572', 'grad_norm': '32', 'learning_rate': '7.975e-06', 'num_tokens': '7.667e+05', 'mean_token_accuracy': '0.6875', 'epoch': '0.2027'}
+{'loss': '1.879', 'grad_norm': '16.62', 'learning_rate': '7.973e-06', 'num_tokens': '7.673e+05', 'mean_token_accuracy': '0.6052', 'epoch': '0.2028'}
+{'loss': '0.7266', 'grad_norm': '11.88', 'learning_rate': '7.972e-06', 'num_tokens': '7.68e+05', 'mean_token_accuracy': '0.7789', 'epoch': '0.203'}
+{'loss': '1.529', 'grad_norm': '36.75', 'learning_rate': '7.97e-06', 'num_tokens': '7.681e+05', 'mean_token_accuracy': '0.6768', 'epoch': '0.2032'}
+{'loss': '1.309', 'grad_norm': '13.19', 'learning_rate': '7.968e-06', 'num_tokens': '7.688e+05', 'mean_token_accuracy': '0.6347', 'epoch': '0.2034'}
+{'loss': '1.435', 'grad_norm': '15.44', 'learning_rate': '7.966e-06', 'num_tokens': '7.694e+05', 'mean_token_accuracy': '0.7163', 'epoch': '0.2035'}
+{'loss': '1.643', 'grad_norm': '11.19', 'learning_rate': '7.965e-06', 'num_tokens': '7.705e+05', 'mean_token_accuracy': '0.5863', 'epoch': '0.2037'}
+{'loss': '1.422', 'grad_norm': '16.5', 'learning_rate': '7.963e-06', 'num_tokens': '7.708e+05', 'mean_token_accuracy': '0.6918', 'epoch': '0.2039'}
+{'loss': '0.919', 'grad_norm': '11.12', 'learning_rate': '7.961e-06', 'num_tokens': '7.715e+05', 'mean_token_accuracy': '0.786', 'epoch': '0.2041'}
+{'loss': '1.322', 'grad_norm': '12.81', 'learning_rate': '7.959e-06', 'num_tokens': '7.721e+05', 'mean_token_accuracy': '0.6575', 'epoch': '0.2042'}
+{'loss': '1.178', 'grad_norm': '35.25', 'learning_rate': '7.958e-06', 'num_tokens': '7.723e+05', 'mean_token_accuracy': '0.7', 'epoch': '0.2044'}
+{'loss': '1.831', 'grad_norm': '36', 'learning_rate': '7.956e-06', 'num_tokens': '7.724e+05', 'mean_token_accuracy': '0.6277', 'epoch': '0.2046'}
+{'loss': '1.366', 'grad_norm': '11.94', 'learning_rate': '7.954e-06', 'num_tokens': '7.732e+05', 'mean_token_accuracy': '0.6411', 'epoch': '0.2048'}
+{'loss': '1.073', 'grad_norm': '33.75', 'learning_rate': '7.952e-06', 'num_tokens': '7.733e+05', 'mean_token_accuracy': '0.8037', 'epoch': '0.2049'}
+{'loss': '1.709', 'grad_norm': '9.75', 'learning_rate': '7.951e-06', 'num_tokens': '7.742e+05', 'mean_token_accuracy': '0.6558', 'epoch': '0.2051'}
+{'loss': '0.6316', 'grad_norm': '7.719', 'learning_rate': '7.949e-06', 'num_tokens': '7.755e+05', 'mean_token_accuracy': '0.8085', 'epoch': '0.2053'}
+{'loss': '1.85', 'grad_norm': '22.12', 'learning_rate': '7.947e-06', 'num_tokens': '7.757e+05', 'mean_token_accuracy': '0.5974', 'epoch': '0.2055'}
+{'loss': '1.081', 'grad_norm': '10.81', 'learning_rate': '7.945e-06', 'num_tokens': '7.765e+05', 'mean_token_accuracy': '0.6879', 'epoch': '0.2056'}
+{'loss': '1.421', 'grad_norm': '12.12', 'learning_rate': '7.944e-06', 'num_tokens': '7.772e+05', 'mean_token_accuracy': '0.7186', 'epoch': '0.2058'}
+{'loss': '1.041', 'grad_norm': '11.94', 'learning_rate': '7.942e-06', 'num_tokens': '7.779e+05', 'mean_token_accuracy': '0.695', 'epoch': '0.206'}
+{'loss': '2.091', 'grad_norm': '16.75', 'learning_rate': '7.94e-06', 'num_tokens': '7.785e+05', 'mean_token_accuracy': '0.5845', 'epoch': '0.2061'}
+{'loss': '1.313', 'grad_norm': '19.25', 'learning_rate': '7.939e-06', 'num_tokens': '7.786e+05', 'mean_token_accuracy': '0.7102', 'epoch': '0.2063'}
+{'loss': '2.596', 'grad_norm': '14.56', 'learning_rate': '7.937e-06', 'num_tokens': '7.792e+05', 'mean_token_accuracy': '0.4732', 'epoch': '0.2065'}
+{'loss': '2.204', 'grad_norm': '17.25', 'learning_rate': '7.935e-06', 'num_tokens': '7.798e+05', 'mean_token_accuracy': '0.5495', 'epoch': '0.2067'}
+{'loss': '1.412', 'grad_norm': '11.12', 'learning_rate': '7.933e-06', 'num_tokens': '7.807e+05', 'mean_token_accuracy': '0.6327', 'epoch': '0.2068'}
+{'loss': '1.37', 'grad_norm': '18', 'learning_rate': '7.932e-06', 'num_tokens': '7.813e+05', 'mean_token_accuracy': '0.6329', 'epoch': '0.207'}
+{'loss': '1.242', 'grad_norm': '13.25', 'learning_rate': '7.93e-06', 'num_tokens': '7.818e+05', 'mean_token_accuracy': '0.652', 'epoch': '0.2072'}
+{'loss': '1.16', 'grad_norm': '11.06', 'learning_rate': '7.928e-06', 'num_tokens': '7.827e+05', 'mean_token_accuracy': '0.6797', 'epoch': '0.2074'}
+{'loss': '1.821', 'grad_norm': '22.62', 'learning_rate': '7.926e-06', 'num_tokens': '7.83e+05', 'mean_token_accuracy': '0.6123', 'epoch': '0.2075'}
+{'loss': '1.117', 'grad_norm': '12.88', 'learning_rate': '7.925e-06', 'num_tokens': '7.835e+05', 'mean_token_accuracy': '0.6958', 'epoch': '0.2077'}
+{'loss': '2.326', 'grad_norm': '11.06', 'learning_rate': '7.923e-06', 'num_tokens': '7.843e+05', 'mean_token_accuracy': '0.5358', 'epoch': '0.2079'}
+{'loss': '1.338', 'grad_norm': '12.81', 'learning_rate': '7.921e-06', 'num_tokens': '7.851e+05', 'mean_token_accuracy': '0.6115', 'epoch': '0.2081'}
+{'loss': '1.732', 'grad_norm': '9.938', 'learning_rate': '7.919e-06', 'num_tokens': '7.859e+05', 'mean_token_accuracy': '0.6588', 'epoch': '0.2082'}
+{'loss': '1.796', 'grad_norm': '12.69', 'learning_rate': '7.918e-06', 'num_tokens': '7.865e+05', 'mean_token_accuracy': '0.6189', 'epoch': '0.2084'}
+{'loss': '2.612', 'grad_norm': '12.44', 'learning_rate': '7.916e-06', 'num_tokens': '7.873e+05', 'mean_token_accuracy': '0.5204', 'epoch': '0.2086'}
+{'loss': '2.662', 'grad_norm': '22.88', 'learning_rate': '7.914e-06', 'num_tokens': '7.875e+05', 'mean_token_accuracy': '0.5184', 'epoch': '0.2088'}
+{'loss': '0.6572', 'grad_norm': '7.625', 'learning_rate': '7.912e-06', 'num_tokens': '7.887e+05', 'mean_token_accuracy': '0.8109', 'epoch': '0.2089'}
+{'loss': '1.122', 'grad_norm': '12.5', 'learning_rate': '7.911e-06', 'num_tokens': '7.894e+05', 'mean_token_accuracy': '0.7017', 'epoch': '0.2091'}
+{'loss': '1.974', 'grad_norm': '14.56', 'learning_rate': '7.909e-06', 'num_tokens': '7.898e+05', 'mean_token_accuracy': '0.5908', 'epoch': '0.2093'}
+{'loss': '1.774', 'grad_norm': '11.06', 'learning_rate': '7.907e-06', 'num_tokens': '7.905e+05', 'mean_token_accuracy': '0.644', 'epoch': '0.2094'}
+{'loss': '1.001', 'grad_norm': '13', 'learning_rate': '7.906e-06', 'num_tokens': '7.912e+05', 'mean_token_accuracy': '0.7151', 'epoch': '0.2096'}
+{'loss': '1.508', 'grad_norm': '10.94', 'learning_rate': '7.904e-06', 'num_tokens': '7.92e+05', 'mean_token_accuracy': '0.6483', 'epoch': '0.2098'}
+{'loss': '1.475', 'grad_norm': '19.75', 'learning_rate': '7.902e-06', 'num_tokens': '7.922e+05', 'mean_token_accuracy': '0.7273', 'epoch': '0.21'}
+{'loss': '1.019', 'grad_norm': '9.812', 'learning_rate': '7.9e-06', 'num_tokens': '7.931e+05', 'mean_token_accuracy': '0.7553', 'epoch': '0.2101'}
+{'loss': '1.226', 'grad_norm': '16.88', 'learning_rate': '7.899e-06', 'num_tokens': '7.934e+05', 'mean_token_accuracy': '0.737', 'epoch': '0.2103'}
+{'loss': '1.572', 'grad_norm': '25.5', 'learning_rate': '7.897e-06', 'num_tokens': '7.936e+05', 'mean_token_accuracy': '0.663', 'epoch': '0.2105'}
+{'loss': '2.202', 'grad_norm': '10.81', 'learning_rate': '7.895e-06', 'num_tokens': '7.944e+05', 'mean_token_accuracy': '0.5512', 'epoch': '0.2107'}
+{'loss': '0.4744', 'grad_norm': '30.12', 'learning_rate': '7.893e-06', 'num_tokens': '7.945e+05', 'mean_token_accuracy': '0.8649', 'epoch': '0.2108'}
+{'loss': '1.177', 'grad_norm': '13.44', 'learning_rate': '7.892e-06', 'num_tokens': '7.951e+05', 'mean_token_accuracy': '0.6716', 'epoch': '0.211'}
+{'loss': '0.8149', 'grad_norm': '22', 'learning_rate': '7.89e-06', 'num_tokens': '7.953e+05', 'mean_token_accuracy': '0.7869', 'epoch': '0.2112'}
+{'loss': '2.504', 'grad_norm': '26', 'learning_rate': '7.888e-06', 'num_tokens': '7.955e+05', 'mean_token_accuracy': '0.5232', 'epoch': '0.2114'}
+{'loss': '1.292', 'grad_norm': '11.19', 'learning_rate': '7.886e-06', 'num_tokens': '7.964e+05', 'mean_token_accuracy': '0.6567', 'epoch': '0.2115'}
+{'loss': '0.6785', 'grad_norm': '6.875', 'learning_rate': '7.885e-06', 'num_tokens': '7.98e+05', 'mean_token_accuracy': '0.7974', 'epoch': '0.2117'}
+{'loss': '1.921', 'grad_norm': '8.75', 'learning_rate': '7.883e-06', 'num_tokens': '7.993e+05', 'mean_token_accuracy': '0.6011', 'epoch': '0.2119'}
+
 21%|██        | 1221/5758 [06:59<23:30,  3.22it/s]
 21%|██        | 1222/5758 [06:59<23:29,  3.22it/s]
                                                   

 21%|██        | 1222/5758 [06:59<23:29,  3.22it/s]
 21%|██        | 1223/5758 [07:00<26:33,  2.85it/s]
                                                   

 21%|██        | 1223/5758 [07:00<26:33,  2.85it/s]
 21%|██▏       | 1224/5758 [07:00<25:52,  2.92it/s]
                                                   

 21%|██▏       | 1224/5758 [07:00<25:52,  2.92it/s]
 21%|██▏       | 1225/5758 [07:00<24:41,  3.06it/s]
                                                   

 21%|██▏       | 1225/5758 [07:00<24:41,  3.06it/s]
 21%|██▏       | 1226/5758 [07:01<24:04,  3.14it/s]
                                                   

 21%|██▏       | 1226/5758 [07:01<24:04,  3.14it/s]
 21%|██▏       | 1227/5758 [07:01<23:28,  3.22it/s]
                                                   

 21%|██▏       | 1227/5758 [07:01<23:28,  3.22it/s]
 21%|██▏       | 1228/5758 [07:01<24:06,  3.13it/s]
                                                   

 21%|██▏       | 1228/5758 [07:01<24:06,  3.13it/s]
 21%|██▏       | 1229/5758 [07:02<24:18,  3.11it/s]
                                                   

 21%|██▏       | 1229/5758 [07:02<24:18,  3.11it/s]
 21%|██▏       | 1230/5758 [07:02<24:30,  3.08it/s]
                                                   

 21%|██▏       | 1230/5758 [07:02<24:30,  3.08it/s]
 21%|██▏       | 1231/5758 [07:02<25:30,  2.96it/s]
                                                   

 21%|██▏       | 1231/5758 [07:02<25:30,  2.96it/s]
 21%|██▏       | 1232/5758 [07:03<25:31,  2.96it/s]
                                                   

 21%|██▏       | 1232/5758 [07:03<25:31,  2.96it/s]
 21%|██▏       | 1233/5758 [07:03<24:30,  3.08it/s]
                                                   

 21%|██▏       | 1233/5758 [07:03<24:30,  3.08it/s]
 21%|██▏       | 1234/5758 [07:03<25:18,  2.98it/s]
                                                   

 21%|██▏       | 1234/5758 [07:03<25:18,  2.98it/s]
 21%|██▏       | 1235/5758 [07:04<24:22,  3.09it/s]
                                                   

 21%|██▏       | 1235/5758 [07:04<24:22,  3.09it/s]
 21%|██▏       | 1236/5758 [07:04<24:49,  3.04it/s]
                                                   

 21%|██▏       | 1236/5758 [07:04<24:49,  3.04it/s]
 21%|██▏       | 1237/5758 [07:04<24:23,  3.09it/s]
                                                   

 21%|██▏       | 1237/5758 [07:04<24:23,  3.09it/s]
 22%|██▏       | 1238/5758 [07:05<25:34,  2.94it/s]
                                                   

 22%|██▏       | 1238/5758 [07:05<25:34,  2.94it/s]
 22%|██▏       | 1239/5758 [07:05<26:18,  2.86it/s]
                                                   

 22%|██▏       | 1239/5758 [07:05<26:18,  2.86it/s]
 22%|██▏       | 1240/5758 [07:05<25:03,  3.01it/s]
                                                   

 22%|██▏       | 1240/5758 [07:05<25:03,  3.01it/s]
 22%|██▏       | 1241/5758 [07:06<26:02,  2.89it/s]
                                                   

 22%|██▏       | 1241/5758 [07:06<26:02,  2.89it/s]
 22%|██▏       | 1242/5758 [07:06<24:56,  3.02it/s]
                                                   

 22%|██▏       | 1242/5758 [07:06<24:56,  3.02it/s]
 22%|██▏       | 1243/5758 [07:06<24:08,  3.12it/s]
                                                   

 22%|██▏       | 1243/5758 [07:06<24:08,  3.12it/s]
 22%|██▏       | 1244/5758 [07:07<23:36,  3.19it/s]
                                                   

 22%|██▏       | 1244/5758 [07:07<23:36,  3.19it/s]
 22%|██▏       | 1245/5758 [07:07<24:39,  3.05it/s]
                                                   

 22%|██▏       | 1245/5758 [07:07<24:39,  3.05it/s]
 22%|██▏       | 1246/5758 [07:07<25:13,  2.98it/s]
                                                   

 22%|██▏       | 1246/5758 [07:07<25:13,  2.98it/s]
 22%|██▏       | 1247/5758 [07:08<24:27,  3.07it/s]
                                                   

 22%|██▏       | 1247/5758 [07:08<24:27,  3.07it/s]
 22%|██▏       | 1248/5758 [07:08<23:54,  3.14it/s]
                                                   

 22%|██▏       | 1248/5758 [07:08<23:54,  3.14it/s]
 22%|██▏       | 1249/5758 [07:08<26:55,  2.79it/s]
                                                   

 22%|██▏       | 1249/5758 [07:08<26:55,  2.79it/s]
 22%|██▏       | 1250/5758 [07:09<27:05,  2.77it/s]
                                                   

 22%|██▏       | 1250/5758 [07:09<27:05,  2.77it/s]
 22%|██▏       | 1251/5758 [07:09<26:32,  2.83it/s]
                                                   

 22%|██▏       | 1251/5758 [07:09<26:32,  2.83it/s]
 22%|██▏       | 1252/5758 [07:09<25:37,  2.93it/s]
                                                   

 22%|██▏       | 1252/5758 [07:09<25:37,  2.93it/s]
 22%|██▏       | 1253/5758 [07:10<24:30,  3.06it/s]
                                                   

 22%|██▏       | 1253/5758 [07:10<24:30,  3.06it/s]
 22%|██▏       | 1254/5758 [07:10<24:24,  3.08it/s]
                                                   

 22%|██▏       | 1254/5758 [07:10<24:24,  3.08it/s]
 22%|██▏       | 1255/5758 [07:10<25:28,  2.95it/s]
                                                   

 22%|██▏       | 1255/5758 [07:10<25:28,  2.95it/s]
 22%|██▏       | 1256/5758 [07:11<24:36,  3.05it/s]
                                                   

 22%|██▏       | 1256/5758 [07:11<24:36,  3.05it/s]
 22%|██▏       | 1257/5758 [07:11<23:50,  3.15it/s]
                                                   

 22%|██▏       | 1257/5758 [07:11<23:50,  3.15it/s]
 22%|██▏       | 1258/5758 [07:11<23:36,  3.18it/s]
                                                   

 22%|██▏       | 1258/5758 [07:11<23:36,  3.18it/s]
 22%|██▏       | 1259/5758 [07:12<24:19,  3.08it/s]
                                                   

 22%|██▏       | 1259/5758 [07:12<24:19,  3.08it/s]
 22%|██▏       | 1260/5758 [07:12<24:05,  3.11it/s]
                                                   

 22%|██▏       | 1260/5758 [07:12<24:05,  3.11it/s]
 22%|██▏       | 1261/5758 [07:12<24:51,  3.02it/s]
                                                   

 22%|██▏       | 1261/5758 [07:12<24:51,  3.02it/s]
 22%|██▏       | 1262/5758 [07:13<26:03,  2.87it/s]
                                                   

 22%|██▏       | 1262/5758 [07:13<26:03,  2.87it/s]
 22%|██▏       | 1263/5758 [07:13<25:13,  2.97it/s]
                                                   

 22%|██▏       | 1263/5758 [07:13<25:13,  2.97it/s]
 22%|██▏       | 1264/5758 [07:13<24:47,  3.02it/s]
                                                   

 22%|██▏       | 1264/5758 [07:13<24:47,  3.02it/s]
 22%|██▏       | 1265/5758 [07:14<25:27,  2.94it/s]
                                                   

 22%|██▏       | 1265/5758 [07:14<25:27,  2.94it/s]
 22%|██▏       | 1266/5758 [07:14<24:25,  3.06it/s]
                                                   

 22%|██▏       | 1266/5758 [07:14<24:25,  3.06it/s]
 22%|██▏       | 1267/5758 [07:14<23:36,  3.17it/s]
                                                   

 22%|██▏       | 1267/5758 [07:14<23:36,  3.17it/s]
 22%|██▏       | 1268/5758 [07:15<23:03,  3.25it/s]
                                                   

 22%|██▏       | 1268/5758 [07:15<23:03,  3.25it/s]
 22%|██▏       | 1269/5758 [07:15<25:16,  2.96it/s]
                                                   

 22%|██▏       | 1269/5758 [07:15<25:16,  2.96it/s]
 22%|██▏       | 1270/5758 [07:15<24:18,  3.08it/s]
                                                   

 22%|██▏       | 1270/5758 [07:15<24:18,  3.08it/s]
 22%|██▏       | 1271/5758 [07:16<25:47,  2.90it/s]
                                                   

 22%|██▏       | 1271/5758 [07:16<25:47,  2.90it/s]
 22%|██▏       | 1272/5758 [07:16<25:20,  2.95it/s]
                                                   

 22%|██▏       | 1272/5758 [07:16<25:20,  2.95it/s]
 22%|██▏       | 1273/5758 [07:16<24:21,  3.07it/s]
                                                   

 22%|██▏       | 1273/5758 [07:16<24:21,  3.07it/s]
 22%|██▏       | 1274/5758 [07:17<24:18,  3.07it/s]
                                                   

 22%|██▏       | 1274/5758 [07:17<24:18,  3.07it/s]
 22%|██▏       | 1275/5758 [07:17<24:43,  3.02it/s]
                                                   

 22%|██▏       | 1275/5758 [07:17<24:43,  3.02it/s]
 22%|██▏       | 1276/5758 [07:17<26:37,  2.81it/s]
                                                   
{'loss': '0.5576', 'grad_norm': '14.25', 'learning_rate': '7.881e-06', 'num_tokens': '7.996e+05', 'mean_token_accuracy': '0.877', 'epoch': '0.2121'}
+{'loss': '1.302', 'grad_norm': '13.5', 'learning_rate': '7.879e-06', 'num_tokens': '8.003e+05', 'mean_token_accuracy': '0.6289', 'epoch': '0.2122'}
+{'loss': '0.9337', 'grad_norm': '15.12', 'learning_rate': '7.878e-06', 'num_tokens': '8.008e+05', 'mean_token_accuracy': '0.7605', 'epoch': '0.2124'}
+{'loss': '1.143', 'grad_norm': '12.06', 'learning_rate': '7.876e-06', 'num_tokens': '8.015e+05', 'mean_token_accuracy': '0.6944', 'epoch': '0.2126'}
+{'loss': '2.152', 'grad_norm': '18.62', 'learning_rate': '7.874e-06', 'num_tokens': '8.022e+05', 'mean_token_accuracy': '0.5868', 'epoch': '0.2127'}
+{'loss': '1.06', 'grad_norm': '12.44', 'learning_rate': '7.873e-06', 'num_tokens': '8.028e+05', 'mean_token_accuracy': '0.6896', 'epoch': '0.2129'}
+{'loss': '1.269', 'grad_norm': '8.562', 'learning_rate': '7.871e-06', 'num_tokens': '8.041e+05', 'mean_token_accuracy': '0.7002', 'epoch': '0.2131'}
+{'loss': '1.496', 'grad_norm': '14', 'learning_rate': '7.869e-06', 'num_tokens': '8.045e+05', 'mean_token_accuracy': '0.6773', 'epoch': '0.2133'}
+{'loss': '1.498', 'grad_norm': '11.12', 'learning_rate': '7.867e-06', 'num_tokens': '8.053e+05', 'mean_token_accuracy': '0.6022', 'epoch': '0.2134'}
+{'loss': '2.346', 'grad_norm': '19.88', 'learning_rate': '7.866e-06', 'num_tokens': '8.058e+05', 'mean_token_accuracy': '0.5379', 'epoch': '0.2136'}
+{'loss': '2.068', 'grad_norm': '16.25', 'learning_rate': '7.864e-06', 'num_tokens': '8.063e+05', 'mean_token_accuracy': '0.5992', 'epoch': '0.2138'}
+{'loss': '1.773', 'grad_norm': '14.81', 'learning_rate': '7.862e-06', 'num_tokens': '8.067e+05', 'mean_token_accuracy': '0.6496', 'epoch': '0.214'}
+{'loss': '1.177', 'grad_norm': '13.5', 'learning_rate': '7.86e-06', 'num_tokens': '8.074e+05', 'mean_token_accuracy': '0.6804', 'epoch': '0.2141'}
+{'loss': '0.8825', 'grad_norm': '13.81', 'learning_rate': '7.859e-06', 'num_tokens': '8.079e+05', 'mean_token_accuracy': '0.7968', 'epoch': '0.2143'}
+{'loss': '1.129', 'grad_norm': '12.88', 'learning_rate': '7.857e-06', 'num_tokens': '8.085e+05', 'mean_token_accuracy': '0.6672', 'epoch': '0.2145'}
+{'loss': '2.003', 'grad_norm': '13.56', 'learning_rate': '7.855e-06', 'num_tokens': '8.09e+05', 'mean_token_accuracy': '0.5996', 'epoch': '0.2147'}
+{'loss': '1.072', 'grad_norm': '32.5', 'learning_rate': '7.853e-06', 'num_tokens': '8.091e+05', 'mean_token_accuracy': '0.7531', 'epoch': '0.2148'}
+{'loss': '2.167', 'grad_norm': '18.12', 'learning_rate': '7.852e-06', 'num_tokens': '8.096e+05', 'mean_token_accuracy': '0.5887', 'epoch': '0.215'}
+{'loss': '2.522', 'grad_norm': '38.75', 'learning_rate': '7.85e-06', 'num_tokens': '8.098e+05', 'mean_token_accuracy': '0.5242', 'epoch': '0.2152'}
+{'loss': '1.89', 'grad_norm': '26.62', 'learning_rate': '7.848e-06', 'num_tokens': '8.099e+05', 'mean_token_accuracy': '0.6564', 'epoch': '0.2154'}
+{'loss': '1.397', 'grad_norm': '23.5', 'learning_rate': '7.846e-06', 'num_tokens': '8.102e+05', 'mean_token_accuracy': '0.7107', 'epoch': '0.2155'}
+{'loss': '1.091', 'grad_norm': '13', 'learning_rate': '7.845e-06', 'num_tokens': '8.109e+05', 'mean_token_accuracy': '0.6847', 'epoch': '0.2157'}
+{'loss': '1.211', 'grad_norm': '12.5', 'learning_rate': '7.843e-06', 'num_tokens': '8.115e+05', 'mean_token_accuracy': '0.7053', 'epoch': '0.2159'}
+{'loss': '0.5603', 'grad_norm': '7.406', 'learning_rate': '7.841e-06', 'num_tokens': '8.128e+05', 'mean_token_accuracy': '0.8328', 'epoch': '0.216'}
+{'loss': '1.309', 'grad_norm': '21.25', 'learning_rate': '7.84e-06', 'num_tokens': '8.131e+05', 'mean_token_accuracy': '0.754', 'epoch': '0.2162'}
+{'loss': '1.888', 'grad_norm': '13.69', 'learning_rate': '7.838e-06', 'num_tokens': '8.136e+05', 'mean_token_accuracy': '0.6138', 'epoch': '0.2164'}
+{'loss': '1.239', 'grad_norm': '14.25', 'learning_rate': '7.836e-06', 'num_tokens': '8.142e+05', 'mean_token_accuracy': '0.6702', 'epoch': '0.2166'}
+{'loss': '2.016', 'grad_norm': '16', 'learning_rate': '7.834e-06', 'num_tokens': '8.149e+05', 'mean_token_accuracy': '0.5716', 'epoch': '0.2167'}
+{'loss': '1.667', 'grad_norm': '42.5', 'learning_rate': '7.833e-06', 'num_tokens': '8.15e+05', 'mean_token_accuracy': '0.6196', 'epoch': '0.2169'}
+{'loss': '0.8252', 'grad_norm': '18.88', 'learning_rate': '7.831e-06', 'num_tokens': '8.152e+05', 'mean_token_accuracy': '0.8185', 'epoch': '0.2171'}
+{'loss': '0.9668', 'grad_norm': '11.44', 'learning_rate': '7.829e-06', 'num_tokens': '8.161e+05', 'mean_token_accuracy': '0.7357', 'epoch': '0.2173'}
+{'loss': '1.332', 'grad_norm': '12.56', 'learning_rate': '7.827e-06', 'num_tokens': '8.168e+05', 'mean_token_accuracy': '0.6159', 'epoch': '0.2174'}
+{'loss': '0.8456', 'grad_norm': '17', 'learning_rate': '7.826e-06', 'num_tokens': '8.171e+05', 'mean_token_accuracy': '0.8046', 'epoch': '0.2176'}
+{'loss': '1.27', 'grad_norm': '12.75', 'learning_rate': '7.824e-06', 'num_tokens': '8.179e+05', 'mean_token_accuracy': '0.656', 'epoch': '0.2178'}
+{'loss': '1.122', 'grad_norm': '15.31', 'learning_rate': '7.822e-06', 'num_tokens': '8.184e+05', 'mean_token_accuracy': '0.6916', 'epoch': '0.218'}
+{'loss': '0.6776', 'grad_norm': '13.25', 'learning_rate': '7.82e-06', 'num_tokens': '8.191e+05', 'mean_token_accuracy': '0.8323', 'epoch': '0.2181'}
+{'loss': '1.49', 'grad_norm': '14.25', 'learning_rate': '7.819e-06', 'num_tokens': '8.197e+05', 'mean_token_accuracy': '0.6228', 'epoch': '0.2183'}
+{'loss': '1.569', 'grad_norm': '13.19', 'learning_rate': '7.817e-06', 'num_tokens': '8.204e+05', 'mean_token_accuracy': '0.5917', 'epoch': '0.2185'}
+{'loss': '1.119', 'grad_norm': '18.38', 'learning_rate': '7.815e-06', 'num_tokens': '8.207e+05', 'mean_token_accuracy': '0.7551', 'epoch': '0.2187'}
+{'loss': '1.44', 'grad_norm': '8.438', 'learning_rate': '7.813e-06', 'num_tokens': '8.222e+05', 'mean_token_accuracy': '0.6671', 'epoch': '0.2188'}
+{'loss': '2.189', 'grad_norm': '18', 'learning_rate': '7.812e-06', 'num_tokens': '8.227e+05', 'mean_token_accuracy': '0.5364', 'epoch': '0.219'}
+{'loss': '1.799', 'grad_norm': '17.88', 'learning_rate': '7.81e-06', 'num_tokens': '8.23e+05', 'mean_token_accuracy': '0.6203', 'epoch': '0.2192'}
+{'loss': '1.156', 'grad_norm': '8.438', 'learning_rate': '7.808e-06', 'num_tokens': '8.245e+05', 'mean_token_accuracy': '0.7306', 'epoch': '0.2193'}
+{'loss': '0.7805', 'grad_norm': '7.844', 'learning_rate': '7.807e-06', 'num_tokens': '8.26e+05', 'mean_token_accuracy': '0.7831', 'epoch': '0.2195'}
+{'loss': '0.3846', 'grad_norm': '21.88', 'learning_rate': '7.805e-06', 'num_tokens': '8.261e+05', 'mean_token_accuracy': '0.9217', 'epoch': '0.2197'}
+{'loss': '1.294', 'grad_norm': '13.56', 'learning_rate': '7.803e-06', 'num_tokens': '8.267e+05', 'mean_token_accuracy': '0.6541', 'epoch': '0.2199'}
+{'loss': '1.667', 'grad_norm': '26.25', 'learning_rate': '7.801e-06', 'num_tokens': '8.269e+05', 'mean_token_accuracy': '0.6797', 'epoch': '0.22'}
+{'loss': '0.7147', 'grad_norm': '6.25', 'learning_rate': '7.8e-06', 'num_tokens': '8.294e+05', 'mean_token_accuracy': '0.7864', 'epoch': '0.2202'}
+{'loss': '0.6837', 'grad_norm': '20.25', 'learning_rate': '7.798e-06', 'num_tokens': '8.296e+05', 'mean_token_accuracy': '0.8151', 'epoch': '0.2204'}
+{'loss': '1.241', 'grad_norm': '13.06', 'learning_rate': '7.796e-06', 'num_tokens': '8.302e+05', 'mean_token_accuracy': '0.6581', 'epoch': '0.2206'}
+{'loss': '2.581', 'grad_norm': '10', 'learning_rate': '7.794e-06', 'num_tokens': '8.314e+05', 'mean_token_accuracy': '0.5055', 'epoch': '0.2207'}
+{'loss': '1.101', 'grad_norm': '23.88', 'learning_rate': '7.793e-06', 'num_tokens': '8.316e+05', 'mean_token_accuracy': '0.73', 'epoch': '0.2209'}
+{'loss': '2.124', 'grad_norm': '18.88', 'learning_rate': '7.791e-06', 'num_tokens': '8.319e+05', 'mean_token_accuracy': '0.5861', 'epoch': '0.2211'}
+{'loss': '1.546', 'grad_norm': '11.5', 'learning_rate': '7.789e-06', 'num_tokens': '8.327e+05', 'mean_token_accuracy': '0.6446', 'epoch': '0.2213'}
+{'loss': '1.358', 'grad_norm': '11.44', 'learning_rate': '7.787e-06', 'num_tokens': '8.336e+05', 'mean_token_accuracy': '0.6278', 'epoch': '0.2214'}
+
 22%|██▏       | 1276/5758 [07:17<26:37,  2.81it/s]
 22%|██▏       | 1277/5758 [07:18<27:07,  2.75it/s]
                                                   

 22%|██▏       | 1277/5758 [07:18<27:07,  2.75it/s]
 22%|██▏       | 1278/5758 [07:18<25:43,  2.90it/s]
                                                   

 22%|██▏       | 1278/5758 [07:18<25:43,  2.90it/s]
 22%|██▏       | 1279/5758 [07:18<24:59,  2.99it/s]
                                                   

 22%|██▏       | 1279/5758 [07:18<24:59,  2.99it/s]
 22%|██▏       | 1280/5758 [07:19<23:57,  3.12it/s]
                                                   

 22%|██▏       | 1280/5758 [07:19<23:57,  3.12it/s]
 22%|██▏       | 1281/5758 [07:19<23:25,  3.18it/s]
                                                   

 22%|██▏       | 1281/5758 [07:19<23:25,  3.18it/s]
 22%|██▏       | 1282/5758 [07:19<23:47,  3.13it/s]
                                                   

 22%|██▏       | 1282/5758 [07:19<23:47,  3.13it/s]
 22%|██▏       | 1283/5758 [07:20<24:29,  3.05it/s]
                                                   

 22%|██▏       | 1283/5758 [07:20<24:29,  3.05it/s]
 22%|██▏       | 1284/5758 [07:20<25:12,  2.96it/s]
                                                   

 22%|██▏       | 1284/5758 [07:20<25:12,  2.96it/s]
 22%|██▏       | 1285/5758 [07:20<24:19,  3.07it/s]
                                                   

 22%|██▏       | 1285/5758 [07:20<24:19,  3.07it/s]
 22%|██▏       | 1286/5758 [07:21<25:36,  2.91it/s]
                                                   

 22%|██▏       | 1286/5758 [07:21<25:36,  2.91it/s]
 22%|██▏       | 1287/5758 [07:21<24:22,  3.06it/s]
                                                   

 22%|██▏       | 1287/5758 [07:21<24:22,  3.06it/s]
 22%|██▏       | 1288/5758 [07:21<23:25,  3.18it/s]
                                                   

 22%|██▏       | 1288/5758 [07:21<23:25,  3.18it/s]
 22%|██▏       | 1289/5758 [07:22<23:08,  3.22it/s]
                                                   

 22%|██▏       | 1289/5758 [07:22<23:08,  3.22it/s]
 22%|██▏       | 1290/5758 [07:22<23:04,  3.23it/s]
                                                   

 22%|██▏       | 1290/5758 [07:22<23:04,  3.23it/s]
 22%|██▏       | 1291/5758 [07:22<24:28,  3.04it/s]
                                                   

 22%|██▏       | 1291/5758 [07:22<24:28,  3.04it/s]
 22%|██▏       | 1292/5758 [07:23<24:02,  3.10it/s]
                                                   

 22%|██▏       | 1292/5758 [07:23<24:02,  3.10it/s]
 22%|██▏       | 1293/5758 [07:23<24:36,  3.03it/s]
                                                   

 22%|██▏       | 1293/5758 [07:23<24:36,  3.03it/s]
 22%|██▏       | 1294/5758 [07:23<24:08,  3.08it/s]
                                                   

 22%|██▏       | 1294/5758 [07:23<24:08,  3.08it/s]
 22%|██▏       | 1295/5758 [07:24<23:40,  3.14it/s]
                                                   

 22%|██▏       | 1295/5758 [07:24<23:40,  3.14it/s]
 23%|██▎       | 1296/5758 [07:24<25:02,  2.97it/s]
                                                   

 23%|██▎       | 1296/5758 [07:24<25:02,  2.97it/s]
 23%|██▎       | 1297/5758 [07:24<25:45,  2.89it/s]
                                                   

 23%|██▎       | 1297/5758 [07:24<25:45,  2.89it/s]
 23%|██▎       | 1298/5758 [07:25<24:31,  3.03it/s]
                                                   

 23%|██▎       | 1298/5758 [07:25<24:31,  3.03it/s]
 23%|██▎       | 1299/5758 [07:25<25:10,  2.95it/s]
                                                   

 23%|██▎       | 1299/5758 [07:25<25:10,  2.95it/s]
 23%|██▎       | 1300/5758 [07:25<25:22,  2.93it/s]
                                                   

 23%|██▎       | 1300/5758 [07:25<25:22,  2.93it/s]
 23%|██▎       | 1301/5758 [07:26<24:54,  2.98it/s]
                                                   

 23%|██▎       | 1301/5758 [07:26<24:54,  2.98it/s]
 23%|██▎       | 1302/5758 [07:26<24:43,  3.00it/s]
                                                   

 23%|██▎       | 1302/5758 [07:26<24:43,  3.00it/s]
 23%|██▎       | 1303/5758 [07:26<26:42,  2.78it/s]
                                                   

 23%|██▎       | 1303/5758 [07:26<26:42,  2.78it/s]
 23%|██▎       | 1304/5758 [07:27<26:11,  2.83it/s]
                                                   

 23%|██▎       | 1304/5758 [07:27<26:11,  2.83it/s]
 23%|██▎       | 1305/5758 [07:27<25:04,  2.96it/s]
                                                   

 23%|██▎       | 1305/5758 [07:27<25:04,  2.96it/s]
 23%|██▎       | 1306/5758 [07:27<24:06,  3.08it/s]
                                                   

 23%|██▎       | 1306/5758 [07:27<24:06,  3.08it/s]
 23%|██▎       | 1307/5758 [07:28<24:36,  3.02it/s]
                                                   

 23%|██▎       | 1307/5758 [07:28<24:36,  3.02it/s]
 23%|██▎       | 1308/5758 [07:28<24:39,  3.01it/s]
                                                   

 23%|██▎       | 1308/5758 [07:28<24:39,  3.01it/s]
 23%|██▎       | 1309/5758 [07:28<23:44,  3.12it/s]
                                                   

 23%|██▎       | 1309/5758 [07:28<23:44,  3.12it/s]
 23%|██▎       | 1310/5758 [07:29<23:56,  3.10it/s]
                                                   

 23%|██▎       | 1310/5758 [07:29<23:56,  3.10it/s]
 23%|██▎       | 1311/5758 [07:29<23:48,  3.11it/s]
                                                   

 23%|██▎       | 1311/5758 [07:29<23:48,  3.11it/s]
 23%|██▎       | 1312/5758 [07:29<23:34,  3.14it/s]
                                                   

 23%|██▎       | 1312/5758 [07:29<23:34,  3.14it/s]
 23%|██▎       | 1313/5758 [07:30<22:58,  3.22it/s]
                                                   

 23%|██▎       | 1313/5758 [07:30<22:58,  3.22it/s]
 23%|██▎       | 1314/5758 [07:30<23:38,  3.13it/s]
                                                   

 23%|██▎       | 1314/5758 [07:30<23:38,  3.13it/s]
 23%|██▎       | 1315/5758 [07:30<24:59,  2.96it/s]
                                                   

 23%|██▎       | 1315/5758 [07:30<24:59,  2.96it/s]
 23%|██▎       | 1316/5758 [07:31<25:22,  2.92it/s]
                                                   

 23%|██▎       | 1316/5758 [07:31<25:22,  2.92it/s]
 23%|██▎       | 1317/5758 [07:31<24:37,  3.01it/s]
                                                   

 23%|██▎       | 1317/5758 [07:31<24:37,  3.01it/s]
 23%|██▎       | 1318/5758 [07:31<25:59,  2.85it/s]
                                                   

 23%|██▎       | 1318/5758 [07:31<25:59,  2.85it/s]
 23%|██▎       | 1319/5758 [07:32<26:23,  2.80it/s]
                                                   

 23%|██▎       | 1319/5758 [07:32<26:23,  2.80it/s]
 23%|██▎       | 1320/5758 [07:32<28:35,  2.59it/s]
                                                   

 23%|██▎       | 1320/5758 [07:32<28:35,  2.59it/s]
 23%|██▎       | 1321/5758 [07:32<26:48,  2.76it/s]
                                                   

 23%|██▎       | 1321/5758 [07:32<26:48,  2.76it/s]
 23%|██▎       | 1322/5758 [07:33<26:41,  2.77it/s]
                                                   

 23%|██▎       | 1322/5758 [07:33<26:41,  2.77it/s]
 23%|██▎       | 1323/5758 [07:33<25:51,  2.86it/s]
                                                   

 23%|██▎       | 1323/5758 [07:33<25:51,  2.86it/s]
 23%|██▎       | 1324/5758 [07:33<25:23,  2.91it/s]
                                                   

 23%|██▎       | 1324/5758 [07:33<25:23,  2.91it/s]
 23%|██▎       | 1325/5758 [07:34<24:07,  3.06it/s]
                                                   

 23%|██▎       | 1325/5758 [07:34<24:07,  3.06it/s]
 23%|██▎       | 1326/5758 [07:34<23:18,  3.17it/s]
                                                   

 23%|██▎       | 1326/5758 [07:34<23:18,  3.17it/s]
 23%|██▎       | 1327/5758 [07:34<24:45,  2.98it/s]
                                                   

 23%|██▎       | 1327/5758 [07:34<24:45,  2.98it/s]
 23%|██▎       | 1328/5758 [07:35<24:05,  3.06it/s]
                                                   

 23%|██▎       | 1328/5758 [07:35<24:05,  3.06it/s]
 23%|██▎       | 1329/5758 [07:35<25:45,  2.87it/s]
                                                   

 23%|██▎       | 1329/5758 [07:35<25:45,  2.87it/s]
 23%|██▎       | 1330/5758 [07:35<25:25,  2.90it/s]
                                                   

 23%|██▎       | 1330/5758 [07:35<25:25,  2.90it/s]
 23%|██▎       | 1331/5758 [07:36<25:59,  2.84it/s]
                                                   
{'loss': '0.8327', 'grad_norm': '10.44', 'learning_rate': '7.786e-06', 'num_tokens': '8.342e+05', 'mean_token_accuracy': '0.7776', 'epoch': '0.2216'}
+{'loss': '1.725', 'grad_norm': '13.75', 'learning_rate': '7.784e-06', 'num_tokens': '8.348e+05', 'mean_token_accuracy': '0.6341', 'epoch': '0.2218'}
+{'loss': '1.289', 'grad_norm': '11.88', 'learning_rate': '7.782e-06', 'num_tokens': '8.355e+05', 'mean_token_accuracy': '0.6448', 'epoch': '0.222'}
+{'loss': '1.163', 'grad_norm': '12.62', 'learning_rate': '7.78e-06', 'num_tokens': '8.362e+05', 'mean_token_accuracy': '0.6736', 'epoch': '0.2221'}
+{'loss': '2.197', 'grad_norm': '17.62', 'learning_rate': '7.779e-06', 'num_tokens': '8.368e+05', 'mean_token_accuracy': '0.5247', 'epoch': '0.2223'}
+{'loss': '1.133', 'grad_norm': '12.31', 'learning_rate': '7.777e-06', 'num_tokens': '8.375e+05', 'mean_token_accuracy': '0.7083', 'epoch': '0.2225'}
+{'loss': '1.102', 'grad_norm': '17.88', 'learning_rate': '7.775e-06', 'num_tokens': '8.377e+05', 'mean_token_accuracy': '0.7315', 'epoch': '0.2226'}
+{'loss': '1.253', 'grad_norm': '11.44', 'learning_rate': '7.774e-06', 'num_tokens': '8.386e+05', 'mean_token_accuracy': '0.6674', 'epoch': '0.2228'}
+{'loss': '1.55', 'grad_norm': '13.44', 'learning_rate': '7.772e-06', 'num_tokens': '8.392e+05', 'mean_token_accuracy': '0.6629', 'epoch': '0.223'}
+{'loss': '1.617', 'grad_norm': '30.25', 'learning_rate': '7.77e-06', 'num_tokens': '8.393e+05', 'mean_token_accuracy': '0.5758', 'epoch': '0.2232'}
+{'loss': '2.065', 'grad_norm': '12.25', 'learning_rate': '7.768e-06', 'num_tokens': '8.399e+05', 'mean_token_accuracy': '0.5932', 'epoch': '0.2233'}
+{'loss': '1.183', 'grad_norm': '12.81', 'learning_rate': '7.767e-06', 'num_tokens': '8.405e+05', 'mean_token_accuracy': '0.6812', 'epoch': '0.2235'}
+{'loss': '1.305', 'grad_norm': '57.75', 'learning_rate': '7.765e-06', 'num_tokens': '8.406e+05', 'mean_token_accuracy': '0.7015', 'epoch': '0.2237'}
+{'loss': '1.039', 'grad_norm': '25.25', 'learning_rate': '7.763e-06', 'num_tokens': '8.408e+05', 'mean_token_accuracy': '0.7557', 'epoch': '0.2239'}
+{'loss': '1.239', 'grad_norm': '12.31', 'learning_rate': '7.761e-06', 'num_tokens': '8.415e+05', 'mean_token_accuracy': '0.6643', 'epoch': '0.224'}
+{'loss': '1.404', 'grad_norm': '34', 'learning_rate': '7.76e-06', 'num_tokens': '8.416e+05', 'mean_token_accuracy': '0.7097', 'epoch': '0.2242'}
+{'loss': '1.238', 'grad_norm': '11.81', 'learning_rate': '7.758e-06', 'num_tokens': '8.423e+05', 'mean_token_accuracy': '0.6548', 'epoch': '0.2244'}
+{'loss': '0.5999', 'grad_norm': '6.625', 'learning_rate': '7.756e-06', 'num_tokens': '8.442e+05', 'mean_token_accuracy': '0.8181', 'epoch': '0.2246'}
+{'loss': '0.6704', 'grad_norm': '14.25', 'learning_rate': '7.754e-06', 'num_tokens': '8.446e+05', 'mean_token_accuracy': '0.8343', 'epoch': '0.2247'}
+{'loss': '1.042', 'grad_norm': '12.19', 'learning_rate': '7.753e-06', 'num_tokens': '8.453e+05', 'mean_token_accuracy': '0.6979', 'epoch': '0.2249'}
+{'loss': '1.377', 'grad_norm': '30.75', 'learning_rate': '7.751e-06', 'num_tokens': '8.454e+05', 'mean_token_accuracy': '0.7077', 'epoch': '0.2251'}
+{'loss': '0.896', 'grad_norm': '29.62', 'learning_rate': '7.749e-06', 'num_tokens': '8.455e+05', 'mean_token_accuracy': '0.7951', 'epoch': '0.2253'}
+{'loss': '1.124', 'grad_norm': '12.31', 'learning_rate': '7.747e-06', 'num_tokens': '8.462e+05', 'mean_token_accuracy': '0.6848', 'epoch': '0.2254'}
+{'loss': '0.5317', 'grad_norm': '16', 'learning_rate': '7.746e-06', 'num_tokens': '8.464e+05', 'mean_token_accuracy': '0.8633', 'epoch': '0.2256'}
+{'loss': '0.9809', 'grad_norm': '19.12', 'learning_rate': '7.744e-06', 'num_tokens': '8.467e+05', 'mean_token_accuracy': '0.827', 'epoch': '0.2258'}
+{'loss': '0.875', 'grad_norm': '38.75', 'learning_rate': '7.742e-06', 'num_tokens': '8.467e+05', 'mean_token_accuracy': '0.7955', 'epoch': '0.2259'}
+{'loss': '1.929', 'grad_norm': '8.875', 'learning_rate': '7.741e-06', 'num_tokens': '8.484e+05', 'mean_token_accuracy': '0.6091', 'epoch': '0.2261'}
+{'loss': '1.312', 'grad_norm': '9.812', 'learning_rate': '7.739e-06', 'num_tokens': '8.498e+05', 'mean_token_accuracy': '0.6664', 'epoch': '0.2263'}
+{'loss': '1.833', 'grad_norm': '19.88', 'learning_rate': '7.737e-06', 'num_tokens': '8.502e+05', 'mean_token_accuracy': '0.6247', 'epoch': '0.2265'}
+{'loss': '1.192', 'grad_norm': '8.688', 'learning_rate': '7.735e-06', 'num_tokens': '8.514e+05', 'mean_token_accuracy': '0.7347', 'epoch': '0.2266'}
+{'loss': '1.4', 'grad_norm': '12.69', 'learning_rate': '7.734e-06', 'num_tokens': '8.521e+05', 'mean_token_accuracy': '0.6395', 'epoch': '0.2268'}
+{'loss': '1.597', 'grad_norm': '21', 'learning_rate': '7.732e-06', 'num_tokens': '8.523e+05', 'mean_token_accuracy': '0.6624', 'epoch': '0.227'}
+{'loss': '0.9891', 'grad_norm': '11', 'learning_rate': '7.73e-06', 'num_tokens': '8.532e+05', 'mean_token_accuracy': '0.7065', 'epoch': '0.2272'}
+{'loss': '1.29', 'grad_norm': '25.62', 'learning_rate': '7.728e-06', 'num_tokens': '8.534e+05', 'mean_token_accuracy': '0.7452', 'epoch': '0.2273'}
+{'loss': '2.408', 'grad_norm': '8.188', 'learning_rate': '7.727e-06', 'num_tokens': '8.551e+05', 'mean_token_accuracy': '0.5051', 'epoch': '0.2275'}
+{'loss': '1.699', 'grad_norm': '24.62', 'learning_rate': '7.725e-06', 'num_tokens': '8.553e+05', 'mean_token_accuracy': '0.6417', 'epoch': '0.2277'}
+{'loss': '0.5551', 'grad_norm': '6.844', 'learning_rate': '7.723e-06', 'num_tokens': '8.567e+05', 'mean_token_accuracy': '0.8302', 'epoch': '0.2279'}
+{'loss': '1.341', 'grad_norm': '8.5', 'learning_rate': '7.721e-06', 'num_tokens': '8.579e+05', 'mean_token_accuracy': '0.6861', 'epoch': '0.228'}
+{'loss': '1.992', 'grad_norm': '9.812', 'learning_rate': '7.72e-06', 'num_tokens': '8.588e+05', 'mean_token_accuracy': '0.5911', 'epoch': '0.2282'}
+{'loss': '1.59', 'grad_norm': '16.5', 'learning_rate': '7.718e-06', 'num_tokens': '8.593e+05', 'mean_token_accuracy': '0.6277', 'epoch': '0.2284'}
+{'loss': '1.544', 'grad_norm': '12.19', 'learning_rate': '7.716e-06', 'num_tokens': '8.602e+05', 'mean_token_accuracy': '0.6024', 'epoch': '0.2286'}
+{'loss': '1.915', 'grad_norm': '11', 'learning_rate': '7.714e-06', 'num_tokens': '8.608e+05', 'mean_token_accuracy': '0.613', 'epoch': '0.2287'}
+{'loss': '1.096', 'grad_norm': '39', 'learning_rate': '7.713e-06', 'num_tokens': '8.609e+05', 'mean_token_accuracy': '0.7447', 'epoch': '0.2289'}
+{'loss': '1.756', 'grad_norm': '11.56', 'learning_rate': '7.711e-06', 'num_tokens': '8.616e+05', 'mean_token_accuracy': '0.6229', 'epoch': '0.2291'}
+{'loss': '1.451', 'grad_norm': '14.62', 'learning_rate': '7.709e-06', 'num_tokens': '8.621e+05', 'mean_token_accuracy': '0.6313', 'epoch': '0.2292'}
+{'loss': '1.936', 'grad_norm': '17.88', 'learning_rate': '7.708e-06', 'num_tokens': '8.627e+05', 'mean_token_accuracy': '0.5958', 'epoch': '0.2294'}
+{'loss': '1.158', 'grad_norm': '6.812', 'learning_rate': '7.706e-06', 'num_tokens': '8.647e+05', 'mean_token_accuracy': '0.7278', 'epoch': '0.2296'}
+{'loss': '1.403', 'grad_norm': '12.31', 'learning_rate': '7.704e-06', 'num_tokens': '8.654e+05', 'mean_token_accuracy': '0.6175', 'epoch': '0.2298'}
+{'loss': '1.351', 'grad_norm': '11.44', 'learning_rate': '7.702e-06', 'num_tokens': '8.663e+05', 'mean_token_accuracy': '0.6363', 'epoch': '0.2299'}
+{'loss': '1.186', 'grad_norm': '9.312', 'learning_rate': '7.701e-06', 'num_tokens': '8.676e+05', 'mean_token_accuracy': '0.7108', 'epoch': '0.2301'}
+{'loss': '1.699', 'grad_norm': '17.25', 'learning_rate': '7.699e-06', 'num_tokens': '8.679e+05', 'mean_token_accuracy': '0.6321', 'epoch': '0.2303'}
+{'loss': '2.065', 'grad_norm': '19.25', 'learning_rate': '7.697e-06', 'num_tokens': '8.682e+05', 'mean_token_accuracy': '0.5567', 'epoch': '0.2305'}
+{'loss': '1.276', 'grad_norm': '17.12', 'learning_rate': '7.695e-06', 'num_tokens': '8.685e+05', 'mean_token_accuracy': '0.7221', 'epoch': '0.2306'}
+{'loss': '0.8746', 'grad_norm': '25.88', 'learning_rate': '7.694e-06', 'num_tokens': '8.687e+05', 'mean_token_accuracy': '0.8042', 'epoch': '0.2308'}
+{'loss': '1.29', 'grad_norm': '7.781', 'learning_rate': '7.692e-06', 'num_tokens': '8.704e+05', 'mean_token_accuracy': '0.6733', 'epoch': '0.231'}
+
 23%|██▎       | 1331/5758 [07:36<25:59,  2.84it/s]
 23%|██▎       | 1332/5758 [07:36<29:20,  2.51it/s]
                                                   

 23%|██▎       | 1332/5758 [07:36<29:20,  2.51it/s]
 23%|██▎       | 1333/5758 [07:37<28:43,  2.57it/s]
                                                   

 23%|██▎       | 1333/5758 [07:37<28:43,  2.57it/s]
 23%|██▎       | 1334/5758 [07:37<27:07,  2.72it/s]
                                                   

 23%|██▎       | 1334/5758 [07:37<27:07,  2.72it/s]
 23%|██▎       | 1335/5758 [07:37<26:57,  2.73it/s]
                                                   

 23%|██▎       | 1335/5758 [07:37<26:57,  2.73it/s]
 23%|██▎       | 1336/5758 [07:38<27:18,  2.70it/s]
                                                   

 23%|██▎       | 1336/5758 [07:38<27:18,  2.70it/s]
 23%|██▎       | 1337/5758 [07:38<25:44,  2.86it/s]
                                                   

 23%|██▎       | 1337/5758 [07:38<25:44,  2.86it/s]
 23%|██▎       | 1338/5758 [07:38<24:58,  2.95it/s]
                                                   

 23%|██▎       | 1338/5758 [07:38<24:58,  2.95it/s]
 23%|██▎       | 1339/5758 [07:39<23:52,  3.08it/s]
                                                   

 23%|██▎       | 1339/5758 [07:39<23:52,  3.08it/s]
 23%|██▎       | 1340/5758 [07:39<24:36,  2.99it/s]
                                                   

 23%|██▎       | 1340/5758 [07:39<24:36,  2.99it/s]
 23%|██▎       | 1341/5758 [07:39<24:55,  2.95it/s]
                                                   

 23%|██▎       | 1341/5758 [07:39<24:55,  2.95it/s]
 23%|██▎       | 1342/5758 [07:40<25:24,  2.90it/s]
                                                   

 23%|██▎       | 1342/5758 [07:40<25:24,  2.90it/s]
 23%|██▎       | 1343/5758 [07:40<24:45,  2.97it/s]
                                                   

 23%|██▎       | 1343/5758 [07:40<24:45,  2.97it/s]
 23%|██▎       | 1344/5758 [07:40<24:09,  3.05it/s]
                                                   

 23%|██▎       | 1344/5758 [07:40<24:09,  3.05it/s]
 23%|██▎       | 1345/5758 [07:41<23:38,  3.11it/s]
                                                   

 23%|██▎       | 1345/5758 [07:41<23:38,  3.11it/s]
 23%|██▎       | 1346/5758 [07:41<23:19,  3.15it/s]
                                                   

 23%|██▎       | 1346/5758 [07:41<23:19,  3.15it/s]
 23%|██▎       | 1347/5758 [07:41<24:29,  3.00it/s]
                                                   

 23%|██▎       | 1347/5758 [07:41<24:29,  3.00it/s]
 23%|██▎       | 1348/5758 [07:42<23:36,  3.11it/s]
                                                   

 23%|██▎       | 1348/5758 [07:42<23:36,  3.11it/s]
 23%|██▎       | 1349/5758 [07:42<22:56,  3.20it/s]
                                                   

 23%|██▎       | 1349/5758 [07:42<22:56,  3.20it/s]
 23%|██▎       | 1350/5758 [07:42<23:15,  3.16it/s]
                                                   

 23%|██▎       | 1350/5758 [07:42<23:15,  3.16it/s]
 23%|██▎       | 1351/5758 [07:43<24:23,  3.01it/s]
                                                   

 23%|██▎       | 1351/5758 [07:43<24:23,  3.01it/s]
 23%|██▎       | 1352/5758 [07:43<24:49,  2.96it/s]
                                                   

 23%|██▎       | 1352/5758 [07:43<24:49,  2.96it/s]
 23%|██▎       | 1353/5758 [07:43<25:32,  2.87it/s]
                                                   

 23%|██▎       | 1353/5758 [07:43<25:32,  2.87it/s]
 24%|██▎       | 1354/5758 [07:44<24:26,  3.00it/s]
                                                   

 24%|██▎       | 1354/5758 [07:44<24:26,  3.00it/s]
 24%|██▎       | 1355/5758 [07:44<23:48,  3.08it/s]
                                                   

 24%|██▎       | 1355/5758 [07:44<23:48,  3.08it/s]
 24%|██▎       | 1356/5758 [07:44<23:19,  3.14it/s]
                                                   

 24%|██▎       | 1356/5758 [07:44<23:19,  3.14it/s]
 24%|██▎       | 1357/5758 [07:45<25:37,  2.86it/s]
                                                   

 24%|██▎       | 1357/5758 [07:45<25:37,  2.86it/s]
 24%|██▎       | 1358/5758 [07:45<25:18,  2.90it/s]
                                                   

 24%|██▎       | 1358/5758 [07:45<25:18,  2.90it/s]
 24%|██▎       | 1359/5758 [07:45<25:02,  2.93it/s]
                                                   

 24%|██▎       | 1359/5758 [07:45<25:02,  2.93it/s]
 24%|██▎       | 1360/5758 [07:46<24:19,  3.01it/s]
                                                   

 24%|██▎       | 1360/5758 [07:46<24:19,  3.01it/s]
 24%|██▎       | 1361/5758 [07:46<23:16,  3.15it/s]
                                                   

 24%|██▎       | 1361/5758 [07:46<23:16,  3.15it/s]
 24%|██▎       | 1362/5758 [07:46<23:49,  3.08it/s]
                                                   

 24%|██▎       | 1362/5758 [07:46<23:49,  3.08it/s]
 24%|██▎       | 1363/5758 [07:47<23:07,  3.17it/s]
                                                   

 24%|██▎       | 1363/5758 [07:47<23:07,  3.17it/s]
 24%|██▎       | 1364/5758 [07:47<22:55,  3.19it/s]
                                                   

 24%|██▎       | 1364/5758 [07:47<22:55,  3.19it/s]
 24%|██▎       | 1365/5758 [07:47<23:29,  3.12it/s]
                                                   

 24%|██▎       | 1365/5758 [07:47<23:29,  3.12it/s]
 24%|██▎       | 1366/5758 [07:48<23:56,  3.06it/s]
                                                   

 24%|██▎       | 1366/5758 [07:48<23:56,  3.06it/s]
 24%|██▎       | 1367/5758 [07:48<22:59,  3.18it/s]
                                                   

 24%|██▎       | 1367/5758 [07:48<22:59,  3.18it/s]
 24%|██▍       | 1368/5758 [07:48<22:23,  3.27it/s]
                                                   

 24%|██▍       | 1368/5758 [07:48<22:23,  3.27it/s]
 24%|██▍       | 1369/5758 [07:48<24:12,  3.02it/s]
                                                   

 24%|██▍       | 1369/5758 [07:48<24:12,  3.02it/s]
 24%|██▍       | 1370/5758 [07:49<23:14,  3.15it/s]
                                                   

 24%|██▍       | 1370/5758 [07:49<23:14,  3.15it/s]
 24%|██▍       | 1371/5758 [07:49<24:14,  3.02it/s]
                                                   

 24%|██▍       | 1371/5758 [07:49<24:14,  3.02it/s]
 24%|██▍       | 1372/5758 [07:49<23:26,  3.12it/s]
                                                   

 24%|██▍       | 1372/5758 [07:49<23:26,  3.12it/s]
 24%|██▍       | 1373/5758 [07:50<22:41,  3.22it/s]
                                                   

 24%|██▍       | 1373/5758 [07:50<22:41,  3.22it/s]
 24%|██▍       | 1374/5758 [07:50<23:39,  3.09it/s]
                                                   

 24%|██▍       | 1374/5758 [07:50<23:39,  3.09it/s]
 24%|██▍       | 1375/5758 [07:50<23:04,  3.16it/s]
                                                   

 24%|██▍       | 1375/5758 [07:50<23:04,  3.16it/s]
 24%|██▍       | 1376/5758 [07:51<22:56,  3.18it/s]
                                                   

 24%|██▍       | 1376/5758 [07:51<22:56,  3.18it/s]
 24%|██▍       | 1377/5758 [07:51<23:15,  3.14it/s]
                                                   

 24%|██▍       | 1377/5758 [07:51<23:15,  3.14it/s]
 24%|██▍       | 1378/5758 [07:51<22:41,  3.22it/s]
                                                   

 24%|██▍       | 1378/5758 [07:51<22:41,  3.22it/s]
 24%|██▍       | 1379/5758 [07:52<22:57,  3.18it/s]
                                                   

 24%|██▍       | 1379/5758 [07:52<22:57,  3.18it/s]
 24%|██▍       | 1380/5758 [07:52<22:25,  3.25it/s]
                                                   

 24%|██▍       | 1380/5758 [07:52<22:25,  3.25it/s]
 24%|██▍       | 1381/5758 [07:52<23:55,  3.05it/s]
                                                   

 24%|██▍       | 1381/5758 [07:52<23:55,  3.05it/s]
 24%|██▍       | 1382/5758 [07:53<24:57,  2.92it/s]
                                                   

 24%|██▍       | 1382/5758 [07:53<24:57,  2.92it/s]
 24%|██▍       | 1383/5758 [07:53<27:44,  2.63it/s]
                                                   

 24%|██▍       | 1383/5758 [07:53<27:44,  2.63it/s]
 24%|██▍       | 1384/5758 [07:53<26:08,  2.79it/s]
                                                   

 24%|██▍       | 1384/5758 [07:53<26:08,  2.79it/s]
 24%|██▍       | 1385/5758 [07:54<24:52,  2.93it/s]
                                                   

 24%|██▍       | 1385/5758 [07:54<24:52,  2.93it/s]
 24%|██▍       | 1386/5758 [07:54<24:45,  2.94it/s]
                                                   
{'loss': '0.5431', 'grad_norm': '26.12', 'learning_rate': '7.69e-06', 'num_tokens': '8.705e+05', 'mean_token_accuracy': '0.8583', 'epoch': '0.2312'}
+{'loss': '2.315', 'grad_norm': '18.5', 'learning_rate': '7.688e-06', 'num_tokens': '8.71e+05', 'mean_token_accuracy': '0.5522', 'epoch': '0.2313'}
+{'loss': '1.547', 'grad_norm': '16.38', 'learning_rate': '7.687e-06', 'num_tokens': '8.716e+05', 'mean_token_accuracy': '0.6449', 'epoch': '0.2315'}
+{'loss': '2.148', 'grad_norm': '15.69', 'learning_rate': '7.685e-06', 'num_tokens': '8.72e+05', 'mean_token_accuracy': '0.5733', 'epoch': '0.2317'}
+{'loss': '1.592', 'grad_norm': '19.75', 'learning_rate': '7.683e-06', 'num_tokens': '8.725e+05', 'mean_token_accuracy': '0.6905', 'epoch': '0.2319'}
+{'loss': '1.506', 'grad_norm': '20.5', 'learning_rate': '7.681e-06', 'num_tokens': '8.727e+05', 'mean_token_accuracy': '0.6429', 'epoch': '0.232'}
+{'loss': '1.818', 'grad_norm': '16.62', 'learning_rate': '7.68e-06', 'num_tokens': '8.734e+05', 'mean_token_accuracy': '0.625', 'epoch': '0.2322'}
+{'loss': '1.138', 'grad_norm': '11.81', 'learning_rate': '7.678e-06', 'num_tokens': '8.742e+05', 'mean_token_accuracy': '0.683', 'epoch': '0.2324'}
+{'loss': '1.034', 'grad_norm': '21', 'learning_rate': '7.676e-06', 'num_tokens': '8.743e+05', 'mean_token_accuracy': '0.745', 'epoch': '0.2325'}
+{'loss': '0.7831', 'grad_norm': '11.38', 'learning_rate': '7.675e-06', 'num_tokens': '8.753e+05', 'mean_token_accuracy': '0.7662', 'epoch': '0.2327'}
+{'loss': '1.854', 'grad_norm': '17.5', 'learning_rate': '7.673e-06', 'num_tokens': '8.758e+05', 'mean_token_accuracy': '0.6286', 'epoch': '0.2329'}
+{'loss': '1.143', 'grad_norm': '10.12', 'learning_rate': '7.671e-06', 'num_tokens': '8.768e+05', 'mean_token_accuracy': '0.7371', 'epoch': '0.2331'}
+{'loss': '1.15', 'grad_norm': '11.25', 'learning_rate': '7.669e-06', 'num_tokens': '8.776e+05', 'mean_token_accuracy': '0.7037', 'epoch': '0.2332'}
+{'loss': '1.959', 'grad_norm': '51.75', 'learning_rate': '7.668e-06', 'num_tokens': '8.776e+05', 'mean_token_accuracy': '0.5926', 'epoch': '0.2334'}
+{'loss': '1.181', 'grad_norm': '8.938', 'learning_rate': '7.666e-06', 'num_tokens': '8.79e+05', 'mean_token_accuracy': '0.711', 'epoch': '0.2336'}
+{'loss': '0.4566', 'grad_norm': '6.906', 'learning_rate': '7.664e-06', 'num_tokens': '8.804e+05', 'mean_token_accuracy': '0.8668', 'epoch': '0.2338'}
+{'loss': '0.606', 'grad_norm': '9.25', 'learning_rate': '7.662e-06', 'num_tokens': '8.814e+05', 'mean_token_accuracy': '0.8243', 'epoch': '0.2339'}
+{'loss': '1.672', 'grad_norm': '42.75', 'learning_rate': '7.661e-06', 'num_tokens': '8.815e+05', 'mean_token_accuracy': '0.6912', 'epoch': '0.2341'}
+{'loss': '1.168', 'grad_norm': '12.88', 'learning_rate': '7.659e-06', 'num_tokens': '8.821e+05', 'mean_token_accuracy': '0.6924', 'epoch': '0.2343'}
+{'loss': '1.702', 'grad_norm': '13.94', 'learning_rate': '7.657e-06', 'num_tokens': '8.829e+05', 'mean_token_accuracy': '0.6112', 'epoch': '0.2345'}
+{'loss': '1.742', 'grad_norm': '13.25', 'learning_rate': '7.655e-06', 'num_tokens': '8.834e+05', 'mean_token_accuracy': '0.6667', 'epoch': '0.2346'}
+{'loss': '1.249', 'grad_norm': '31.62', 'learning_rate': '7.654e-06', 'num_tokens': '8.836e+05', 'mean_token_accuracy': '0.8333', 'epoch': '0.2348'}
+{'loss': '0.7377', 'grad_norm': '20.5', 'learning_rate': '7.652e-06', 'num_tokens': '8.838e+05', 'mean_token_accuracy': '0.8052', 'epoch': '0.235'}
+{'loss': '1.588', 'grad_norm': '44', 'learning_rate': '7.65e-06', 'num_tokens': '8.839e+05', 'mean_token_accuracy': '0.6667', 'epoch': '0.2352'}
+{'loss': '0.6685', 'grad_norm': '8.938', 'learning_rate': '7.648e-06', 'num_tokens': '8.853e+05', 'mean_token_accuracy': '0.7984', 'epoch': '0.2353'}
+{'loss': '1.256', 'grad_norm': '12.81', 'learning_rate': '7.647e-06', 'num_tokens': '8.859e+05', 'mean_token_accuracy': '0.6758', 'epoch': '0.2355'}
+{'loss': '1.138', 'grad_norm': '12.44', 'learning_rate': '7.645e-06', 'num_tokens': '8.866e+05', 'mean_token_accuracy': '0.7278', 'epoch': '0.2357'}
+{'loss': '1.237', 'grad_norm': '15.81', 'learning_rate': '7.643e-06', 'num_tokens': '8.871e+05', 'mean_token_accuracy': '0.6562', 'epoch': '0.2358'}
+{'loss': '1.555', 'grad_norm': '12.31', 'learning_rate': '7.642e-06', 'num_tokens': '8.88e+05', 'mean_token_accuracy': '0.5718', 'epoch': '0.236'}
+{'loss': '1.766', 'grad_norm': '12.69', 'learning_rate': '7.64e-06', 'num_tokens': '8.886e+05', 'mean_token_accuracy': '0.6285', 'epoch': '0.2362'}
+{'loss': '0.8045', 'grad_norm': '40.5', 'learning_rate': '7.638e-06', 'num_tokens': '8.887e+05', 'mean_token_accuracy': '0.8358', 'epoch': '0.2364'}
+{'loss': '1.73', 'grad_norm': '21.12', 'learning_rate': '7.636e-06', 'num_tokens': '8.892e+05', 'mean_token_accuracy': '0.6194', 'epoch': '0.2365'}
+{'loss': '1.124', 'grad_norm': '12.12', 'learning_rate': '7.635e-06', 'num_tokens': '8.898e+05', 'mean_token_accuracy': '0.6832', 'epoch': '0.2367'}
+{'loss': '0.9948', 'grad_norm': '40', 'learning_rate': '7.633e-06', 'num_tokens': '8.899e+05', 'mean_token_accuracy': '0.7952', 'epoch': '0.2369'}
+{'loss': '1.968', 'grad_norm': '9.75', 'learning_rate': '7.631e-06', 'num_tokens': '8.908e+05', 'mean_token_accuracy': '0.5611', 'epoch': '0.2371'}
+{'loss': '0.9948', 'grad_norm': '32.5', 'learning_rate': '7.629e-06', 'num_tokens': '8.909e+05', 'mean_token_accuracy': '0.75', 'epoch': '0.2372'}
+{'loss': '0.8987', 'grad_norm': '13', 'learning_rate': '7.628e-06', 'num_tokens': '8.916e+05', 'mean_token_accuracy': '0.7467', 'epoch': '0.2374'}
+{'loss': '1.806', 'grad_norm': '18.25', 'learning_rate': '7.626e-06', 'num_tokens': '8.922e+05', 'mean_token_accuracy': '0.6145', 'epoch': '0.2376'}
+{'loss': '0.7136', 'grad_norm': '20.5', 'learning_rate': '7.624e-06', 'num_tokens': '8.923e+05', 'mean_token_accuracy': '0.8321', 'epoch': '0.2378'}
+{'loss': '1.083', 'grad_norm': '7.594', 'learning_rate': '7.622e-06', 'num_tokens': '8.936e+05', 'mean_token_accuracy': '0.7422', 'epoch': '0.2379'}
+{'loss': '2.05', 'grad_norm': '20', 'learning_rate': '7.621e-06', 'num_tokens': '8.941e+05', 'mean_token_accuracy': '0.6041', 'epoch': '0.2381'}
+{'loss': '1.07', 'grad_norm': '12.19', 'learning_rate': '7.619e-06', 'num_tokens': '8.947e+05', 'mean_token_accuracy': '0.6978', 'epoch': '0.2383'}
+{'loss': '1.496', 'grad_norm': '33.25', 'learning_rate': '7.617e-06', 'num_tokens': '8.948e+05', 'mean_token_accuracy': '0.7183', 'epoch': '0.2385'}
+{'loss': '1.972', 'grad_norm': '35.25', 'learning_rate': '7.615e-06', 'num_tokens': '8.949e+05', 'mean_token_accuracy': '0.5478', 'epoch': '0.2386'}
+{'loss': '2.023', 'grad_norm': '15.38', 'learning_rate': '7.614e-06', 'num_tokens': '8.953e+05', 'mean_token_accuracy': '0.5988', 'epoch': '0.2388'}
+{'loss': '1.961', 'grad_norm': '18.25', 'learning_rate': '7.612e-06', 'num_tokens': '8.956e+05', 'mean_token_accuracy': '0.6137', 'epoch': '0.239'}
+{'loss': '1.212', 'grad_norm': '11.69', 'learning_rate': '7.61e-06', 'num_tokens': '8.965e+05', 'mean_token_accuracy': '0.661', 'epoch': '0.2391'}
+{'loss': '2.094', 'grad_norm': '17.88', 'learning_rate': '7.609e-06', 'num_tokens': '8.971e+05', 'mean_token_accuracy': '0.5854', 'epoch': '0.2393'}
+{'loss': '1.157', 'grad_norm': '15.06', 'learning_rate': '7.607e-06', 'num_tokens': '8.975e+05', 'mean_token_accuracy': '0.7614', 'epoch': '0.2395'}
+{'loss': '1.149', 'grad_norm': '13.25', 'learning_rate': '7.605e-06', 'num_tokens': '8.981e+05', 'mean_token_accuracy': '0.6753', 'epoch': '0.2397'}
+{'loss': '1.436', 'grad_norm': '14.56', 'learning_rate': '7.603e-06', 'num_tokens': '8.987e+05', 'mean_token_accuracy': '0.6384', 'epoch': '0.2398'}
+{'loss': '0.9541', 'grad_norm': '19', 'learning_rate': '7.602e-06', 'num_tokens': '8.99e+05', 'mean_token_accuracy': '0.7778', 'epoch': '0.24'}
+{'loss': '1.049', 'grad_norm': '28.5', 'learning_rate': '7.6e-06', 'num_tokens': '8.991e+05', 'mean_token_accuracy': '0.7818', 'epoch': '0.2402'}
+{'loss': '1.037', 'grad_norm': '12.31', 'learning_rate': '7.598e-06', 'num_tokens': '8.998e+05', 'mean_token_accuracy': '0.6919', 'epoch': '0.2404'}
+{'loss': '0.9046', 'grad_norm': '23.12', 'learning_rate': '7.596e-06', 'num_tokens': '9e+05', 'mean_token_accuracy': '0.7633', 'epoch': '0.2405'}
+
 24%|██▍       | 1386/5758 [07:54<24:45,  2.94it/s]
 24%|██▍       | 1387/5758 [07:54<24:09,  3.01it/s]
                                                   

 24%|██▍       | 1387/5758 [07:54<24:09,  3.01it/s]
 24%|██▍       | 1388/5758 [07:55<23:16,  3.13it/s]
                                                   

 24%|██▍       | 1388/5758 [07:55<23:16,  3.13it/s]
 24%|██▍       | 1389/5758 [07:55<24:18,  3.00it/s]
                                                   

 24%|██▍       | 1389/5758 [07:55<24:18,  3.00it/s]
 24%|██▍       | 1390/5758 [07:55<23:18,  3.12it/s]
                                                   

 24%|██▍       | 1390/5758 [07:55<23:18,  3.12it/s]
 24%|██▍       | 1391/5758 [07:56<23:40,  3.07it/s]
                                                   

 24%|██▍       | 1391/5758 [07:56<23:40,  3.07it/s]
 24%|██▍       | 1392/5758 [07:56<22:36,  3.22it/s]
                                                   

 24%|██▍       | 1392/5758 [07:56<22:36,  3.22it/s]
 24%|██▍       | 1393/5758 [07:56<25:12,  2.89it/s]
                                                   

 24%|██▍       | 1393/5758 [07:56<25:12,  2.89it/s]
 24%|██▍       | 1394/5758 [07:57<25:28,  2.85it/s]
                                                   

 24%|██▍       | 1394/5758 [07:57<25:28,  2.85it/s]
 24%|██▍       | 1395/5758 [07:57<26:17,  2.77it/s]
                                                   

 24%|██▍       | 1395/5758 [07:57<26:17,  2.77it/s]
 24%|██▍       | 1396/5758 [07:57<25:14,  2.88it/s]
                                                   

 24%|██▍       | 1396/5758 [07:57<25:14,  2.88it/s]
 24%|██▍       | 1397/5758 [07:58<25:45,  2.82it/s]
                                                   

 24%|██▍       | 1397/5758 [07:58<25:45,  2.82it/s]
 24%|██▍       | 1398/5758 [07:58<24:27,  2.97it/s]
                                                   

 24%|██▍       | 1398/5758 [07:58<24:27,  2.97it/s]
 24%|██▍       | 1399/5758 [07:58<24:32,  2.96it/s]
                                                   

 24%|██▍       | 1399/5758 [07:58<24:32,  2.96it/s]
 24%|██▍       | 1400/5758 [07:59<24:08,  3.01it/s]
                                                   

 24%|██▍       | 1400/5758 [07:59<24:08,  3.01it/s]
 24%|██▍       | 1401/5758 [07:59<23:12,  3.13it/s]
                                                   

 24%|██▍       | 1401/5758 [07:59<23:12,  3.13it/s]
 24%|██▍       | 1402/5758 [07:59<22:33,  3.22it/s]
                                                   

 24%|██▍       | 1402/5758 [07:59<22:33,  3.22it/s]
 24%|██▍       | 1403/5758 [08:00<24:17,  2.99it/s]
                                                   

 24%|██▍       | 1403/5758 [08:00<24:17,  2.99it/s]
 24%|██▍       | 1404/5758 [08:00<24:12,  3.00it/s]
                                                   

 24%|██▍       | 1404/5758 [08:00<24:12,  3.00it/s]
 24%|██▍       | 1405/5758 [08:00<25:22,  2.86it/s]
                                                   

 24%|██▍       | 1405/5758 [08:00<25:22,  2.86it/s]
 24%|██▍       | 1406/5758 [08:01<25:26,  2.85it/s]
                                                   

 24%|██▍       | 1406/5758 [08:01<25:26,  2.85it/s]
 24%|██▍       | 1407/5758 [08:01<24:29,  2.96it/s]
                                                   

 24%|██▍       | 1407/5758 [08:01<24:29,  2.96it/s]
 24%|██▍       | 1408/5758 [08:02<25:29,  2.84it/s]
                                                   

 24%|██▍       | 1408/5758 [08:02<25:29,  2.84it/s]
 24%|██▍       | 1409/5758 [08:02<27:59,  2.59it/s]
                                                   

 24%|██▍       | 1409/5758 [08:02<27:59,  2.59it/s]
 24%|██▍       | 1410/5758 [08:02<26:15,  2.76it/s]
                                                   

 24%|██▍       | 1410/5758 [08:02<26:15,  2.76it/s]
 25%|██▍       | 1411/5758 [08:03<26:42,  2.71it/s]
                                                   

 25%|██▍       | 1411/5758 [08:03<26:42,  2.71it/s]
 25%|██▍       | 1412/5758 [08:03<25:16,  2.87it/s]
                                                   

 25%|██▍       | 1412/5758 [08:03<25:16,  2.87it/s]
 25%|██▍       | 1413/5758 [08:03<24:05,  3.01it/s]
                                                   

 25%|██▍       | 1413/5758 [08:03<24:05,  3.01it/s]
 25%|██▍       | 1414/5758 [08:04<23:25,  3.09it/s]
                                                   

 25%|██▍       | 1414/5758 [08:04<23:25,  3.09it/s]
 25%|██▍       | 1415/5758 [08:04<23:10,  3.12it/s]
                                                   

 25%|██▍       | 1415/5758 [08:04<23:10,  3.12it/s]
 25%|██▍       | 1416/5758 [08:04<24:20,  2.97it/s]
                                                   

 25%|██▍       | 1416/5758 [08:04<24:20,  2.97it/s]
 25%|██▍       | 1417/5758 [08:05<23:31,  3.08it/s]
                                                   

 25%|██▍       | 1417/5758 [08:05<23:31,  3.08it/s]
 25%|██▍       | 1418/5758 [08:05<23:25,  3.09it/s]
                                                   

 25%|██▍       | 1418/5758 [08:05<23:25,  3.09it/s]
 25%|██▍       | 1419/5758 [08:05<23:38,  3.06it/s]
                                                   

 25%|██▍       | 1419/5758 [08:05<23:38,  3.06it/s]
 25%|██▍       | 1420/5758 [08:06<23:51,  3.03it/s]
                                                   

 25%|██▍       | 1420/5758 [08:06<23:51,  3.03it/s]
 25%|██▍       | 1421/5758 [08:06<24:24,  2.96it/s]
                                                   

 25%|██▍       | 1421/5758 [08:06<24:24,  2.96it/s]
 25%|██▍       | 1422/5758 [08:06<23:30,  3.08it/s]
                                                   

 25%|██▍       | 1422/5758 [08:06<23:30,  3.08it/s]
 25%|██▍       | 1423/5758 [08:07<23:37,  3.06it/s]
                                                   

 25%|██▍       | 1423/5758 [08:07<23:37,  3.06it/s]
 25%|██▍       | 1424/5758 [08:07<23:17,  3.10it/s]
                                                   

 25%|██▍       | 1424/5758 [08:07<23:17,  3.10it/s]
 25%|██▍       | 1425/5758 [08:07<23:56,  3.02it/s]
                                                   

 25%|██▍       | 1425/5758 [08:07<23:56,  3.02it/s]
 25%|██▍       | 1426/5758 [08:08<26:39,  2.71it/s]
                                                   

 25%|██▍       | 1426/5758 [08:08<26:39,  2.71it/s]
 25%|██▍       | 1427/5758 [08:08<25:44,  2.81it/s]
                                                   

 25%|██▍       | 1427/5758 [08:08<25:44,  2.81it/s]
 25%|██▍       | 1428/5758 [08:08<24:13,  2.98it/s]
                                                   

 25%|██▍       | 1428/5758 [08:08<24:13,  2.98it/s]
 25%|██▍       | 1429/5758 [08:09<23:35,  3.06it/s]
                                                   

 25%|██▍       | 1429/5758 [08:09<23:35,  3.06it/s]
 25%|██▍       | 1430/5758 [08:09<24:05,  2.99it/s]
                                                   

 25%|██▍       | 1430/5758 [08:09<24:05,  2.99it/s]
 25%|██▍       | 1431/5758 [08:09<24:15,  2.97it/s]
                                                   

 25%|██▍       | 1431/5758 [08:09<24:15,  2.97it/s]
 25%|██▍       | 1432/5758 [08:10<24:06,  2.99it/s]
                                                   

 25%|██▍       | 1432/5758 [08:10<24:06,  2.99it/s]
 25%|██▍       | 1433/5758 [08:10<23:05,  3.12it/s]
                                                   

 25%|██▍       | 1433/5758 [08:10<23:05,  3.12it/s]
 25%|██▍       | 1434/5758 [08:10<22:53,  3.15it/s]
                                                   

 25%|██▍       | 1434/5758 [08:10<22:53,  3.15it/s]
 25%|██▍       | 1435/5758 [08:11<22:49,  3.16it/s]
                                                   

 25%|██▍       | 1435/5758 [08:11<22:49,  3.16it/s]
 25%|██▍       | 1436/5758 [08:11<23:48,  3.03it/s]
                                                   

 25%|██▍       | 1436/5758 [08:11<23:48,  3.03it/s]
 25%|██▍       | 1437/5758 [08:11<25:55,  2.78it/s]
                                                   

 25%|██▍       | 1437/5758 [08:11<25:55,  2.78it/s]
 25%|██▍       | 1438/5758 [08:12<25:11,  2.86it/s]
                                                   

 25%|██▍       | 1438/5758 [08:12<25:11,  2.86it/s]
 25%|██▍       | 1439/5758 [08:12<24:02,  2.99it/s]
                                                   

 25%|██▍       | 1439/5758 [08:12<24:02,  2.99it/s]
 25%|██▌       | 1440/5758 [08:12<24:14,  2.97it/s]
                                                   

 25%|██▌       | 1440/5758 [08:12<24:14,  2.97it/s]
 25%|██▌       | 1441/5758 [08:13<23:47,  3.02it/s]
                                                   
{'loss': '1.226', 'grad_norm': '12.06', 'learning_rate': '7.595e-06', 'num_tokens': '9.009e+05', 'mean_token_accuracy': '0.6727', 'epoch': '0.2407'}
+{'loss': '2.347', 'grad_norm': '13', 'learning_rate': '7.593e-06', 'num_tokens': '9.016e+05', 'mean_token_accuracy': '0.5103', 'epoch': '0.2409'}
+{'loss': '1.055', 'grad_norm': '13.31', 'learning_rate': '7.591e-06', 'num_tokens': '9.022e+05', 'mean_token_accuracy': '0.708', 'epoch': '0.2411'}
+{'loss': '1.458', 'grad_norm': '15.44', 'learning_rate': '7.589e-06', 'num_tokens': '9.027e+05', 'mean_token_accuracy': '0.6315', 'epoch': '0.2412'}
+{'loss': '0.836', 'grad_norm': '17.62', 'learning_rate': '7.588e-06', 'num_tokens': '9.03e+05', 'mean_token_accuracy': '0.8333', 'epoch': '0.2414'}
+{'loss': '1.155', 'grad_norm': '15.81', 'learning_rate': '7.586e-06', 'num_tokens': '9.035e+05', 'mean_token_accuracy': '0.7111', 'epoch': '0.2416'}
+{'loss': '0.5513', 'grad_norm': '29.88', 'learning_rate': '7.584e-06', 'num_tokens': '9.035e+05', 'mean_token_accuracy': '0.85', 'epoch': '0.2418'}
+{'loss': '1.575', 'grad_norm': '6.062', 'learning_rate': '7.582e-06', 'num_tokens': '9.073e+05', 'mean_token_accuracy': '0.6437', 'epoch': '0.2419'}
+{'loss': '1.831', 'grad_norm': '11', 'learning_rate': '7.581e-06', 'num_tokens': '9.083e+05', 'mean_token_accuracy': '0.6197', 'epoch': '0.2421'}
+{'loss': '1.16', 'grad_norm': '13.31', 'learning_rate': '7.579e-06', 'num_tokens': '9.089e+05', 'mean_token_accuracy': '0.7065', 'epoch': '0.2423'}
+{'loss': '1.018', 'grad_norm': '11.94', 'learning_rate': '7.577e-06', 'num_tokens': '9.097e+05', 'mean_token_accuracy': '0.6954', 'epoch': '0.2424'}
+{'loss': '2.329', 'grad_norm': '15.38', 'learning_rate': '7.576e-06', 'num_tokens': '9.102e+05', 'mean_token_accuracy': '0.5653', 'epoch': '0.2426'}
+{'loss': '1.934', 'grad_norm': '14.31', 'learning_rate': '7.574e-06', 'num_tokens': '9.108e+05', 'mean_token_accuracy': '0.5764', 'epoch': '0.2428'}
+{'loss': '1.52', 'grad_norm': '7.062', 'learning_rate': '7.572e-06', 'num_tokens': '9.126e+05', 'mean_token_accuracy': '0.6605', 'epoch': '0.243'}
+{'loss': '1.003', 'grad_norm': '25', 'learning_rate': '7.57e-06', 'num_tokens': '9.128e+05', 'mean_token_accuracy': '0.8054', 'epoch': '0.2431'}
+{'loss': '1.967', 'grad_norm': '12.75', 'learning_rate': '7.569e-06', 'num_tokens': '9.134e+05', 'mean_token_accuracy': '0.6104', 'epoch': '0.2433'}
+{'loss': '1.297', 'grad_norm': '11', 'learning_rate': '7.567e-06', 'num_tokens': '9.14e+05', 'mean_token_accuracy': '0.7022', 'epoch': '0.2435'}
+{'loss': '1.148', 'grad_norm': '13.62', 'learning_rate': '7.565e-06', 'num_tokens': '9.146e+05', 'mean_token_accuracy': '0.6803', 'epoch': '0.2437'}
+{'loss': '1.182', 'grad_norm': '10.88', 'learning_rate': '7.563e-06', 'num_tokens': '9.154e+05', 'mean_token_accuracy': '0.6817', 'epoch': '0.2438'}
+{'loss': '0.8188', 'grad_norm': '9.625', 'learning_rate': '7.562e-06', 'num_tokens': '9.166e+05', 'mean_token_accuracy': '0.7464', 'epoch': '0.244'}
+{'loss': '1.522', 'grad_norm': '30', 'learning_rate': '7.56e-06', 'num_tokens': '9.167e+05', 'mean_token_accuracy': '0.7615', 'epoch': '0.2442'}
+{'loss': '1.16', 'grad_norm': '11.5', 'learning_rate': '7.558e-06', 'num_tokens': '9.174e+05', 'mean_token_accuracy': '0.7024', 'epoch': '0.2444'}
+{'loss': '0.6788', 'grad_norm': '26.25', 'learning_rate': '7.556e-06', 'num_tokens': '9.175e+05', 'mean_token_accuracy': '0.8605', 'epoch': '0.2445'}
+{'loss': '1.464', 'grad_norm': '13.81', 'learning_rate': '7.555e-06', 'num_tokens': '9.18e+05', 'mean_token_accuracy': '0.7188', 'epoch': '0.2447'}
+{'loss': '0.9644', 'grad_norm': '19.62', 'learning_rate': '7.553e-06', 'num_tokens': '9.182e+05', 'mean_token_accuracy': '0.7821', 'epoch': '0.2449'}
+{'loss': '2.549', 'grad_norm': '9.812', 'learning_rate': '7.551e-06', 'num_tokens': '9.194e+05', 'mean_token_accuracy': '0.5064', 'epoch': '0.2451'}
+{'loss': '1.996', 'grad_norm': '9.188', 'learning_rate': '7.549e-06', 'num_tokens': '9.207e+05', 'mean_token_accuracy': '0.5907', 'epoch': '0.2452'}
+{'loss': '0.9291', 'grad_norm': '46.75', 'learning_rate': '7.548e-06', 'num_tokens': '9.208e+05', 'mean_token_accuracy': '0.7838', 'epoch': '0.2454'}
+{'loss': '1.168', 'grad_norm': '12.12', 'learning_rate': '7.546e-06', 'num_tokens': '9.215e+05', 'mean_token_accuracy': '0.6696', 'epoch': '0.2456'}
+{'loss': '1.111', 'grad_norm': '7.281', 'learning_rate': '7.544e-06', 'num_tokens': '9.23e+05', 'mean_token_accuracy': '0.7359', 'epoch': '0.2457'}
+{'loss': '1.275', 'grad_norm': '15', 'learning_rate': '7.543e-06', 'num_tokens': '9.235e+05', 'mean_token_accuracy': '0.6526', 'epoch': '0.2459'}
+{'loss': '1.392', 'grad_norm': '13.94', 'learning_rate': '7.541e-06', 'num_tokens': '9.242e+05', 'mean_token_accuracy': '0.6375', 'epoch': '0.2461'}
+{'loss': '2.074', 'grad_norm': '18.25', 'learning_rate': '7.539e-06', 'num_tokens': '9.246e+05', 'mean_token_accuracy': '0.6015', 'epoch': '0.2463'}
+{'loss': '1.663', 'grad_norm': '42.25', 'learning_rate': '7.537e-06', 'num_tokens': '9.247e+05', 'mean_token_accuracy': '0.61', 'epoch': '0.2464'}
+{'loss': '1.528', 'grad_norm': '9.625', 'learning_rate': '7.536e-06', 'num_tokens': '9.256e+05', 'mean_token_accuracy': '0.6565', 'epoch': '0.2466'}
+{'loss': '0.7858', 'grad_norm': '10.38', 'learning_rate': '7.534e-06', 'num_tokens': '9.266e+05', 'mean_token_accuracy': '0.7657', 'epoch': '0.2468'}
+{'loss': '1.261', 'grad_norm': '25.12', 'learning_rate': '7.532e-06', 'num_tokens': '9.268e+05', 'mean_token_accuracy': '0.7317', 'epoch': '0.247'}
+{'loss': '0.8328', 'grad_norm': '15.31', 'learning_rate': '7.53e-06', 'num_tokens': '9.272e+05', 'mean_token_accuracy': '0.7982', 'epoch': '0.2471'}
+{'loss': '1.502', 'grad_norm': '14.06', 'learning_rate': '7.529e-06', 'num_tokens': '9.279e+05', 'mean_token_accuracy': '0.6008', 'epoch': '0.2473'}
+{'loss': '1.367', 'grad_norm': '18', 'learning_rate': '7.527e-06', 'num_tokens': '9.282e+05', 'mean_token_accuracy': '0.704', 'epoch': '0.2475'}
+{'loss': '2.019', 'grad_norm': '6.688', 'learning_rate': '7.525e-06', 'num_tokens': '9.304e+05', 'mean_token_accuracy': '0.5776', 'epoch': '0.2477'}
+{'loss': '1.864', 'grad_norm': '15.25', 'learning_rate': '7.523e-06', 'num_tokens': '9.308e+05', 'mean_token_accuracy': '0.6418', 'epoch': '0.2478'}
+{'loss': '0.7846', 'grad_norm': '24.5', 'learning_rate': '7.522e-06', 'num_tokens': '9.31e+05', 'mean_token_accuracy': '0.8212', 'epoch': '0.248'}
+{'loss': '1.261', 'grad_norm': '16.25', 'learning_rate': '7.52e-06', 'num_tokens': '9.313e+05', 'mean_token_accuracy': '0.7479', 'epoch': '0.2482'}
+{'loss': '1.206', 'grad_norm': '21.88', 'learning_rate': '7.518e-06', 'num_tokens': '9.316e+05', 'mean_token_accuracy': '0.7449', 'epoch': '0.2484'}
+{'loss': '1.886', 'grad_norm': '10.81', 'learning_rate': '7.516e-06', 'num_tokens': '9.325e+05', 'mean_token_accuracy': '0.6057', 'epoch': '0.2485'}
+{'loss': '0.8838', 'grad_norm': '13.25', 'learning_rate': '7.515e-06', 'num_tokens': '9.33e+05', 'mean_token_accuracy': '0.7977', 'epoch': '0.2487'}
+{'loss': '0.6493', 'grad_norm': '10', 'learning_rate': '7.513e-06', 'num_tokens': '9.342e+05', 'mean_token_accuracy': '0.8076', 'epoch': '0.2489'}
+{'loss': '1.225', 'grad_norm': '12.31', 'learning_rate': '7.511e-06', 'num_tokens': '9.349e+05', 'mean_token_accuracy': '0.6793', 'epoch': '0.249'}
+{'loss': '0.8784', 'grad_norm': '22.5', 'learning_rate': '7.51e-06', 'num_tokens': '9.351e+05', 'mean_token_accuracy': '0.8063', 'epoch': '0.2492'}
+{'loss': '1.654', 'grad_norm': '12.88', 'learning_rate': '7.508e-06', 'num_tokens': '9.357e+05', 'mean_token_accuracy': '0.6698', 'epoch': '0.2494'}
+{'loss': '1.399', 'grad_norm': '14', 'learning_rate': '7.506e-06', 'num_tokens': '9.364e+05', 'mean_token_accuracy': '0.6577', 'epoch': '0.2496'}
+{'loss': '0.5909', 'grad_norm': '6.781', 'learning_rate': '7.504e-06', 'num_tokens': '9.38e+05', 'mean_token_accuracy': '0.8233', 'epoch': '0.2497'}
+{'loss': '1.138', 'grad_norm': '12.75', 'learning_rate': '7.503e-06', 'num_tokens': '9.387e+05', 'mean_token_accuracy': '0.6771', 'epoch': '0.2499'}
+{'loss': '2.097', 'grad_norm': '22.25', 'learning_rate': '7.501e-06', 'num_tokens': '9.391e+05', 'mean_token_accuracy': '0.5913', 'epoch': '0.2501'}
+
 25%|██▌       | 1441/5758 [08:13<23:47,  3.02it/s]
 25%|██▌       | 1442/5758 [08:13<23:01,  3.12it/s]
                                                   

 25%|██▌       | 1442/5758 [08:13<23:01,  3.12it/s]
 25%|██▌       | 1443/5758 [08:13<23:56,  3.00it/s]
                                                   

 25%|██▌       | 1443/5758 [08:13<23:56,  3.00it/s]
 25%|██▌       | 1444/5758 [08:14<24:10,  2.97it/s]
                                                   

 25%|██▌       | 1444/5758 [08:14<24:10,  2.97it/s]
 25%|██▌       | 1445/5758 [08:14<23:35,  3.05it/s]
                                                   

 25%|██▌       | 1445/5758 [08:14<23:35,  3.05it/s]
 25%|██▌       | 1446/5758 [08:14<22:40,  3.17it/s]
                                                   

 25%|██▌       | 1446/5758 [08:14<22:40,  3.17it/s]
 25%|██▌       | 1447/5758 [08:14<22:39,  3.17it/s]
                                                   

 25%|██▌       | 1447/5758 [08:15<22:39,  3.17it/s]
 25%|██▌       | 1448/5758 [08:15<22:11,  3.24it/s]
                                                   

 25%|██▌       | 1448/5758 [08:15<22:11,  3.24it/s]
 25%|██▌       | 1449/5758 [08:15<23:25,  3.07it/s]
                                                   

 25%|██▌       | 1449/5758 [08:15<23:25,  3.07it/s]
 25%|██▌       | 1450/5758 [08:15<22:53,  3.14it/s]
                                                   

 25%|██▌       | 1450/5758 [08:15<22:53,  3.14it/s]
 25%|██▌       | 1451/5758 [08:16<23:05,  3.11it/s]
                                                   

 25%|██▌       | 1451/5758 [08:16<23:05,  3.11it/s]
 25%|██▌       | 1452/5758 [08:16<21:56,  3.27it/s]
                                                   

 25%|██▌       | 1452/5758 [08:16<21:56,  3.27it/s]
 25%|██▌       | 1453/5758 [08:16<22:12,  3.23it/s]
                                                   

 25%|██▌       | 1453/5758 [08:16<22:12,  3.23it/s]
 25%|██▌       | 1454/5758 [08:17<23:03,  3.11it/s]
                                                   

 25%|██▌       | 1454/5758 [08:17<23:03,  3.11it/s]
 25%|██▌       | 1455/5758 [08:17<22:44,  3.15it/s]
                                                   

 25%|██▌       | 1455/5758 [08:17<22:44,  3.15it/s]
 25%|██▌       | 1456/5758 [08:17<23:16,  3.08it/s]
                                                   

 25%|██▌       | 1456/5758 [08:17<23:16,  3.08it/s]
 25%|██▌       | 1457/5758 [08:18<22:32,  3.18it/s]
                                                   

 25%|██▌       | 1457/5758 [08:18<22:32,  3.18it/s]
 25%|██▌       | 1458/5758 [08:18<23:04,  3.10it/s]
                                                   

 25%|██▌       | 1458/5758 [08:18<23:04,  3.10it/s]
 25%|██▌       | 1459/5758 [08:18<22:48,  3.14it/s]
                                                   

 25%|██▌       | 1459/5758 [08:18<22:48,  3.14it/s]
 25%|██▌       | 1460/5758 [08:19<23:23,  3.06it/s]
                                                   

 25%|██▌       | 1460/5758 [08:19<23:23,  3.06it/s]
 25%|██▌       | 1461/5758 [08:19<23:39,  3.03it/s]
                                                   

 25%|██▌       | 1461/5758 [08:19<23:39,  3.03it/s]
 25%|██▌       | 1462/5758 [08:19<23:47,  3.01it/s]
                                                   

 25%|██▌       | 1462/5758 [08:19<23:47,  3.01it/s]
 25%|██▌       | 1463/5758 [08:20<27:33,  2.60it/s]
                                                   

 25%|██▌       | 1463/5758 [08:20<27:33,  2.60it/s]
 25%|██▌       | 1464/5758 [08:20<25:37,  2.79it/s]
                                                   

 25%|██▌       | 1464/5758 [08:20<25:37,  2.79it/s]
 25%|██▌       | 1465/5758 [08:20<24:56,  2.87it/s]
                                                   

 25%|██▌       | 1465/5758 [08:20<24:56,  2.87it/s]
 25%|██▌       | 1466/5758 [08:21<23:37,  3.03it/s]
                                                   

 25%|██▌       | 1466/5758 [08:21<23:37,  3.03it/s]
 25%|██▌       | 1467/5758 [08:21<23:51,  3.00it/s]
                                                   

 25%|██▌       | 1467/5758 [08:21<23:51,  3.00it/s]
 25%|██▌       | 1468/5758 [08:21<25:00,  2.86it/s]
                                                   

 25%|██▌       | 1468/5758 [08:21<25:00,  2.86it/s]
 26%|██▌       | 1469/5758 [08:22<24:11,  2.96it/s]
                                                   

 26%|██▌       | 1469/5758 [08:22<24:11,  2.96it/s]
 26%|██▌       | 1470/5758 [08:22<23:22,  3.06it/s]
                                                   

 26%|██▌       | 1470/5758 [08:22<23:22,  3.06it/s]
 26%|██▌       | 1471/5758 [08:22<23:46,  3.00it/s]
                                                   

 26%|██▌       | 1471/5758 [08:22<23:46,  3.00it/s]
 26%|██▌       | 1472/5758 [08:23<23:00,  3.10it/s]
                                                   

 26%|██▌       | 1472/5758 [08:23<23:00,  3.10it/s]
 26%|██▌       | 1473/5758 [08:23<23:21,  3.06it/s]
                                                   

 26%|██▌       | 1473/5758 [08:23<23:21,  3.06it/s]
 26%|██▌       | 1474/5758 [08:23<23:42,  3.01it/s]
                                                   

 26%|██▌       | 1474/5758 [08:23<23:42,  3.01it/s]
 26%|██▌       | 1475/5758 [08:24<22:57,  3.11it/s]
                                                   

 26%|██▌       | 1475/5758 [08:24<22:57,  3.11it/s]
 26%|██▌       | 1476/5758 [08:24<23:25,  3.05it/s]
                                                   

 26%|██▌       | 1476/5758 [08:24<23:25,  3.05it/s]
 26%|██▌       | 1477/5758 [08:24<23:44,  3.00it/s]
                                                   

 26%|██▌       | 1477/5758 [08:24<23:44,  3.00it/s]
 26%|██▌       | 1478/5758 [08:25<24:13,  2.94it/s]
                                                   

 26%|██▌       | 1478/5758 [08:25<24:13,  2.94it/s]
 26%|██▌       | 1479/5758 [08:25<23:20,  3.06it/s]
                                                   

 26%|██▌       | 1479/5758 [08:25<23:20,  3.06it/s]
 26%|██▌       | 1480/5758 [08:25<24:16,  2.94it/s]
                                                   

 26%|██▌       | 1480/5758 [08:25<24:16,  2.94it/s]
 26%|██▌       | 1481/5758 [08:26<23:39,  3.01it/s]
                                                   

 26%|██▌       | 1481/5758 [08:26<23:39,  3.01it/s]
 26%|██▌       | 1482/5758 [08:26<24:22,  2.92it/s]
                                                   

 26%|██▌       | 1482/5758 [08:26<24:22,  2.92it/s]
 26%|██▌       | 1483/5758 [08:26<23:04,  3.09it/s]
                                                   

 26%|██▌       | 1483/5758 [08:26<23:04,  3.09it/s]
 26%|██▌       | 1484/5758 [08:27<24:54,  2.86it/s]
                                                   

 26%|██▌       | 1484/5758 [08:27<24:54,  2.86it/s]
 26%|██▌       | 1485/5758 [08:27<24:14,  2.94it/s]
                                                   

 26%|██▌       | 1485/5758 [08:27<24:14,  2.94it/s]
 26%|██▌       | 1486/5758 [08:27<24:07,  2.95it/s]
                                                   

 26%|██▌       | 1486/5758 [08:27<24:07,  2.95it/s]
 26%|██▌       | 1487/5758 [08:28<24:44,  2.88it/s]
                                                   

 26%|██▌       | 1487/5758 [08:28<24:44,  2.88it/s]
 26%|██▌       | 1488/5758 [08:28<24:00,  2.96it/s]
                                                   

 26%|██▌       | 1488/5758 [08:28<24:00,  2.96it/s]
 26%|██▌       | 1489/5758 [08:29<27:09,  2.62it/s]
                                                   

 26%|██▌       | 1489/5758 [08:29<27:09,  2.62it/s]
 26%|██▌       | 1490/5758 [08:29<27:35,  2.58it/s]
                                                   

 26%|██▌       | 1490/5758 [08:29<27:35,  2.58it/s]
 26%|██▌       | 1491/5758 [08:29<25:23,  2.80it/s]
                                                   

 26%|██▌       | 1491/5758 [08:29<25:23,  2.80it/s]
 26%|██▌       | 1492/5758 [08:30<25:17,  2.81it/s]
                                                   

 26%|██▌       | 1492/5758 [08:30<25:17,  2.81it/s]
 26%|██▌       | 1493/5758 [08:30<25:21,  2.80it/s]
                                                   

 26%|██▌       | 1493/5758 [08:30<25:21,  2.80it/s]
 26%|██▌       | 1494/5758 [08:30<25:21,  2.80it/s]
                                                   

 26%|██▌       | 1494/5758 [08:30<25:21,  2.80it/s]
 26%|██▌       | 1495/5758 [08:31<24:42,  2.88it/s]
                                                   

 26%|██▌       | 1495/5758 [08:31<24:42,  2.88it/s]
 26%|██▌       | 1496/5758 [08:31<23:21,  3.04it/s]
                                                   
{'loss': '1.002', 'grad_norm': '10.94', 'learning_rate': '7.499e-06', 'num_tokens': '9.399e+05', 'mean_token_accuracy': '0.7513', 'epoch': '0.2503'}
+{'loss': '0.6114', 'grad_norm': '20.12', 'learning_rate': '7.497e-06', 'num_tokens': '9.401e+05', 'mean_token_accuracy': '0.8727', 'epoch': '0.2504'}
+{'loss': '0.849', 'grad_norm': '16.62', 'learning_rate': '7.496e-06', 'num_tokens': '9.404e+05', 'mean_token_accuracy': '0.8308', 'epoch': '0.2506'}
+{'loss': '0.9828', 'grad_norm': '15', 'learning_rate': '7.494e-06', 'num_tokens': '9.408e+05', 'mean_token_accuracy': '0.7636', 'epoch': '0.2508'}
+{'loss': '0.7683', 'grad_norm': '8.875', 'learning_rate': '7.492e-06', 'num_tokens': '9.423e+05', 'mean_token_accuracy': '0.7559', 'epoch': '0.251'}
+{'loss': '2.113', 'grad_norm': '17.25', 'learning_rate': '7.49e-06', 'num_tokens': '9.429e+05', 'mean_token_accuracy': '0.5905', 'epoch': '0.2511'}
+{'loss': '0.9985', 'grad_norm': '11.12', 'learning_rate': '7.489e-06', 'num_tokens': '9.438e+05', 'mean_token_accuracy': '0.7097', 'epoch': '0.2513'}
+{'loss': '0.8528', 'grad_norm': '16.62', 'learning_rate': '7.487e-06', 'num_tokens': '9.441e+05', 'mean_token_accuracy': '0.8133', 'epoch': '0.2515'}
+{'loss': '2.054', 'grad_norm': '18.5', 'learning_rate': '7.485e-06', 'num_tokens': '9.446e+05', 'mean_token_accuracy': '0.6008', 'epoch': '0.2516'}
+{'loss': '1.036', 'grad_norm': '13.06', 'learning_rate': '7.484e-06', 'num_tokens': '9.453e+05', 'mean_token_accuracy': '0.7068', 'epoch': '0.2518'}
+{'loss': '1.785', 'grad_norm': '10.38', 'learning_rate': '7.482e-06', 'num_tokens': '9.461e+05', 'mean_token_accuracy': '0.6484', 'epoch': '0.252'}
+{'loss': '0.6279', 'grad_norm': '49.25', 'learning_rate': '7.48e-06', 'num_tokens': '9.462e+05', 'mean_token_accuracy': '0.8364', 'epoch': '0.2522'}
+{'loss': '1.574', 'grad_norm': '12.25', 'learning_rate': '7.478e-06', 'num_tokens': '9.469e+05', 'mean_token_accuracy': '0.6688', 'epoch': '0.2523'}
+{'loss': '1.05', 'grad_norm': '13.19', 'learning_rate': '7.477e-06', 'num_tokens': '9.474e+05', 'mean_token_accuracy': '0.765', 'epoch': '0.2525'}
+{'loss': '2.336', 'grad_norm': '13.75', 'learning_rate': '7.475e-06', 'num_tokens': '9.481e+05', 'mean_token_accuracy': '0.5275', 'epoch': '0.2527'}
+{'loss': '1.258', 'grad_norm': '7.719', 'learning_rate': '7.473e-06', 'num_tokens': '9.499e+05', 'mean_token_accuracy': '0.689', 'epoch': '0.2529'}
+{'loss': '2.419', 'grad_norm': '15', 'learning_rate': '7.471e-06', 'num_tokens': '9.505e+05', 'mean_token_accuracy': '0.5644', 'epoch': '0.253'}
+{'loss': '2.14', 'grad_norm': '28', 'learning_rate': '7.47e-06', 'num_tokens': '9.508e+05', 'mean_token_accuracy': '0.5419', 'epoch': '0.2532'}
+{'loss': '0.6574', 'grad_norm': '7.844', 'learning_rate': '7.468e-06', 'num_tokens': '9.522e+05', 'mean_token_accuracy': '0.8083', 'epoch': '0.2534'}
+{'loss': '1.42', 'grad_norm': '11.44', 'learning_rate': '7.466e-06', 'num_tokens': '9.531e+05', 'mean_token_accuracy': '0.6203', 'epoch': '0.2536'}
+{'loss': '1.269', 'grad_norm': '11.69', 'learning_rate': '7.464e-06', 'num_tokens': '9.54e+05', 'mean_token_accuracy': '0.659', 'epoch': '0.2537'}
+{'loss': '2.152', 'grad_norm': '14.12', 'learning_rate': '7.463e-06', 'num_tokens': '9.549e+05', 'mean_token_accuracy': '0.5556', 'epoch': '0.2539'}
+{'loss': '1.872', 'grad_norm': '20', 'learning_rate': '7.461e-06', 'num_tokens': '9.551e+05', 'mean_token_accuracy': '0.5918', 'epoch': '0.2541'}
+{'loss': '1.593', 'grad_norm': '23.5', 'learning_rate': '7.459e-06', 'num_tokens': '9.555e+05', 'mean_token_accuracy': '0.6938', 'epoch': '0.2543'}
+{'loss': '1.105', 'grad_norm': '20.12', 'learning_rate': '7.457e-06', 'num_tokens': '9.557e+05', 'mean_token_accuracy': '0.7536', 'epoch': '0.2544'}
+{'loss': '1.154', 'grad_norm': '13.06', 'learning_rate': '7.456e-06', 'num_tokens': '9.563e+05', 'mean_token_accuracy': '0.6683', 'epoch': '0.2546'}
+{'loss': '1.636', 'grad_norm': '10.25', 'learning_rate': '7.454e-06', 'num_tokens': '9.572e+05', 'mean_token_accuracy': '0.6667', 'epoch': '0.2548'}
+{'loss': '1.892', 'grad_norm': '15.94', 'learning_rate': '7.452e-06', 'num_tokens': '9.575e+05', 'mean_token_accuracy': '0.6528', 'epoch': '0.2549'}
+{'loss': '0.6591', 'grad_norm': '21.38', 'learning_rate': '7.451e-06', 'num_tokens': '9.577e+05', 'mean_token_accuracy': '0.8634', 'epoch': '0.2551'}
+{'loss': '1.065', 'grad_norm': '13.5', 'learning_rate': '7.449e-06', 'num_tokens': '9.583e+05', 'mean_token_accuracy': '0.6868', 'epoch': '0.2553'}
+{'loss': '0.5961', 'grad_norm': '13.06', 'learning_rate': '7.447e-06', 'num_tokens': '9.588e+05', 'mean_token_accuracy': '0.8268', 'epoch': '0.2555'}
+{'loss': '1.408', 'grad_norm': '13.38', 'learning_rate': '7.445e-06', 'num_tokens': '9.594e+05', 'mean_token_accuracy': '0.6055', 'epoch': '0.2556'}
+{'loss': '0.7445', 'grad_norm': '13.75', 'learning_rate': '7.444e-06', 'num_tokens': '9.599e+05', 'mean_token_accuracy': '0.8037', 'epoch': '0.2558'}
+{'loss': '0.548', 'grad_norm': '14.56', 'learning_rate': '7.442e-06', 'num_tokens': '9.601e+05', 'mean_token_accuracy': '0.8795', 'epoch': '0.256'}
+{'loss': '1.168', 'grad_norm': '13.12', 'learning_rate': '7.44e-06', 'num_tokens': '9.607e+05', 'mean_token_accuracy': '0.6636', 'epoch': '0.2562'}
+{'loss': '0.9513', 'grad_norm': '19.5', 'learning_rate': '7.438e-06', 'num_tokens': '9.61e+05', 'mean_token_accuracy': '0.7655', 'epoch': '0.2563'}
+{'loss': '2.212', 'grad_norm': '20.62', 'learning_rate': '7.437e-06', 'num_tokens': '9.614e+05', 'mean_token_accuracy': '0.5534', 'epoch': '0.2565'}
+{'loss': '1.85', 'grad_norm': '22.38', 'learning_rate': '7.435e-06', 'num_tokens': '9.617e+05', 'mean_token_accuracy': '0.6281', 'epoch': '0.2567'}
+{'loss': '1.144', 'grad_norm': '19.38', 'learning_rate': '7.433e-06', 'num_tokens': '9.62e+05', 'mean_token_accuracy': '0.7559', 'epoch': '0.2569'}
+{'loss': '1.212', 'grad_norm': '14.44', 'learning_rate': '7.431e-06', 'num_tokens': '9.626e+05', 'mean_token_accuracy': '0.6685', 'epoch': '0.257'}
+{'loss': '1.535', 'grad_norm': '12.12', 'learning_rate': '7.43e-06', 'num_tokens': '9.634e+05', 'mean_token_accuracy': '0.6321', 'epoch': '0.2572'}
+{'loss': '1.783', 'grad_norm': '10.06', 'learning_rate': '7.428e-06', 'num_tokens': '9.644e+05', 'mean_token_accuracy': '0.6414', 'epoch': '0.2574'}
+{'loss': '1.067', 'grad_norm': '13.19', 'learning_rate': '7.426e-06', 'num_tokens': '9.65e+05', 'mean_token_accuracy': '0.7088', 'epoch': '0.2576'}
+{'loss': '1.163', 'grad_norm': '29', 'learning_rate': '7.424e-06', 'num_tokens': '9.652e+05', 'mean_token_accuracy': '0.7007', 'epoch': '0.2577'}
+{'loss': '1.247', 'grad_norm': '11.5', 'learning_rate': '7.423e-06', 'num_tokens': '9.66e+05', 'mean_token_accuracy': '0.6478', 'epoch': '0.2579'}
+{'loss': '1.194', 'grad_norm': '10.69', 'learning_rate': '7.421e-06', 'num_tokens': '9.668e+05', 'mean_token_accuracy': '0.6854', 'epoch': '0.2581'}
+{'loss': '0.9911', 'grad_norm': '14', 'learning_rate': '7.419e-06', 'num_tokens': '9.674e+05', 'mean_token_accuracy': '0.719', 'epoch': '0.2582'}
+{'loss': '1.457', 'grad_norm': '14.12', 'learning_rate': '7.418e-06', 'num_tokens': '9.681e+05', 'mean_token_accuracy': '0.6123', 'epoch': '0.2584'}
+{'loss': '1.73', 'grad_norm': '21.12', 'learning_rate': '7.416e-06', 'num_tokens': '9.683e+05', 'mean_token_accuracy': '0.6448', 'epoch': '0.2586'}
+{'loss': '1.155', 'grad_norm': '26.25', 'learning_rate': '7.414e-06', 'num_tokens': '9.685e+05', 'mean_token_accuracy': '0.7606', 'epoch': '0.2588'}
+{'loss': '0.6987', 'grad_norm': '9.188', 'learning_rate': '7.412e-06', 'num_tokens': '9.697e+05', 'mean_token_accuracy': '0.7868', 'epoch': '0.2589'}
+{'loss': '0.5654', 'grad_norm': '13.81', 'learning_rate': '7.411e-06', 'num_tokens': '9.699e+05', 'mean_token_accuracy': '0.8811', 'epoch': '0.2591'}
+{'loss': '0.904', 'grad_norm': '17', 'learning_rate': '7.409e-06', 'num_tokens': '9.702e+05', 'mean_token_accuracy': '0.7747', 'epoch': '0.2593'}
+{'loss': '0.9308', 'grad_norm': '21.12', 'learning_rate': '7.407e-06', 'num_tokens': '9.704e+05', 'mean_token_accuracy': '0.8063', 'epoch': '0.2595'}
+{'loss': '1.515', 'grad_norm': '13.19', 'learning_rate': '7.405e-06', 'num_tokens': '9.713e+05', 'mean_token_accuracy': '0.615', 'epoch': '0.2596'}
+
 26%|██▌       | 1496/5758 [08:31<23:21,  3.04it/s]
 26%|██▌       | 1497/5758 [08:31<24:10,  2.94it/s]
                                                   

 26%|██▌       | 1497/5758 [08:31<24:10,  2.94it/s]
 26%|██▌       | 1498/5758 [08:32<25:06,  2.83it/s]
                                                   

 26%|██▌       | 1498/5758 [08:32<25:06,  2.83it/s]
 26%|██▌       | 1499/5758 [08:32<24:35,  2.89it/s]
                                                   

 26%|██▌       | 1499/5758 [08:32<24:35,  2.89it/s]
 26%|██▌       | 1500/5758 [08:32<24:55,  2.85it/s]
                                                   

 26%|██▌       | 1500/5758 [08:32<24:55,  2.85it/s]{'loss': '0.9191', 'grad_norm': '20', 'learning_rate': '7.404e-06', 'num_tokens': '9.716e+05', 'mean_token_accuracy': '0.7867', 'epoch': '0.2598'}
+{'loss': '1.296', 'grad_norm': '10.19', 'learning_rate': '7.402e-06', 'num_tokens': '9.727e+05', 'mean_token_accuracy': '0.6385', 'epoch': '0.26'}
+{'loss': '1.251', 'grad_norm': '13.81', 'learning_rate': '7.4e-06', 'num_tokens': '9.732e+05', 'mean_token_accuracy': '0.6559', 'epoch': '0.2602'}
+{'loss': '0.9334', 'grad_norm': '16.88', 'learning_rate': '7.398e-06', 'num_tokens': '9.737e+05', 'mean_token_accuracy': '0.8195', 'epoch': '0.2603'}
+{'loss': '0.6263', 'grad_norm': '5.406', 'learning_rate': '7.397e-06', 'num_tokens': '9.778e+05', 'mean_token_accuracy': '0.8212', 'epoch': '0.2605'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.60it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.60it/s]
+
 26%|██▌       | 1501/5758 [08:34<1:00:38,  1.17it/s]
                                                     

 26%|██▌       | 1501/5758 [08:34<1:00:38,  1.17it/s]
 26%|██▌       | 1502/5758 [08:35<50:15,  1.41it/s]  
                                                   

 26%|██▌       | 1502/5758 [08:35<50:15,  1.41it/s]
 26%|██▌       | 1503/5758 [08:35<41:30,  1.71it/s]
                                                   

 26%|██▌       | 1503/5758 [08:35<41:30,  1.71it/s]
 26%|██▌       | 1504/5758 [08:35<35:51,  1.98it/s]
                                                   

 26%|██▌       | 1504/5758 [08:35<35:51,  1.98it/s]
 26%|██▌       | 1505/5758 [08:36<32:13,  2.20it/s]
                                                   

 26%|██▌       | 1505/5758 [08:36<32:13,  2.20it/s]
 26%|██▌       | 1506/5758 [08:36<28:42,  2.47it/s]
                                                   

 26%|██▌       | 1506/5758 [08:36<28:42,  2.47it/s]
 26%|██▌       | 1507/5758 [08:36<27:22,  2.59it/s]
                                                   

 26%|██▌       | 1507/5758 [08:36<27:22,  2.59it/s]
 26%|██▌       | 1508/5758 [08:37<26:51,  2.64it/s]
                                                   

 26%|██▌       | 1508/5758 [08:37<26:51,  2.64it/s]
 26%|██▌       | 1509/5758 [08:37<25:07,  2.82it/s]
                                                   

 26%|██▌       | 1509/5758 [08:37<25:07,  2.82it/s]
 26%|██▌       | 1510/5758 [08:37<23:49,  2.97it/s]
                                                   

 26%|██▌       | 1510/5758 [08:37<23:49,  2.97it/s]
 26%|██▌       | 1511/5758 [08:38<23:30,  3.01it/s]
                                                   

 26%|██▌       | 1511/5758 [08:38<23:30,  3.01it/s]
 26%|██▋       | 1512/5758 [08:38<23:22,  3.03it/s]
                                                   

 26%|██▋       | 1512/5758 [08:38<23:22,  3.03it/s]
 26%|██▋       | 1513/5758 [08:38<23:05,  3.06it/s]
                                                   

 26%|██▋       | 1513/5758 [08:38<23:05,  3.06it/s]
 26%|██▋       | 1514/5758 [08:39<23:38,  2.99it/s]
                                                   

 26%|██▋       | 1514/5758 [08:39<23:38,  2.99it/s]
 26%|██▋       | 1515/5758 [08:39<23:06,  3.06it/s]
                                                   

 26%|██▋       | 1515/5758 [08:39<23:06,  3.06it/s]
 26%|██▋       | 1516/5758 [08:39<26:42,  2.65it/s]
                                                   

 26%|██▋       | 1516/5758 [08:39<26:42,  2.65it/s]
 26%|██▋       | 1517/5758 [08:40<25:27,  2.78it/s]
                                                   

 26%|██▋       | 1517/5758 [08:40<25:27,  2.78it/s]
 26%|██▋       | 1518/5758 [08:40<25:16,  2.80it/s]
                                                   

 26%|██▋       | 1518/5758 [08:40<25:16,  2.80it/s]
 26%|██▋       | 1519/5758 [08:41<25:31,  2.77it/s]
                                                   

 26%|██▋       | 1519/5758 [08:41<25:31,  2.77it/s]
 26%|██▋       | 1520/5758 [08:41<25:18,  2.79it/s]
                                                   

 26%|██▋       | 1520/5758 [08:41<25:18,  2.79it/s]
 26%|██▋       | 1521/5758 [08:41<23:53,  2.96it/s]
                                                   

 26%|██▋       | 1521/5758 [08:41<23:53,  2.96it/s]
 26%|██▋       | 1522/5758 [08:42<23:39,  2.98it/s]
                                                   

 26%|██▋       | 1522/5758 [08:42<23:39,  2.98it/s]
 26%|██▋       | 1523/5758 [08:42<22:53,  3.08it/s]
                                                   

 26%|██▋       | 1523/5758 [08:42<22:53,  3.08it/s]
 26%|██▋       | 1524/5758 [08:42<22:38,  3.12it/s]
                                                   

 26%|██▋       | 1524/5758 [08:42<22:38,  3.12it/s]
 26%|██▋       | 1525/5758 [08:42<22:05,  3.19it/s]
                                                   

 26%|██▋       | 1525/5758 [08:42<22:05,  3.19it/s]
 27%|██▋       | 1526/5758 [08:43<21:39,  3.26it/s]
                                                   

 27%|██▋       | 1526/5758 [08:43<21:39,  3.26it/s]
 27%|██▋       | 1527/5758 [08:43<22:53,  3.08it/s]
                                                   

 27%|██▋       | 1527/5758 [08:43<22:53,  3.08it/s]
 27%|██▋       | 1528/5758 [08:43<23:36,  2.99it/s]
                                                   

 27%|██▋       | 1528/5758 [08:43<23:36,  2.99it/s]
 27%|██▋       | 1529/5758 [08:44<25:10,  2.80it/s]
                                                   

 27%|██▋       | 1529/5758 [08:44<25:10,  2.80it/s]
 27%|██▋       | 1530/5758 [08:44<26:16,  2.68it/s]
                                                   

 27%|██▋       | 1530/5758 [08:44<26:16,  2.68it/s]
 27%|██▋       | 1531/5758 [08:45<25:14,  2.79it/s]
                                                   

 27%|██▋       | 1531/5758 [08:45<25:14,  2.79it/s]
 27%|██▋       | 1532/5758 [08:45<24:19,  2.90it/s]
                                                   

 27%|██▋       | 1532/5758 [08:45<24:19,  2.90it/s]
 27%|██▋       | 1533/5758 [08:45<25:21,  2.78it/s]
                                                   

 27%|██▋       | 1533/5758 [08:45<25:21,  2.78it/s]
 27%|██▋       | 1534/5758 [08:46<24:23,  2.89it/s]
                                                   

 27%|██▋       | 1534/5758 [08:46<24:23,  2.89it/s]
 27%|██▋       | 1535/5758 [08:46<24:33,  2.86it/s]
                                                   

 27%|██▋       | 1535/5758 [08:46<24:33,  2.86it/s]
 27%|██▋       | 1536/5758 [08:46<23:44,  2.96it/s]
                                                   

 27%|██▋       | 1536/5758 [08:46<23:44,  2.96it/s]
 27%|██▋       | 1537/5758 [08:47<22:50,  3.08it/s]
                                                   

 27%|██▋       | 1537/5758 [08:47<22:50,  3.08it/s]
 27%|██▋       | 1538/5758 [08:47<22:07,  3.18it/s]
                                                   

 27%|██▋       | 1538/5758 [08:47<22:07,  3.18it/s]
 27%|██▋       | 1539/5758 [08:47<22:39,  3.10it/s]
                                                   

 27%|██▋       | 1539/5758 [08:47<22:39,  3.10it/s]
 27%|██▋       | 1540/5758 [08:48<22:53,  3.07it/s]
                                                   

 27%|██▋       | 1540/5758 [08:48<22:53,  3.07it/s]
 27%|██▋       | 1541/5758 [08:48<22:15,  3.16it/s]
                                                   

 27%|██▋       | 1541/5758 [08:48<22:15,  3.16it/s]
 27%|██▋       | 1542/5758 [08:48<21:45,  3.23it/s]
                                                   

 27%|██▋       | 1542/5758 [08:48<21:45,  3.23it/s]
 27%|██▋       | 1543/5758 [08:48<23:11,  3.03it/s]
                                                   

 27%|██▋       | 1543/5758 [08:48<23:11,  3.03it/s]
 27%|██▋       | 1544/5758 [08:49<24:47,  2.83it/s]
                                                   

 27%|██▋       | 1544/5758 [08:49<24:47,  2.83it/s]
 27%|██▋       | 1545/5758 [08:49<26:36,  2.64it/s]
                                                   

 27%|██▋       | 1545/5758 [08:49<26:36,  2.64it/s]
 27%|██▋       | 1546/5758 [08:50<25:37,  2.74it/s]
                                                   

 27%|██▋       | 1546/5758 [08:50<25:37,  2.74it/s]
 27%|██▋       | 1547/5758 [08:50<26:09,  2.68it/s]
                                                   

 27%|██▋       | 1547/5758 [08:50<26:09,  2.68it/s]
 27%|██▋       | 1548/5758 [08:50<25:16,  2.78it/s]
                                                   

 27%|██▋       | 1548/5758 [08:50<25:16,  2.78it/s]
 27%|██▋       | 1549/5758 [08:51<24:38,  2.85it/s]
                                                   

 27%|██▋       | 1549/5758 [08:51<24:38,  2.85it/s]
 27%|██▋       | 1550/5758 [08:51<23:53,  2.94it/s]
                                                   

 27%|██▋       | 1550/5758 [08:51<23:53,  2.94it/s]
 27%|██▋       | 1551/5758 [08:51<24:00,  2.92it/s]
                                                   

 27%|██▋       | 1551/5758 [08:51<24:00,  2.92it/s]
 27%|██▋       | 1552/5758 [08:52<25:35,  2.74it/s]
                                                   

 27%|██▋       | 1552/5758 [08:52<25:35,  2.74it/s]
 27%|██▋       | 1553/5758 [08:52<25:14,  2.78it/s]
                                                   

 27%|██▋       | 1553/5758 [08:52<25:14,  2.78it/s]
 27%|██▋       | 1554/5758 [08:53<25:53,  2.71it/s]
                                                   

 27%|██▋       | 1554/5758 [08:53<25:53,  2.71it/s]
 27%|██▋       | 1555/5758 [08:53<24:58,  2.80it/s]
                                                   

 27%|██▋       | 1555/5758 [08:53<24:58,  2.80it/s]
 27%|██▋       | 1556/5758 [08:53<24:56,  2.81it/s]
                                                   
{'loss': '1.386', 'grad_norm': '12.5', 'learning_rate': '7.395e-06', 'num_tokens': '9.784e+05', 'mean_token_accuracy': '0.6382', 'epoch': '0.2607'}
+{'loss': '0.6051', 'grad_norm': '7.062', 'learning_rate': '7.393e-06', 'num_tokens': '9.805e+05', 'mean_token_accuracy': '0.8194', 'epoch': '0.2609'}
+{'loss': '1.464', 'grad_norm': '40.5', 'learning_rate': '7.391e-06', 'num_tokens': '9.806e+05', 'mean_token_accuracy': '0.6571', 'epoch': '0.261'}
+{'loss': '1.287', 'grad_norm': '11.75', 'learning_rate': '7.39e-06', 'num_tokens': '9.814e+05', 'mean_token_accuracy': '0.6658', 'epoch': '0.2612'}
+{'loss': '1.977', 'grad_norm': '10.5', 'learning_rate': '7.388e-06', 'num_tokens': '9.822e+05', 'mean_token_accuracy': '0.5915', 'epoch': '0.2614'}
+{'loss': '0.9727', 'grad_norm': '37.25', 'learning_rate': '7.386e-06', 'num_tokens': '9.823e+05', 'mean_token_accuracy': '0.7746', 'epoch': '0.2615'}
+{'loss': '2.117', 'grad_norm': '22.38', 'learning_rate': '7.385e-06', 'num_tokens': '9.828e+05', 'mean_token_accuracy': '0.5594', 'epoch': '0.2617'}
+{'loss': '2.027', 'grad_norm': '11.19', 'learning_rate': '7.383e-06', 'num_tokens': '9.838e+05', 'mean_token_accuracy': '0.5898', 'epoch': '0.2619'}
+{'loss': '0.9302', 'grad_norm': '25.5', 'learning_rate': '7.381e-06', 'num_tokens': '9.84e+05', 'mean_token_accuracy': '0.7853', 'epoch': '0.2621'}
+{'loss': '1.55', 'grad_norm': '13.56', 'learning_rate': '7.379e-06', 'num_tokens': '9.846e+05', 'mean_token_accuracy': '0.6042', 'epoch': '0.2622'}
+{'loss': '0.7398', 'grad_norm': '13.25', 'learning_rate': '7.378e-06', 'num_tokens': '9.85e+05', 'mean_token_accuracy': '0.8382', 'epoch': '0.2624'}
+{'loss': '2.084', 'grad_norm': '28', 'learning_rate': '7.376e-06', 'num_tokens': '9.852e+05', 'mean_token_accuracy': '0.6294', 'epoch': '0.2626'}
+{'loss': '0.9554', 'grad_norm': '34', 'learning_rate': '7.374e-06', 'num_tokens': '9.852e+05', 'mean_token_accuracy': '0.8068', 'epoch': '0.2628'}
+{'loss': '1.228', 'grad_norm': '17.25', 'learning_rate': '7.372e-06', 'num_tokens': '9.855e+05', 'mean_token_accuracy': '0.7542', 'epoch': '0.2629'}
+{'loss': '1.257', 'grad_norm': '13.12', 'learning_rate': '7.371e-06', 'num_tokens': '9.862e+05', 'mean_token_accuracy': '0.6297', 'epoch': '0.2631'}
+{'loss': '0.6332', 'grad_norm': '8.812', 'learning_rate': '7.369e-06', 'num_tokens': '9.873e+05', 'mean_token_accuracy': '0.7982', 'epoch': '0.2633'}
+{'loss': '1.417', 'grad_norm': '13', 'learning_rate': '7.367e-06', 'num_tokens': '9.881e+05', 'mean_token_accuracy': '0.6218', 'epoch': '0.2635'}
+{'loss': '1.78', 'grad_norm': '17.75', 'learning_rate': '7.365e-06', 'num_tokens': '9.886e+05', 'mean_token_accuracy': '0.6453', 'epoch': '0.2636'}
+{'loss': '0.9952', 'grad_norm': '9.125', 'learning_rate': '7.364e-06', 'num_tokens': '9.897e+05', 'mean_token_accuracy': '0.7581', 'epoch': '0.2638'}
+{'loss': '1.623', 'grad_norm': '14.94', 'learning_rate': '7.362e-06', 'num_tokens': '9.902e+05', 'mean_token_accuracy': '0.6206', 'epoch': '0.264'}
+{'loss': '1.141', 'grad_norm': '8.938', 'learning_rate': '7.36e-06', 'num_tokens': '9.914e+05', 'mean_token_accuracy': '0.7203', 'epoch': '0.2642'}
+{'loss': '1.039', 'grad_norm': '10.94', 'learning_rate': '7.358e-06', 'num_tokens': '9.922e+05', 'mean_token_accuracy': '0.702', 'epoch': '0.2643'}
+{'loss': '1.387', 'grad_norm': '19.38', 'learning_rate': '7.357e-06', 'num_tokens': '9.926e+05', 'mean_token_accuracy': '0.7287', 'epoch': '0.2645'}
+{'loss': '1.397', 'grad_norm': '12.94', 'learning_rate': '7.355e-06', 'num_tokens': '9.933e+05', 'mean_token_accuracy': '0.6226', 'epoch': '0.2647'}
+{'loss': '1.316', 'grad_norm': '14', 'learning_rate': '7.353e-06', 'num_tokens': '9.939e+05', 'mean_token_accuracy': '0.6242', 'epoch': '0.2648'}
+{'loss': '2.394', 'grad_norm': '21.88', 'learning_rate': '7.352e-06', 'num_tokens': '9.943e+05', 'mean_token_accuracy': '0.5405', 'epoch': '0.265'}
+{'loss': '1.232', 'grad_norm': '12.69', 'learning_rate': '7.35e-06', 'num_tokens': '9.951e+05', 'mean_token_accuracy': '0.6551', 'epoch': '0.2652'}
+{'loss': '0.5918', 'grad_norm': '6.656', 'learning_rate': '7.348e-06', 'num_tokens': '9.969e+05', 'mean_token_accuracy': '0.8246', 'epoch': '0.2654'}
+{'loss': '1.201', 'grad_norm': '13.94', 'learning_rate': '7.346e-06', 'num_tokens': '9.975e+05', 'mean_token_accuracy': '0.673', 'epoch': '0.2655'}
+{'loss': '1.431', 'grad_norm': '41.75', 'learning_rate': '7.345e-06', 'num_tokens': '9.976e+05', 'mean_token_accuracy': '0.7288', 'epoch': '0.2657'}
+{'loss': '1.385', 'grad_norm': '12.81', 'learning_rate': '7.343e-06', 'num_tokens': '9.983e+05', 'mean_token_accuracy': '0.6502', 'epoch': '0.2659'}
+{'loss': '1.306', 'grad_norm': '11.56', 'learning_rate': '7.341e-06', 'num_tokens': '9.99e+05', 'mean_token_accuracy': '0.6594', 'epoch': '0.2661'}
+{'loss': '1.935', 'grad_norm': '12.88', 'learning_rate': '7.339e-06', 'num_tokens': '9.996e+05', 'mean_token_accuracy': '0.6214', 'epoch': '0.2662'}
+{'loss': '1.28', 'grad_norm': '12.81', 'learning_rate': '7.338e-06', 'num_tokens': '1e+06', 'mean_token_accuracy': '0.6585', 'epoch': '0.2664'}
+{'loss': '1.559', 'grad_norm': '37', 'learning_rate': '7.336e-06', 'num_tokens': '1e+06', 'mean_token_accuracy': '0.7037', 'epoch': '0.2666'}
+{'loss': '1.45', 'grad_norm': '8.75', 'learning_rate': '7.334e-06', 'num_tokens': '1.002e+06', 'mean_token_accuracy': '0.659', 'epoch': '0.2668'}
+{'loss': '0.9658', 'grad_norm': '12.31', 'learning_rate': '7.332e-06', 'num_tokens': '1.002e+06', 'mean_token_accuracy': '0.7176', 'epoch': '0.2669'}
+{'loss': '1.778', 'grad_norm': '13.69', 'learning_rate': '7.331e-06', 'num_tokens': '1.003e+06', 'mean_token_accuracy': '0.6384', 'epoch': '0.2671'}
+{'loss': '0.6179', 'grad_norm': '6.625', 'learning_rate': '7.329e-06', 'num_tokens': '1.005e+06', 'mean_token_accuracy': '0.8122', 'epoch': '0.2673'}
+{'loss': '0.5999', 'grad_norm': '6.781', 'learning_rate': '7.327e-06', 'num_tokens': '1.007e+06', 'mean_token_accuracy': '0.8231', 'epoch': '0.2675'}
+{'loss': '1.711', 'grad_norm': '11.69', 'learning_rate': '7.325e-06', 'num_tokens': '1.007e+06', 'mean_token_accuracy': '0.6322', 'epoch': '0.2676'}
+{'loss': '1.276', 'grad_norm': '10.12', 'learning_rate': '7.324e-06', 'num_tokens': '1.009e+06', 'mean_token_accuracy': '0.6587', 'epoch': '0.2678'}
+{'loss': '1.102', 'grad_norm': '13.38', 'learning_rate': '7.322e-06', 'num_tokens': '1.009e+06', 'mean_token_accuracy': '0.6673', 'epoch': '0.268'}
+{'loss': '1.385', 'grad_norm': '30.88', 'learning_rate': '7.32e-06', 'num_tokens': '1.009e+06', 'mean_token_accuracy': '0.6738', 'epoch': '0.2681'}
+{'loss': '1.238', 'grad_norm': '8.125', 'learning_rate': '7.319e-06', 'num_tokens': '1.011e+06', 'mean_token_accuracy': '0.7092', 'epoch': '0.2683'}
+{'loss': '1.701', 'grad_norm': '11.06', 'learning_rate': '7.317e-06', 'num_tokens': '1.012e+06', 'mean_token_accuracy': '0.6543', 'epoch': '0.2685'}
+{'loss': '2.064', 'grad_norm': '14.38', 'learning_rate': '7.315e-06', 'num_tokens': '1.012e+06', 'mean_token_accuracy': '0.5904', 'epoch': '0.2687'}
+{'loss': '1.397', 'grad_norm': '15.62', 'learning_rate': '7.313e-06', 'num_tokens': '1.013e+06', 'mean_token_accuracy': '0.7136', 'epoch': '0.2688'}
+{'loss': '0.8163', 'grad_norm': '14.88', 'learning_rate': '7.312e-06', 'num_tokens': '1.013e+06', 'mean_token_accuracy': '0.8146', 'epoch': '0.269'}
+{'loss': '1.226', 'grad_norm': '12.81', 'learning_rate': '7.31e-06', 'num_tokens': '1.014e+06', 'mean_token_accuracy': '0.6752', 'epoch': '0.2692'}
+{'loss': '1.507', 'grad_norm': '7.625', 'learning_rate': '7.308e-06', 'num_tokens': '1.015e+06', 'mean_token_accuracy': '0.6719', 'epoch': '0.2694'}
+{'loss': '1.527', 'grad_norm': '15.12', 'learning_rate': '7.306e-06', 'num_tokens': '1.016e+06', 'mean_token_accuracy': '0.6218', 'epoch': '0.2695'}
+{'loss': '2.036', 'grad_norm': '17.12', 'learning_rate': '7.305e-06', 'num_tokens': '1.016e+06', 'mean_token_accuracy': '0.6416', 'epoch': '0.2697'}
+{'loss': '0.6434', 'grad_norm': '24', 'learning_rate': '7.303e-06', 'num_tokens': '1.016e+06', 'mean_token_accuracy': '0.8519', 'epoch': '0.2699'}
+{'loss': '1.727', 'grad_norm': '12', 'learning_rate': '7.301e-06', 'num_tokens': '1.017e+06', 'mean_token_accuracy': '0.5899', 'epoch': '0.2701'}
+
 27%|██▋       | 1556/5758 [08:53<24:56,  2.81it/s]
 27%|██▋       | 1557/5758 [08:54<24:06,  2.90it/s]
                                                   

 27%|██▋       | 1557/5758 [08:54<24:06,  2.90it/s]
 27%|██▋       | 1558/5758 [08:54<24:07,  2.90it/s]
                                                   

 27%|██▋       | 1558/5758 [08:54<24:07,  2.90it/s]
 27%|██▋       | 1559/5758 [08:54<23:21,  3.00it/s]
                                                   

 27%|██▋       | 1559/5758 [08:54<23:21,  3.00it/s]
 27%|██▋       | 1560/5758 [08:55<23:17,  3.00it/s]
                                                   

 27%|██▋       | 1560/5758 [08:55<23:17,  3.00it/s]
 27%|██▋       | 1561/5758 [08:55<23:37,  2.96it/s]
                                                   

 27%|██▋       | 1561/5758 [08:55<23:37,  2.96it/s]
 27%|██▋       | 1562/5758 [08:55<23:00,  3.04it/s]
                                                   

 27%|██▋       | 1562/5758 [08:55<23:00,  3.04it/s]
 27%|██▋       | 1563/5758 [08:55<22:28,  3.11it/s]
                                                   

 27%|██▋       | 1563/5758 [08:55<22:28,  3.11it/s]
 27%|██▋       | 1564/5758 [08:56<24:37,  2.84it/s]
                                                   

 27%|██▋       | 1564/5758 [08:56<24:37,  2.84it/s]
 27%|██▋       | 1565/5758 [08:56<23:23,  2.99it/s]
                                                   

 27%|██▋       | 1565/5758 [08:56<23:23,  2.99it/s]
 27%|██▋       | 1566/5758 [08:57<23:28,  2.98it/s]
                                                   

 27%|██▋       | 1566/5758 [08:57<23:28,  2.98it/s]
 27%|██▋       | 1567/5758 [08:57<26:22,  2.65it/s]
                                                   

 27%|██▋       | 1567/5758 [08:57<26:22,  2.65it/s]
 27%|██▋       | 1568/5758 [08:57<26:36,  2.62it/s]
                                                   

 27%|██▋       | 1568/5758 [08:57<26:36,  2.62it/s]
 27%|██▋       | 1569/5758 [08:58<24:48,  2.81it/s]
                                                   

 27%|██▋       | 1569/5758 [08:58<24:48,  2.81it/s]
 27%|██▋       | 1570/5758 [08:58<25:13,  2.77it/s]
                                                   

 27%|██▋       | 1570/5758 [08:58<25:13,  2.77it/s]
 27%|██▋       | 1571/5758 [08:58<25:09,  2.77it/s]
                                                   

 27%|██▋       | 1571/5758 [08:58<25:09,  2.77it/s]
 27%|██▋       | 1572/5758 [08:59<24:14,  2.88it/s]
                                                   

 27%|██▋       | 1572/5758 [08:59<24:14,  2.88it/s]
 27%|██▋       | 1573/5758 [08:59<23:16,  3.00it/s]
                                                   

 27%|██▋       | 1573/5758 [08:59<23:16,  3.00it/s]
 27%|██▋       | 1574/5758 [08:59<22:47,  3.06it/s]
                                                   

 27%|██▋       | 1574/5758 [08:59<22:47,  3.06it/s]
 27%|██▋       | 1575/5758 [09:00<22:11,  3.14it/s]
                                                   

 27%|██▋       | 1575/5758 [09:00<22:11,  3.14it/s]
 27%|██▋       | 1576/5758 [09:00<21:42,  3.21it/s]
                                                   

 27%|██▋       | 1576/5758 [09:00<21:42,  3.21it/s]
 27%|██▋       | 1577/5758 [09:00<23:03,  3.02it/s]
                                                   

 27%|██▋       | 1577/5758 [09:00<23:03,  3.02it/s]
 27%|██▋       | 1578/5758 [09:01<23:09,  3.01it/s]
                                                   

 27%|██▋       | 1578/5758 [09:01<23:09,  3.01it/s]
 27%|██▋       | 1579/5758 [09:01<22:23,  3.11it/s]
                                                   

 27%|██▋       | 1579/5758 [09:01<22:23,  3.11it/s]
 27%|██▋       | 1580/5758 [09:01<22:24,  3.11it/s]
                                                   

 27%|██▋       | 1580/5758 [09:01<22:24,  3.11it/s]
 27%|██▋       | 1581/5758 [09:02<21:47,  3.19it/s]
                                                   

 27%|██▋       | 1581/5758 [09:02<21:47,  3.19it/s]
 27%|██▋       | 1582/5758 [09:02<22:56,  3.03it/s]
                                                   

 27%|██▋       | 1582/5758 [09:02<22:56,  3.03it/s]
 27%|██▋       | 1583/5758 [09:02<22:46,  3.06it/s]
                                                   

 27%|██▋       | 1583/5758 [09:02<22:46,  3.06it/s]
 28%|██▊       | 1584/5758 [09:03<23:14,  2.99it/s]
                                                   

 28%|██▊       | 1584/5758 [09:03<23:14,  2.99it/s]
 28%|██▊       | 1585/5758 [09:03<24:19,  2.86it/s]
                                                   

 28%|██▊       | 1585/5758 [09:03<24:19,  2.86it/s]
 28%|██▊       | 1586/5758 [09:03<23:09,  3.00it/s]
                                                   

 28%|██▊       | 1586/5758 [09:03<23:09,  3.00it/s]
 28%|██▊       | 1587/5758 [09:04<23:05,  3.01it/s]
                                                   

 28%|██▊       | 1587/5758 [09:04<23:05,  3.01it/s]
 28%|██▊       | 1588/5758 [09:04<22:52,  3.04it/s]
                                                   

 28%|██▊       | 1588/5758 [09:04<22:52,  3.04it/s]
 28%|██▊       | 1589/5758 [09:04<22:20,  3.11it/s]
                                                   

 28%|██▊       | 1589/5758 [09:04<22:20,  3.11it/s]
 28%|██▊       | 1590/5758 [09:05<22:48,  3.05it/s]
                                                   

 28%|██▊       | 1590/5758 [09:05<22:48,  3.05it/s]
 28%|██▊       | 1591/5758 [09:05<23:44,  2.93it/s]
                                                   

 28%|██▊       | 1591/5758 [09:05<23:44,  2.93it/s]
 28%|██▊       | 1592/5758 [09:05<22:34,  3.08it/s]
                                                   

 28%|██▊       | 1592/5758 [09:05<22:34,  3.08it/s]
 28%|██▊       | 1593/5758 [09:06<24:34,  2.82it/s]
                                                   

 28%|██▊       | 1593/5758 [09:06<24:34,  2.82it/s]
 28%|██▊       | 1594/5758 [09:06<23:16,  2.98it/s]
                                                   

 28%|██▊       | 1594/5758 [09:06<23:16,  2.98it/s]
 28%|██▊       | 1595/5758 [09:06<23:35,  2.94it/s]
                                                   

 28%|██▊       | 1595/5758 [09:06<23:35,  2.94it/s]
 28%|██▊       | 1596/5758 [09:07<24:22,  2.85it/s]
                                                   

 28%|██▊       | 1596/5758 [09:07<24:22,  2.85it/s]
 28%|██▊       | 1597/5758 [09:07<23:00,  3.01it/s]
                                                   

 28%|██▊       | 1597/5758 [09:07<23:00,  3.01it/s]
 28%|██▊       | 1598/5758 [09:07<22:26,  3.09it/s]
                                                   

 28%|██▊       | 1598/5758 [09:07<22:26,  3.09it/s]
 28%|██▊       | 1599/5758 [09:08<22:48,  3.04it/s]
                                                   

 28%|██▊       | 1599/5758 [09:08<22:48,  3.04it/s]
 28%|██▊       | 1600/5758 [09:08<23:51,  2.91it/s]
                                                   

 28%|██▊       | 1600/5758 [09:08<23:51,  2.91it/s]
 28%|██▊       | 1601/5758 [09:08<23:57,  2.89it/s]
                                                   

 28%|██▊       | 1601/5758 [09:08<23:57,  2.89it/s]
 28%|██▊       | 1602/5758 [09:09<23:20,  2.97it/s]
                                                   

 28%|██▊       | 1602/5758 [09:09<23:20,  2.97it/s]
 28%|██▊       | 1603/5758 [09:09<22:27,  3.08it/s]
                                                   

 28%|██▊       | 1603/5758 [09:09<22:27,  3.08it/s]
 28%|██▊       | 1604/5758 [09:09<21:58,  3.15it/s]
                                                   

 28%|██▊       | 1604/5758 [09:09<21:58,  3.15it/s]
 28%|██▊       | 1605/5758 [09:10<22:10,  3.12it/s]
                                                   

 28%|██▊       | 1605/5758 [09:10<22:10,  3.12it/s]
 28%|██▊       | 1606/5758 [09:10<23:59,  2.88it/s]
                                                   

 28%|██▊       | 1606/5758 [09:10<23:59,  2.88it/s]
 28%|██▊       | 1607/5758 [09:10<24:43,  2.80it/s]
                                                   

 28%|██▊       | 1607/5758 [09:10<24:43,  2.80it/s]
 28%|██▊       | 1608/5758 [09:11<25:18,  2.73it/s]
                                                   

 28%|██▊       | 1608/5758 [09:11<25:18,  2.73it/s]
 28%|██▊       | 1609/5758 [09:11<23:55,  2.89it/s]
                                                   

 28%|██▊       | 1609/5758 [09:11<23:55,  2.89it/s]
 28%|██▊       | 1610/5758 [09:11<22:56,  3.01it/s]
                                                   

 28%|██▊       | 1610/5758 [09:11<22:56,  3.01it/s]
 28%|██▊       | 1611/5758 [09:12<22:02,  3.14it/s]
                                                   
{'loss': '2.136', 'grad_norm': '10.06', 'learning_rate': '7.299e-06', 'num_tokens': '1.018e+06', 'mean_token_accuracy': '0.5685', 'epoch': '0.2702'}
+{'loss': '0.7171', 'grad_norm': '12.5', 'learning_rate': '7.298e-06', 'num_tokens': '1.019e+06', 'mean_token_accuracy': '0.8473', 'epoch': '0.2704'}
+{'loss': '1.116', 'grad_norm': '10.5', 'learning_rate': '7.296e-06', 'num_tokens': '1.02e+06', 'mean_token_accuracy': '0.7262', 'epoch': '0.2706'}
+{'loss': '1.021', 'grad_norm': '38.5', 'learning_rate': '7.294e-06', 'num_tokens': '1.02e+06', 'mean_token_accuracy': '0.7765', 'epoch': '0.2708'}
+{'loss': '1.189', 'grad_norm': '7.594', 'learning_rate': '7.292e-06', 'num_tokens': '1.021e+06', 'mean_token_accuracy': '0.7359', 'epoch': '0.2709'}
+{'loss': '1.17', 'grad_norm': '15.06', 'learning_rate': '7.291e-06', 'num_tokens': '1.022e+06', 'mean_token_accuracy': '0.6911', 'epoch': '0.2711'}
+{'loss': '1.521', 'grad_norm': '12.75', 'learning_rate': '7.289e-06', 'num_tokens': '1.023e+06', 'mean_token_accuracy': '0.5937', 'epoch': '0.2713'}
+{'loss': '1.187', 'grad_norm': '13.12', 'learning_rate': '7.287e-06', 'num_tokens': '1.023e+06', 'mean_token_accuracy': '0.6535', 'epoch': '0.2714'}
+{'loss': '1.005', 'grad_norm': '5.469', 'learning_rate': '7.286e-06', 'num_tokens': '1.027e+06', 'mean_token_accuracy': '0.7485', 'epoch': '0.2716'}
+{'loss': '1.126', 'grad_norm': '14.5', 'learning_rate': '7.284e-06', 'num_tokens': '1.028e+06', 'mean_token_accuracy': '0.6786', 'epoch': '0.2718'}
+{'loss': '0.6228', 'grad_norm': '6.406', 'learning_rate': '7.282e-06', 'num_tokens': '1.029e+06', 'mean_token_accuracy': '0.8127', 'epoch': '0.272'}
+{'loss': '1.908', 'grad_norm': '12.62', 'learning_rate': '7.28e-06', 'num_tokens': '1.03e+06', 'mean_token_accuracy': '0.5928', 'epoch': '0.2721'}
+{'loss': '1.248', 'grad_norm': '26.62', 'learning_rate': '7.279e-06', 'num_tokens': '1.03e+06', 'mean_token_accuracy': '0.6963', 'epoch': '0.2723'}
+{'loss': '0.9916', 'grad_norm': '29.88', 'learning_rate': '7.277e-06', 'num_tokens': '1.03e+06', 'mean_token_accuracy': '0.7785', 'epoch': '0.2725'}
+{'loss': '0.5871', 'grad_norm': '15.88', 'learning_rate': '7.275e-06', 'num_tokens': '1.031e+06', 'mean_token_accuracy': '0.8655', 'epoch': '0.2727'}
+{'loss': '2.461', 'grad_norm': '22.12', 'learning_rate': '7.273e-06', 'num_tokens': '1.031e+06', 'mean_token_accuracy': '0.5556', 'epoch': '0.2728'}
+{'loss': '1.026', 'grad_norm': '5.688', 'learning_rate': '7.272e-06', 'num_tokens': '1.034e+06', 'mean_token_accuracy': '0.7505', 'epoch': '0.273'}
+{'loss': '1.881', 'grad_norm': '47.75', 'learning_rate': '7.27e-06', 'num_tokens': '1.034e+06', 'mean_token_accuracy': '0.68', 'epoch': '0.2732'}
+{'loss': '1.876', 'grad_norm': '35.75', 'learning_rate': '7.268e-06', 'num_tokens': '1.034e+06', 'mean_token_accuracy': '0.642', 'epoch': '0.2734'}
+{'loss': '1.248', 'grad_norm': '13.5', 'learning_rate': '7.266e-06', 'num_tokens': '1.035e+06', 'mean_token_accuracy': '0.6574', 'epoch': '0.2735'}
+{'loss': '1.209', 'grad_norm': '12.56', 'learning_rate': '7.265e-06', 'num_tokens': '1.035e+06', 'mean_token_accuracy': '0.694', 'epoch': '0.2737'}
+{'loss': '1.889', 'grad_norm': '17.25', 'learning_rate': '7.263e-06', 'num_tokens': '1.036e+06', 'mean_token_accuracy': '0.5952', 'epoch': '0.2739'}
+{'loss': '0.8959', 'grad_norm': '15.75', 'learning_rate': '7.261e-06', 'num_tokens': '1.036e+06', 'mean_token_accuracy': '0.8014', 'epoch': '0.2741'}
+{'loss': '1.424', 'grad_norm': '16.88', 'learning_rate': '7.259e-06', 'num_tokens': '1.036e+06', 'mean_token_accuracy': '0.6851', 'epoch': '0.2742'}
+{'loss': '1.015', 'grad_norm': '18.75', 'learning_rate': '7.258e-06', 'num_tokens': '1.037e+06', 'mean_token_accuracy': '0.7408', 'epoch': '0.2744'}
+{'loss': '0.506', 'grad_norm': '8.438', 'learning_rate': '7.256e-06', 'num_tokens': '1.038e+06', 'mean_token_accuracy': '0.8419', 'epoch': '0.2746'}
+{'loss': '1.122', 'grad_norm': '13.94', 'learning_rate': '7.254e-06', 'num_tokens': '1.039e+06', 'mean_token_accuracy': '0.7036', 'epoch': '0.2747'}
+{'loss': '1.053', 'grad_norm': '11.19', 'learning_rate': '7.253e-06', 'num_tokens': '1.039e+06', 'mean_token_accuracy': '0.6916', 'epoch': '0.2749'}
+{'loss': '1.295', 'grad_norm': '15.5', 'learning_rate': '7.251e-06', 'num_tokens': '1.04e+06', 'mean_token_accuracy': '0.6646', 'epoch': '0.2751'}
+{'loss': '1.227', 'grad_norm': '15', 'learning_rate': '7.249e-06', 'num_tokens': '1.041e+06', 'mean_token_accuracy': '0.6968', 'epoch': '0.2753'}
+{'loss': '1.673', 'grad_norm': '10.81', 'learning_rate': '7.247e-06', 'num_tokens': '1.042e+06', 'mean_token_accuracy': '0.5919', 'epoch': '0.2754'}
+{'loss': '0.5289', 'grad_norm': '6.5', 'learning_rate': '7.246e-06', 'num_tokens': '1.043e+06', 'mean_token_accuracy': '0.8299', 'epoch': '0.2756'}
+{'loss': '1.074', 'grad_norm': '10.88', 'learning_rate': '7.244e-06', 'num_tokens': '1.044e+06', 'mean_token_accuracy': '0.7106', 'epoch': '0.2758'}
+{'loss': '1.064', 'grad_norm': '11.19', 'learning_rate': '7.242e-06', 'num_tokens': '1.045e+06', 'mean_token_accuracy': '0.7666', 'epoch': '0.276'}
+{'loss': '0.4157', 'grad_norm': '5.844', 'learning_rate': '7.24e-06', 'num_tokens': '1.047e+06', 'mean_token_accuracy': '0.874', 'epoch': '0.2761'}
+{'loss': '1.169', 'grad_norm': '11.38', 'learning_rate': '7.239e-06', 'num_tokens': '1.047e+06', 'mean_token_accuracy': '0.7396', 'epoch': '0.2763'}
+{'loss': '0.6469', 'grad_norm': '9.625', 'learning_rate': '7.237e-06', 'num_tokens': '1.049e+06', 'mean_token_accuracy': '0.8021', 'epoch': '0.2765'}
+{'loss': '1.443', 'grad_norm': '17.75', 'learning_rate': '7.235e-06', 'num_tokens': '1.049e+06', 'mean_token_accuracy': '0.6947', 'epoch': '0.2767'}
+{'loss': '0.9628', 'grad_norm': '15.31', 'learning_rate': '7.233e-06', 'num_tokens': '1.049e+06', 'mean_token_accuracy': '0.784', 'epoch': '0.2768'}
+{'loss': '1.323', 'grad_norm': '9.562', 'learning_rate': '7.232e-06', 'num_tokens': '1.05e+06', 'mean_token_accuracy': '0.7088', 'epoch': '0.277'}
+{'loss': '0.7871', 'grad_norm': '30.62', 'learning_rate': '7.23e-06', 'num_tokens': '1.05e+06', 'mean_token_accuracy': '0.816', 'epoch': '0.2772'}
+{'loss': '1.97', 'grad_norm': '18.62', 'learning_rate': '7.228e-06', 'num_tokens': '1.051e+06', 'mean_token_accuracy': '0.6087', 'epoch': '0.2774'}
+{'loss': '2.337', 'grad_norm': '58.5', 'learning_rate': '7.226e-06', 'num_tokens': '1.051e+06', 'mean_token_accuracy': '0.5714', 'epoch': '0.2775'}
+{'loss': '1.208', 'grad_norm': '18.88', 'learning_rate': '7.225e-06', 'num_tokens': '1.051e+06', 'mean_token_accuracy': '0.7401', 'epoch': '0.2777'}
+{'loss': '1.103', 'grad_norm': '21.88', 'learning_rate': '7.223e-06', 'num_tokens': '1.051e+06', 'mean_token_accuracy': '0.7007', 'epoch': '0.2779'}
+{'loss': '1.29', 'grad_norm': '11.31', 'learning_rate': '7.221e-06', 'num_tokens': '1.052e+06', 'mean_token_accuracy': '0.6622', 'epoch': '0.278'}
+{'loss': '1.395', 'grad_norm': '13.44', 'learning_rate': '7.22e-06', 'num_tokens': '1.053e+06', 'mean_token_accuracy': '0.6129', 'epoch': '0.2782'}
+{'loss': '1.425', 'grad_norm': '34.5', 'learning_rate': '7.218e-06', 'num_tokens': '1.053e+06', 'mean_token_accuracy': '0.6943', 'epoch': '0.2784'}
+{'loss': '1.203', 'grad_norm': '21.12', 'learning_rate': '7.216e-06', 'num_tokens': '1.053e+06', 'mean_token_accuracy': '0.756', 'epoch': '0.2786'}
+{'loss': '0.7956', 'grad_norm': '8.312', 'learning_rate': '7.214e-06', 'num_tokens': '1.055e+06', 'mean_token_accuracy': '0.7655', 'epoch': '0.2787'}
+{'loss': '1.715', 'grad_norm': '13.75', 'learning_rate': '7.213e-06', 'num_tokens': '1.055e+06', 'mean_token_accuracy': '0.6597', 'epoch': '0.2789'}
+{'loss': '1.325', 'grad_norm': '6.531', 'learning_rate': '7.211e-06', 'num_tokens': '1.06e+06', 'mean_token_accuracy': '0.694', 'epoch': '0.2791'}
+{'loss': '1.42', 'grad_norm': '14.19', 'learning_rate': '7.209e-06', 'num_tokens': '1.06e+06', 'mean_token_accuracy': '0.7008', 'epoch': '0.2793'}
+{'loss': '1.083', 'grad_norm': '24.62', 'learning_rate': '7.207e-06', 'num_tokens': '1.06e+06', 'mean_token_accuracy': '0.7485', 'epoch': '0.2794'}
+{'loss': '1.314', 'grad_norm': '17.38', 'learning_rate': '7.206e-06', 'num_tokens': '1.061e+06', 'mean_token_accuracy': '0.7382', 'epoch': '0.2796'}
+
 28%|██▊       | 1611/5758 [09:12<22:02,  3.14it/s]
 28%|██▊       | 1612/5758 [09:12<21:45,  3.18it/s]
                                                   

 28%|██▊       | 1612/5758 [09:12<21:45,  3.18it/s]
 28%|██▊       | 1613/5758 [09:12<22:11,  3.11it/s]
                                                   

 28%|██▊       | 1613/5758 [09:12<22:11,  3.11it/s]
 28%|██▊       | 1614/5758 [09:13<21:52,  3.16it/s]
                                                   

 28%|██▊       | 1614/5758 [09:13<21:52,  3.16it/s]
 28%|██▊       | 1615/5758 [09:13<22:51,  3.02it/s]
                                                   

 28%|██▊       | 1615/5758 [09:13<22:51,  3.02it/s]
 28%|██▊       | 1616/5758 [09:13<22:18,  3.09it/s]
                                                   

 28%|██▊       | 1616/5758 [09:13<22:18,  3.09it/s]
 28%|██▊       | 1617/5758 [09:14<23:34,  2.93it/s]
                                                   

 28%|██▊       | 1617/5758 [09:14<23:34,  2.93it/s]
 28%|██▊       | 1618/5758 [09:14<24:00,  2.87it/s]
                                                   

 28%|██▊       | 1618/5758 [09:14<24:00,  2.87it/s]
 28%|██▊       | 1619/5758 [09:14<26:12,  2.63it/s]
                                                   

 28%|██▊       | 1619/5758 [09:14<26:12,  2.63it/s]
 28%|██▊       | 1620/5758 [09:15<25:21,  2.72it/s]
                                                   

 28%|██▊       | 1620/5758 [09:15<25:21,  2.72it/s]
 28%|██▊       | 1621/5758 [09:15<26:34,  2.60it/s]
                                                   

 28%|██▊       | 1621/5758 [09:15<26:34,  2.60it/s]
 28%|██▊       | 1622/5758 [09:16<26:03,  2.65it/s]
                                                   

 28%|██▊       | 1622/5758 [09:16<26:03,  2.65it/s]
 28%|██▊       | 1623/5758 [09:16<24:38,  2.80it/s]
                                                   

 28%|██▊       | 1623/5758 [09:16<24:38,  2.80it/s]
 28%|██▊       | 1624/5758 [09:16<24:09,  2.85it/s]
                                                   

 28%|██▊       | 1624/5758 [09:16<24:09,  2.85it/s]
 28%|██▊       | 1625/5758 [09:17<23:00,  2.99it/s]
                                                   

 28%|██▊       | 1625/5758 [09:17<23:00,  2.99it/s]
 28%|██▊       | 1626/5758 [09:17<22:23,  3.07it/s]
                                                   

 28%|██▊       | 1626/5758 [09:17<22:23,  3.07it/s]
 28%|██▊       | 1627/5758 [09:17<23:11,  2.97it/s]
                                                   

 28%|██▊       | 1627/5758 [09:17<23:11,  2.97it/s]
 28%|██▊       | 1628/5758 [09:18<23:38,  2.91it/s]
                                                   

 28%|██▊       | 1628/5758 [09:18<23:38,  2.91it/s]
 28%|██▊       | 1629/5758 [09:18<22:30,  3.06it/s]
                                                   

 28%|██▊       | 1629/5758 [09:18<22:30,  3.06it/s]
 28%|██▊       | 1630/5758 [09:18<22:19,  3.08it/s]
                                                   

 28%|██▊       | 1630/5758 [09:18<22:19,  3.08it/s]
 28%|██▊       | 1631/5758 [09:18<21:48,  3.15it/s]
                                                   

 28%|██▊       | 1631/5758 [09:18<21:48,  3.15it/s]
 28%|██▊       | 1632/5758 [09:19<23:10,  2.97it/s]
                                                   

 28%|██▊       | 1632/5758 [09:19<23:10,  2.97it/s]
 28%|██▊       | 1633/5758 [09:19<22:52,  3.01it/s]
                                                   

 28%|██▊       | 1633/5758 [09:19<22:52,  3.01it/s]
 28%|██▊       | 1634/5758 [09:20<22:38,  3.04it/s]
                                                   

 28%|██▊       | 1634/5758 [09:20<22:38,  3.04it/s]
 28%|██▊       | 1635/5758 [09:20<22:06,  3.11it/s]
                                                   

 28%|██▊       | 1635/5758 [09:20<22:06,  3.11it/s]
 28%|██▊       | 1636/5758 [09:20<23:14,  2.96it/s]
                                                   

 28%|██▊       | 1636/5758 [09:20<23:14,  2.96it/s]
 28%|██▊       | 1637/5758 [09:21<22:44,  3.02it/s]
                                                   

 28%|██▊       | 1637/5758 [09:21<22:44,  3.02it/s]
 28%|██▊       | 1638/5758 [09:21<23:24,  2.93it/s]
                                                   

 28%|██▊       | 1638/5758 [09:21<23:24,  2.93it/s]
 28%|██▊       | 1639/5758 [09:21<24:46,  2.77it/s]
                                                   

 28%|██▊       | 1639/5758 [09:21<24:46,  2.77it/s]
 28%|██▊       | 1640/5758 [09:22<23:59,  2.86it/s]
                                                   

 28%|██▊       | 1640/5758 [09:22<23:59,  2.86it/s]
 28%|██▊       | 1641/5758 [09:22<23:04,  2.97it/s]
                                                   

 28%|██▊       | 1641/5758 [09:22<23:04,  2.97it/s]
 29%|██▊       | 1642/5758 [09:22<22:18,  3.07it/s]
                                                   

 29%|██▊       | 1642/5758 [09:22<22:18,  3.07it/s]
 29%|██▊       | 1643/5758 [09:23<22:52,  3.00it/s]
                                                   

 29%|██▊       | 1643/5758 [09:23<22:52,  3.00it/s]
 29%|██▊       | 1644/5758 [09:23<23:30,  2.92it/s]
                                                   

 29%|██▊       | 1644/5758 [09:23<23:30,  2.92it/s]
 29%|██▊       | 1645/5758 [09:23<26:24,  2.60it/s]
                                                   

 29%|██▊       | 1645/5758 [09:23<26:24,  2.60it/s]
 29%|██▊       | 1646/5758 [09:24<24:37,  2.78it/s]
                                                   

 29%|██▊       | 1646/5758 [09:24<24:37,  2.78it/s]
 29%|██▊       | 1647/5758 [09:24<24:54,  2.75it/s]
                                                   

 29%|██▊       | 1647/5758 [09:24<24:54,  2.75it/s]
 29%|██▊       | 1648/5758 [09:24<24:02,  2.85it/s]
                                                   

 29%|██▊       | 1648/5758 [09:24<24:02,  2.85it/s]
 29%|██▊       | 1649/5758 [09:25<23:57,  2.86it/s]
                                                   

 29%|██▊       | 1649/5758 [09:25<23:57,  2.86it/s]
 29%|██▊       | 1650/5758 [09:25<23:22,  2.93it/s]
                                                   

 29%|██▊       | 1650/5758 [09:25<23:22,  2.93it/s]
 29%|██▊       | 1651/5758 [09:25<23:44,  2.88it/s]
                                                   

 29%|██▊       | 1651/5758 [09:25<23:44,  2.88it/s]
 29%|██▊       | 1652/5758 [09:26<22:51,  2.99it/s]
                                                   

 29%|██▊       | 1652/5758 [09:26<22:51,  2.99it/s]
 29%|██▊       | 1653/5758 [09:26<23:35,  2.90it/s]
                                                   

 29%|██▊       | 1653/5758 [09:26<23:35,  2.90it/s]
 29%|██▊       | 1654/5758 [09:26<23:41,  2.89it/s]
                                                   

 29%|██▊       | 1654/5758 [09:26<23:41,  2.89it/s]
 29%|██▊       | 1655/5758 [09:27<22:53,  2.99it/s]
                                                   

 29%|██▊       | 1655/5758 [09:27<22:53,  2.99it/s]
 29%|██▉       | 1656/5758 [09:27<22:10,  3.08it/s]
                                                   

 29%|██▉       | 1656/5758 [09:27<22:10,  3.08it/s]
 29%|██▉       | 1657/5758 [09:27<21:35,  3.17it/s]
                                                   

 29%|██▉       | 1657/5758 [09:27<21:35,  3.17it/s]
 29%|██▉       | 1658/5758 [09:28<22:17,  3.07it/s]
                                                   

 29%|██▉       | 1658/5758 [09:28<22:17,  3.07it/s]
 29%|██▉       | 1659/5758 [09:28<21:39,  3.15it/s]
                                                   

 29%|██▉       | 1659/5758 [09:28<21:39,  3.15it/s]
 29%|██▉       | 1660/5758 [09:28<21:52,  3.12it/s]
                                                   

 29%|██▉       | 1660/5758 [09:28<21:52,  3.12it/s]
 29%|██▉       | 1661/5758 [09:29<22:00,  3.10it/s]
                                                   

 29%|██▉       | 1661/5758 [09:29<22:00,  3.10it/s]
 29%|██▉       | 1662/5758 [09:29<22:19,  3.06it/s]
                                                   

 29%|██▉       | 1662/5758 [09:29<22:19,  3.06it/s]
 29%|██▉       | 1663/5758 [09:29<21:44,  3.14it/s]
                                                   

 29%|██▉       | 1663/5758 [09:29<21:44,  3.14it/s]
 29%|██▉       | 1664/5758 [09:30<22:08,  3.08it/s]
                                                   

 29%|██▉       | 1664/5758 [09:30<22:08,  3.08it/s]
 29%|██▉       | 1665/5758 [09:30<21:26,  3.18it/s]
                                                   

 29%|██▉       | 1665/5758 [09:30<21:26,  3.18it/s]
 29%|██▉       | 1666/5758 [09:30<21:22,  3.19it/s]
                                                   
{'loss': '1.199', 'grad_norm': '11.94', 'learning_rate': '7.204e-06', 'num_tokens': '1.061e+06', 'mean_token_accuracy': '0.6962', 'epoch': '0.2798'}
+{'loss': '1.153', 'grad_norm': '11.81', 'learning_rate': '7.202e-06', 'num_tokens': '1.062e+06', 'mean_token_accuracy': '0.7037', 'epoch': '0.28'}
+{'loss': '1.609', 'grad_norm': '14.12', 'learning_rate': '7.2e-06', 'num_tokens': '1.062e+06', 'mean_token_accuracy': '0.6599', 'epoch': '0.2801'}
+{'loss': '1.322', 'grad_norm': '12.75', 'learning_rate': '7.199e-06', 'num_tokens': '1.063e+06', 'mean_token_accuracy': '0.6368', 'epoch': '0.2803'}
+{'loss': '1.035', 'grad_norm': '14.5', 'learning_rate': '7.197e-06', 'num_tokens': '1.064e+06', 'mean_token_accuracy': '0.7031', 'epoch': '0.2805'}
+{'loss': '2.024', 'grad_norm': '12.25', 'learning_rate': '7.195e-06', 'num_tokens': '1.064e+06', 'mean_token_accuracy': '0.5849', 'epoch': '0.2807'}
+{'loss': '1.087', 'grad_norm': '13.44', 'learning_rate': '7.193e-06', 'num_tokens': '1.065e+06', 'mean_token_accuracy': '0.6936', 'epoch': '0.2808'}
+{'loss': '1.103', 'grad_norm': '14', 'learning_rate': '7.192e-06', 'num_tokens': '1.065e+06', 'mean_token_accuracy': '0.7029', 'epoch': '0.281'}
+{'loss': '0.6013', 'grad_norm': '12.06', 'learning_rate': '7.19e-06', 'num_tokens': '1.066e+06', 'mean_token_accuracy': '0.8399', 'epoch': '0.2812'}
+{'loss': '1.91', 'grad_norm': '15.44', 'learning_rate': '7.188e-06', 'num_tokens': '1.066e+06', 'mean_token_accuracy': '0.5897', 'epoch': '0.2813'}
+{'loss': '0.5671', 'grad_norm': '6.094', 'learning_rate': '7.187e-06', 'num_tokens': '1.068e+06', 'mean_token_accuracy': '0.8294', 'epoch': '0.2815'}
+{'loss': '1.451', 'grad_norm': '7.844', 'learning_rate': '7.185e-06', 'num_tokens': '1.07e+06', 'mean_token_accuracy': '0.66', 'epoch': '0.2817'}
+{'loss': '1.35', 'grad_norm': '35', 'learning_rate': '7.183e-06', 'num_tokens': '1.07e+06', 'mean_token_accuracy': '0.7342', 'epoch': '0.2819'}
+{'loss': '0.7487', 'grad_norm': '14.31', 'learning_rate': '7.181e-06', 'num_tokens': '1.071e+06', 'mean_token_accuracy': '0.7967', 'epoch': '0.282'}
+{'loss': '2.477', 'grad_norm': '63.5', 'learning_rate': '7.18e-06', 'num_tokens': '1.071e+06', 'mean_token_accuracy': '0.6232', 'epoch': '0.2822'}
+{'loss': '1.118', 'grad_norm': '12.81', 'learning_rate': '7.178e-06', 'num_tokens': '1.071e+06', 'mean_token_accuracy': '0.6711', 'epoch': '0.2824'}
+{'loss': '1.358', 'grad_norm': '12.5', 'learning_rate': '7.176e-06', 'num_tokens': '1.072e+06', 'mean_token_accuracy': '0.6429', 'epoch': '0.2826'}
+{'loss': '1.19', 'grad_norm': '15.31', 'learning_rate': '7.174e-06', 'num_tokens': '1.073e+06', 'mean_token_accuracy': '0.664', 'epoch': '0.2827'}
+{'loss': '0.6417', 'grad_norm': '20.62', 'learning_rate': '7.173e-06', 'num_tokens': '1.073e+06', 'mean_token_accuracy': '0.8535', 'epoch': '0.2829'}
+{'loss': '1.233', 'grad_norm': '40.5', 'learning_rate': '7.171e-06', 'num_tokens': '1.073e+06', 'mean_token_accuracy': '0.7241', 'epoch': '0.2831'}
+{'loss': '0.5036', 'grad_norm': '17', 'learning_rate': '7.169e-06', 'num_tokens': '1.073e+06', 'mean_token_accuracy': '0.8848', 'epoch': '0.2833'}
+{'loss': '2.345', 'grad_norm': '19.38', 'learning_rate': '7.167e-06', 'num_tokens': '1.074e+06', 'mean_token_accuracy': '0.5365', 'epoch': '0.2834'}
+{'loss': '1.045', 'grad_norm': '22.75', 'learning_rate': '7.166e-06', 'num_tokens': '1.074e+06', 'mean_token_accuracy': '0.806', 'epoch': '0.2836'}
+{'loss': '1.345', 'grad_norm': '13.56', 'learning_rate': '7.164e-06', 'num_tokens': '1.075e+06', 'mean_token_accuracy': '0.6403', 'epoch': '0.2838'}
+{'loss': '1.068', 'grad_norm': '12.25', 'learning_rate': '7.162e-06', 'num_tokens': '1.076e+06', 'mean_token_accuracy': '0.7218', 'epoch': '0.284'}
+{'loss': '1.892', 'grad_norm': '17.12', 'learning_rate': '7.16e-06', 'num_tokens': '1.076e+06', 'mean_token_accuracy': '0.6084', 'epoch': '0.2841'}
+{'loss': '2.282', 'grad_norm': '12.44', 'learning_rate': '7.159e-06', 'num_tokens': '1.077e+06', 'mean_token_accuracy': '0.5694', 'epoch': '0.2843'}
+{'loss': '2.351', 'grad_norm': '11.19', 'learning_rate': '7.157e-06', 'num_tokens': '1.078e+06', 'mean_token_accuracy': '0.542', 'epoch': '0.2845'}
+{'loss': '1.085', 'grad_norm': '25.12', 'learning_rate': '7.155e-06', 'num_tokens': '1.078e+06', 'mean_token_accuracy': '0.7619', 'epoch': '0.2846'}
+{'loss': '1.822', 'grad_norm': '16.25', 'learning_rate': '7.154e-06', 'num_tokens': '1.078e+06', 'mean_token_accuracy': '0.6514', 'epoch': '0.2848'}
+{'loss': '0.9697', 'grad_norm': '11.06', 'learning_rate': '7.152e-06', 'num_tokens': '1.079e+06', 'mean_token_accuracy': '0.7275', 'epoch': '0.285'}
+{'loss': '1.125', 'grad_norm': '7.969', 'learning_rate': '7.15e-06', 'num_tokens': '1.08e+06', 'mean_token_accuracy': '0.7291', 'epoch': '0.2852'}
+{'loss': '1.543', 'grad_norm': '24.5', 'learning_rate': '7.148e-06', 'num_tokens': '1.081e+06', 'mean_token_accuracy': '0.639', 'epoch': '0.2853'}
+{'loss': '1.149', 'grad_norm': '14.44', 'learning_rate': '7.147e-06', 'num_tokens': '1.081e+06', 'mean_token_accuracy': '0.6976', 'epoch': '0.2855'}
+{'loss': '1.698', 'grad_norm': '23.12', 'learning_rate': '7.145e-06', 'num_tokens': '1.081e+06', 'mean_token_accuracy': '0.6748', 'epoch': '0.2857'}
+{'loss': '1.238', 'grad_norm': '20.62', 'learning_rate': '7.143e-06', 'num_tokens': '1.081e+06', 'mean_token_accuracy': '0.7391', 'epoch': '0.2859'}
+{'loss': '0.8095', 'grad_norm': '11.81', 'learning_rate': '7.141e-06', 'num_tokens': '1.082e+06', 'mean_token_accuracy': '0.7967', 'epoch': '0.286'}
+{'loss': '2.953', 'grad_norm': '44.75', 'learning_rate': '7.14e-06', 'num_tokens': '1.082e+06', 'mean_token_accuracy': '0.4783', 'epoch': '0.2862'}
+{'loss': '1.565', 'grad_norm': '12.31', 'learning_rate': '7.138e-06', 'num_tokens': '1.083e+06', 'mean_token_accuracy': '0.6061', 'epoch': '0.2864'}
+{'loss': '0.8256', 'grad_norm': '13.81', 'learning_rate': '7.136e-06', 'num_tokens': '1.084e+06', 'mean_token_accuracy': '0.8261', 'epoch': '0.2866'}
+{'loss': '2.432', 'grad_norm': '39.25', 'learning_rate': '7.134e-06', 'num_tokens': '1.084e+06', 'mean_token_accuracy': '0.5505', 'epoch': '0.2867'}
+{'loss': '0.7669', 'grad_norm': '26.25', 'learning_rate': '7.133e-06', 'num_tokens': '1.084e+06', 'mean_token_accuracy': '0.8323', 'epoch': '0.2869'}
+{'loss': '0.5894', 'grad_norm': '6.188', 'learning_rate': '7.131e-06', 'num_tokens': '1.086e+06', 'mean_token_accuracy': '0.8225', 'epoch': '0.2871'}
+{'loss': '1.549', 'grad_norm': '17.25', 'learning_rate': '7.129e-06', 'num_tokens': '1.086e+06', 'mean_token_accuracy': '0.6649', 'epoch': '0.2873'}
+{'loss': '2.028', 'grad_norm': '17.38', 'learning_rate': '7.127e-06', 'num_tokens': '1.087e+06', 'mean_token_accuracy': '0.6106', 'epoch': '0.2874'}
+{'loss': '0.6854', 'grad_norm': '16.88', 'learning_rate': '7.126e-06', 'num_tokens': '1.087e+06', 'mean_token_accuracy': '0.8567', 'epoch': '0.2876'}
+{'loss': '0.5751', 'grad_norm': '7.125', 'learning_rate': '7.124e-06', 'num_tokens': '1.088e+06', 'mean_token_accuracy': '0.8317', 'epoch': '0.2878'}
+{'loss': '1.103', 'grad_norm': '18.5', 'learning_rate': '7.122e-06', 'num_tokens': '1.089e+06', 'mean_token_accuracy': '0.768', 'epoch': '0.2879'}
+{'loss': '1.934', 'grad_norm': '13.12', 'learning_rate': '7.121e-06', 'num_tokens': '1.089e+06', 'mean_token_accuracy': '0.5912', 'epoch': '0.2881'}
+{'loss': '1.981', 'grad_norm': '38.25', 'learning_rate': '7.119e-06', 'num_tokens': '1.089e+06', 'mean_token_accuracy': '0.6495', 'epoch': '0.2883'}
+{'loss': '1.571', 'grad_norm': '12.06', 'learning_rate': '7.117e-06', 'num_tokens': '1.09e+06', 'mean_token_accuracy': '0.6057', 'epoch': '0.2885'}
+{'loss': '1.594', 'grad_norm': '11.31', 'learning_rate': '7.115e-06', 'num_tokens': '1.091e+06', 'mean_token_accuracy': '0.6393', 'epoch': '0.2886'}
+{'loss': '1.817', 'grad_norm': '12.69', 'learning_rate': '7.114e-06', 'num_tokens': '1.092e+06', 'mean_token_accuracy': '0.6317', 'epoch': '0.2888'}
+{'loss': '1.005', 'grad_norm': '29.75', 'learning_rate': '7.112e-06', 'num_tokens': '1.092e+06', 'mean_token_accuracy': '0.8095', 'epoch': '0.289'}
+{'loss': '1.99', 'grad_norm': '27.75', 'learning_rate': '7.11e-06', 'num_tokens': '1.092e+06', 'mean_token_accuracy': '0.589', 'epoch': '0.2892'}
+
 29%|██▉       | 1666/5758 [09:30<21:22,  3.19it/s]
 29%|██▉       | 1667/5758 [09:31<21:22,  3.19it/s]
                                                   

 29%|██▉       | 1667/5758 [09:31<21:22,  3.19it/s]
 29%|██▉       | 1668/5758 [09:31<21:45,  3.13it/s]
                                                   

 29%|██▉       | 1668/5758 [09:31<21:45,  3.13it/s]
 29%|██▉       | 1669/5758 [09:31<22:00,  3.10it/s]
                                                   

 29%|██▉       | 1669/5758 [09:31<22:00,  3.10it/s]
 29%|██▉       | 1670/5758 [09:32<21:23,  3.19it/s]
                                                   

 29%|██▉       | 1670/5758 [09:32<21:23,  3.19it/s]
 29%|██▉       | 1671/5758 [09:32<21:33,  3.16it/s]
                                                   

 29%|██▉       | 1671/5758 [09:32<21:33,  3.16it/s]
 29%|██▉       | 1672/5758 [09:32<21:29,  3.17it/s]
                                                   

 29%|██▉       | 1672/5758 [09:32<21:29,  3.17it/s]
 29%|██▉       | 1673/5758 [09:33<22:29,  3.03it/s]
                                                   

 29%|██▉       | 1673/5758 [09:33<22:29,  3.03it/s]
 29%|██▉       | 1674/5758 [09:33<24:45,  2.75it/s]
                                                   

 29%|██▉       | 1674/5758 [09:33<24:45,  2.75it/s]
 29%|██▉       | 1675/5758 [09:33<24:54,  2.73it/s]
                                                   

 29%|██▉       | 1675/5758 [09:33<24:54,  2.73it/s]
 29%|██▉       | 1676/5758 [09:34<25:01,  2.72it/s]
                                                   

 29%|██▉       | 1676/5758 [09:34<25:01,  2.72it/s]
 29%|██▉       | 1677/5758 [09:34<23:22,  2.91it/s]
                                                   

 29%|██▉       | 1677/5758 [09:34<23:22,  2.91it/s]
 29%|██▉       | 1678/5758 [09:34<22:24,  3.04it/s]
                                                   

 29%|██▉       | 1678/5758 [09:34<22:24,  3.04it/s]
 29%|██▉       | 1679/5758 [09:35<21:38,  3.14it/s]
                                                   

 29%|██▉       | 1679/5758 [09:35<21:38,  3.14it/s]
 29%|██▉       | 1680/5758 [09:35<20:59,  3.24it/s]
                                                   

 29%|██▉       | 1680/5758 [09:35<20:59,  3.24it/s]
 29%|██▉       | 1681/5758 [09:35<21:40,  3.13it/s]
                                                   

 29%|██▉       | 1681/5758 [09:35<21:40,  3.13it/s]
 29%|██▉       | 1682/5758 [09:35<21:17,  3.19it/s]
                                                   

 29%|██▉       | 1682/5758 [09:35<21:17,  3.19it/s]
 29%|██▉       | 1683/5758 [09:36<22:49,  2.97it/s]
                                                   

 29%|██▉       | 1683/5758 [09:36<22:49,  2.97it/s]
 29%|██▉       | 1684/5758 [09:36<22:24,  3.03it/s]
                                                   

 29%|██▉       | 1684/5758 [09:36<22:24,  3.03it/s]
 29%|██▉       | 1685/5758 [09:37<21:46,  3.12it/s]
                                                   

 29%|██▉       | 1685/5758 [09:37<21:46,  3.12it/s]
 29%|██▉       | 1686/5758 [09:37<22:29,  3.02it/s]
                                                   

 29%|██▉       | 1686/5758 [09:37<22:29,  3.02it/s]
 29%|██▉       | 1687/5758 [09:37<23:20,  2.91it/s]
                                                   

 29%|██▉       | 1687/5758 [09:37<23:20,  2.91it/s]
 29%|██▉       | 1688/5758 [09:38<22:52,  2.97it/s]
                                                   

 29%|██▉       | 1688/5758 [09:38<22:52,  2.97it/s]
 29%|██▉       | 1689/5758 [09:38<23:25,  2.90it/s]
                                                   

 29%|██▉       | 1689/5758 [09:38<23:25,  2.90it/s]
 29%|██▉       | 1690/5758 [09:38<24:00,  2.82it/s]
                                                   

 29%|██▉       | 1690/5758 [09:38<24:00,  2.82it/s]
 29%|██▉       | 1691/5758 [09:39<23:36,  2.87it/s]
                                                   

 29%|██▉       | 1691/5758 [09:39<23:36,  2.87it/s]
 29%|██▉       | 1692/5758 [09:39<23:07,  2.93it/s]
                                                   

 29%|██▉       | 1692/5758 [09:39<23:07,  2.93it/s]
 29%|██▉       | 1693/5758 [09:39<22:15,  3.04it/s]
                                                   

 29%|██▉       | 1693/5758 [09:39<22:15,  3.04it/s]
 29%|██▉       | 1694/5758 [09:40<23:25,  2.89it/s]
                                                   

 29%|██▉       | 1694/5758 [09:40<23:25,  2.89it/s]
 29%|██▉       | 1695/5758 [09:40<23:28,  2.89it/s]
                                                   

 29%|█���▉       | 1695/5758 [09:40<23:28,  2.89it/s]
 29%|██▉       | 1696/5758 [09:40<22:43,  2.98it/s]
                                                   

 29%|██▉       | 1696/5758 [09:40<22:43,  2.98it/s]
 29%|██▉       | 1697/5758 [09:41<22:19,  3.03it/s]
                                                   

 29%|██▉       | 1697/5758 [09:41<22:19,  3.03it/s]
 29%|██▉       | 1698/5758 [09:41<22:52,  2.96it/s]
                                                   

 29%|██▉       | 1698/5758 [09:41<22:52,  2.96it/s]
 30%|██▉       | 1699/5758 [09:41<24:16,  2.79it/s]
                                                   

 30%|██▉       | 1699/5758 [09:41<24:16,  2.79it/s]
 30%|██▉       | 1700/5758 [09:42<22:54,  2.95it/s]
                                                   

 30%|██▉       | 1700/5758 [09:42<22:54,  2.95it/s]
 30%|██▉       | 1701/5758 [09:42<22:33,  3.00it/s]
                                                   

 30%|██▉       | 1701/5758 [09:42<22:33,  3.00it/s]
 30%|██▉       | 1702/5758 [09:42<21:58,  3.08it/s]
                                                   

 30%|██▉       | 1702/5758 [09:42<21:58,  3.08it/s]
 30%|██▉       | 1703/5758 [09:43<22:51,  2.96it/s]
                                                   

 30%|██▉       | 1703/5758 [09:43<22:51,  2.96it/s]
 30%|██▉       | 1704/5758 [09:43<22:13,  3.04it/s]
                                                   

 30%|██▉       | 1704/5758 [09:43<22:13,  3.04it/s]
 30%|██▉       | 1705/5758 [09:43<22:43,  2.97it/s]
                                                   

 30%|██▉       | 1705/5758 [09:43<22:43,  2.97it/s]
 30%|██▉       | 1706/5758 [09:44<23:18,  2.90it/s]
                                                   

 30%|██▉       | 1706/5758 [09:44<23:18,  2.90it/s]
 30%|██▉       | 1707/5758 [09:44<22:23,  3.02it/s]
                                                   

 30%|██▉       | 1707/5758 [09:44<22:23,  3.02it/s]
 30%|██▉       | 1708/5758 [09:44<21:45,  3.10it/s]
                                                   

 30%|██▉       | 1708/5758 [09:44<21:45,  3.10it/s]
 30%|██▉       | 1709/5758 [09:45<21:07,  3.19it/s]
                                                   

 30%|██▉       | 1709/5758 [09:45<21:07,  3.19it/s]
 30%|██▉       | 1710/5758 [09:45<22:15,  3.03it/s]
                                                   

 30%|██▉       | 1710/5758 [09:45<22:15,  3.03it/s]
 30%|██▉       | 1711/5758 [09:45<23:21,  2.89it/s]
                                                   

 30%|██▉       | 1711/5758 [09:45<23:21,  2.89it/s]
 30%|██▉       | 1712/5758 [09:46<22:38,  2.98it/s]
                                                   

 30%|██▉       | 1712/5758 [09:46<22:38,  2.98it/s]
 30%|██▉       | 1713/5758 [09:46<23:13,  2.90it/s]
                                                   

 30%|██▉       | 1713/5758 [09:46<23:13,  2.90it/s]
 30%|██▉       | 1714/5758 [09:46<22:35,  2.98it/s]
                                                   

 30%|██▉       | 1714/5758 [09:46<22:35,  2.98it/s]
 30%|██▉       | 1715/5758 [09:47<22:57,  2.93it/s]
                                                   

 30%|██▉       | 1715/5758 [09:47<22:57,  2.93it/s]
 30%|██▉       | 1716/5758 [09:47<23:21,  2.88it/s]
                                                   

 30%|██▉       | 1716/5758 [09:47<23:21,  2.88it/s]
 30%|██▉       | 1717/5758 [09:47<23:07,  2.91it/s]
                                                   

 30%|██▉       | 1717/5758 [09:47<23:07,  2.91it/s]
 30%|██▉       | 1718/5758 [09:48<22:28,  2.99it/s]
                                                   

 30%|██▉       | 1718/5758 [09:48<22:28,  2.99it/s]
 30%|██▉       | 1719/5758 [09:48<21:44,  3.10it/s]
                                                   

 30%|██▉       | 1719/5758 [09:48<21:44,  3.10it/s]
 30%|██▉       | 1720/5758 [09:48<21:24,  3.14it/s]
                                                   

 30%|██▉       | 1720/5758 [09:48<21:24,  3.14it/s]
 30%|██▉       | 1721/5758 [09:49<21:08,  3.18it/s]
                                                   
{'loss': '3.104', 'grad_norm': '12.5', 'learning_rate': '7.108e-06', 'num_tokens': '1.093e+06', 'mean_token_accuracy': '0.4761', 'epoch': '0.2893'}
+{'loss': '1.308', 'grad_norm': '22.12', 'learning_rate': '7.107e-06', 'num_tokens': '1.093e+06', 'mean_token_accuracy': '0.7245', 'epoch': '0.2895'}
+{'loss': '0.9529', 'grad_norm': '17.25', 'learning_rate': '7.105e-06', 'num_tokens': '1.094e+06', 'mean_token_accuracy': '0.7541', 'epoch': '0.2897'}
+{'loss': '0.8905', 'grad_norm': '12.75', 'learning_rate': '7.103e-06', 'num_tokens': '1.094e+06', 'mean_token_accuracy': '0.7932', 'epoch': '0.2899'}
+{'loss': '1.146', 'grad_norm': '13.62', 'learning_rate': '7.101e-06', 'num_tokens': '1.095e+06', 'mean_token_accuracy': '0.6697', 'epoch': '0.29'}
+{'loss': '1.362', 'grad_norm': '11', 'learning_rate': '7.1e-06', 'num_tokens': '1.095e+06', 'mean_token_accuracy': '0.6927', 'epoch': '0.2902'}
+{'loss': '1.524', 'grad_norm': '12.38', 'learning_rate': '7.098e-06', 'num_tokens': '1.096e+06', 'mean_token_accuracy': '0.6707', 'epoch': '0.2904'}
+{'loss': '0.6091', 'grad_norm': '24.75', 'learning_rate': '7.096e-06', 'num_tokens': '1.096e+06', 'mean_token_accuracy': '0.8655', 'epoch': '0.2906'}
+{'loss': '1.024', 'grad_norm': '11.44', 'learning_rate': '7.094e-06', 'num_tokens': '1.097e+06', 'mean_token_accuracy': '0.7124', 'epoch': '0.2907'}
+{'loss': '1.184', 'grad_norm': '9.688', 'learning_rate': '7.093e-06', 'num_tokens': '1.098e+06', 'mean_token_accuracy': '0.6794', 'epoch': '0.2909'}
+{'loss': '2.046', 'grad_norm': '12.75', 'learning_rate': '7.091e-06', 'num_tokens': '1.099e+06', 'mean_token_accuracy': '0.5924', 'epoch': '0.2911'}
+{'loss': '1.158', 'grad_norm': '13.25', 'learning_rate': '7.089e-06', 'num_tokens': '1.099e+06', 'mean_token_accuracy': '0.6699', 'epoch': '0.2912'}
+{'loss': '1.01', 'grad_norm': '24', 'learning_rate': '7.088e-06', 'num_tokens': '1.099e+06', 'mean_token_accuracy': '0.7588', 'epoch': '0.2914'}
+{'loss': '1.441', 'grad_norm': '13.75', 'learning_rate': '7.086e-06', 'num_tokens': '1.1e+06', 'mean_token_accuracy': '0.7011', 'epoch': '0.2916'}
+{'loss': '1.243', 'grad_norm': '13.31', 'learning_rate': '7.084e-06', 'num_tokens': '1.101e+06', 'mean_token_accuracy': '0.6644', 'epoch': '0.2918'}
+{'loss': '1.586', 'grad_norm': '9.938', 'learning_rate': '7.082e-06', 'num_tokens': '1.102e+06', 'mean_token_accuracy': '0.65', 'epoch': '0.2919'}
+{'loss': '1.322', 'grad_norm': '13.38', 'learning_rate': '7.081e-06', 'num_tokens': '1.102e+06', 'mean_token_accuracy': '0.6235', 'epoch': '0.2921'}
+{'loss': '1.152', 'grad_norm': '12.94', 'learning_rate': '7.079e-06', 'num_tokens': '1.103e+06', 'mean_token_accuracy': '0.683', 'epoch': '0.2923'}
+{'loss': '0.9861', 'grad_norm': '5.781', 'learning_rate': '7.077e-06', 'num_tokens': '1.106e+06', 'mean_token_accuracy': '0.7462', 'epoch': '0.2925'}
+{'loss': '0.9488', 'grad_norm': '12', 'learning_rate': '7.075e-06', 'num_tokens': '1.106e+06', 'mean_token_accuracy': '0.7145', 'epoch': '0.2926'}
+{'loss': '1.334', 'grad_norm': '10.38', 'learning_rate': '7.074e-06', 'num_tokens': '1.107e+06', 'mean_token_accuracy': '0.6236', 'epoch': '0.2928'}
+{'loss': '0.9673', 'grad_norm': '13.56', 'learning_rate': '7.072e-06', 'num_tokens': '1.108e+06', 'mean_token_accuracy': '0.7109', 'epoch': '0.293'}
+{'loss': '1.942', 'grad_norm': '18.25', 'learning_rate': '7.07e-06', 'num_tokens': '1.108e+06', 'mean_token_accuracy': '0.6203', 'epoch': '0.2932'}
+{'loss': '1.822', 'grad_norm': '20', 'learning_rate': '7.068e-06', 'num_tokens': '1.109e+06', 'mean_token_accuracy': '0.6462', 'epoch': '0.2933'}
+{'loss': '1.512', 'grad_norm': '22.62', 'learning_rate': '7.067e-06', 'num_tokens': '1.109e+06', 'mean_token_accuracy': '0.6449', 'epoch': '0.2935'}
+{'loss': '1.401', 'grad_norm': '25', 'learning_rate': '7.065e-06', 'num_tokens': '1.109e+06', 'mean_token_accuracy': '0.6887', 'epoch': '0.2937'}
+{'loss': '1.163', 'grad_norm': '11.88', 'learning_rate': '7.063e-06', 'num_tokens': '1.11e+06', 'mean_token_accuracy': '0.6753', 'epoch': '0.2939'}
+{'loss': '1.293', 'grad_norm': '12.69', 'learning_rate': '7.061e-06', 'num_tokens': '1.111e+06', 'mean_token_accuracy': '0.6761', 'epoch': '0.294'}
+{'loss': '2.313', 'grad_norm': '18.12', 'learning_rate': '7.06e-06', 'num_tokens': '1.111e+06', 'mean_token_accuracy': '0.5273', 'epoch': '0.2942'}
+{'loss': '2.283', 'grad_norm': '20.25', 'learning_rate': '7.058e-06', 'num_tokens': '1.112e+06', 'mean_token_accuracy': '0.556', 'epoch': '0.2944'}
+{'loss': '1.44', 'grad_norm': '8.625', 'learning_rate': '7.056e-06', 'num_tokens': '1.113e+06', 'mean_token_accuracy': '0.6714', 'epoch': '0.2945'}
+{'loss': '1.102', 'grad_norm': '11.06', 'learning_rate': '7.055e-06', 'num_tokens': '1.114e+06', 'mean_token_accuracy': '0.7087', 'epoch': '0.2947'}
+{'loss': '1.985', 'grad_norm': '10.44', 'learning_rate': '7.053e-06', 'num_tokens': '1.115e+06', 'mean_token_accuracy': '0.5763', 'epoch': '0.2949'}
+{'loss': '0.8192', 'grad_norm': '13.38', 'learning_rate': '7.051e-06', 'num_tokens': '1.115e+06', 'mean_token_accuracy': '0.7981', 'epoch': '0.2951'}
+{'loss': '1.148', 'grad_norm': '17.12', 'learning_rate': '7.049e-06', 'num_tokens': '1.116e+06', 'mean_token_accuracy': '0.7781', 'epoch': '0.2952'}
+{'loss': '1.218', 'grad_norm': '12.19', 'learning_rate': '7.048e-06', 'num_tokens': '1.116e+06', 'mean_token_accuracy': '0.6481', 'epoch': '0.2954'}
+{'loss': '0.7999', 'grad_norm': '12', 'learning_rate': '7.046e-06', 'num_tokens': '1.117e+06', 'mean_token_accuracy': '0.7593', 'epoch': '0.2956'}
+{'loss': '1.842', 'grad_norm': '9.438', 'learning_rate': '7.044e-06', 'num_tokens': '1.118e+06', 'mean_token_accuracy': '0.622', 'epoch': '0.2958'}
+{'loss': '0.8899', 'grad_norm': '17.62', 'learning_rate': '7.042e-06', 'num_tokens': '1.118e+06', 'mean_token_accuracy': '0.7977', 'epoch': '0.2959'}
+{'loss': '2.064', 'grad_norm': '30.88', 'learning_rate': '7.041e-06', 'num_tokens': '1.118e+06', 'mean_token_accuracy': '0.5413', 'epoch': '0.2961'}
+{'loss': '1.898', 'grad_norm': '9.062', 'learning_rate': '7.039e-06', 'num_tokens': '1.119e+06', 'mean_token_accuracy': '0.6049', 'epoch': '0.2963'}
+{'loss': '1', 'grad_norm': '29.88', 'learning_rate': '7.037e-06', 'num_tokens': '1.12e+06', 'mean_token_accuracy': '0.7853', 'epoch': '0.2965'}
+{'loss': '1.544', 'grad_norm': '28.75', 'learning_rate': '7.035e-06', 'num_tokens': '1.12e+06', 'mean_token_accuracy': '0.7531', 'epoch': '0.2966'}
+{'loss': '2.044', 'grad_norm': '17.75', 'learning_rate': '7.034e-06', 'num_tokens': '1.12e+06', 'mean_token_accuracy': '0.5893', 'epoch': '0.2968'}
+{'loss': '0.7801', 'grad_norm': '10.56', 'learning_rate': '7.032e-06', 'num_tokens': '1.121e+06', 'mean_token_accuracy': '0.8152', 'epoch': '0.297'}
+{'loss': '0.5141', 'grad_norm': '14.19', 'learning_rate': '7.03e-06', 'num_tokens': '1.121e+06', 'mean_token_accuracy': '0.8927', 'epoch': '0.2972'}
+{'loss': '0.9367', 'grad_norm': '24.88', 'learning_rate': '7.028e-06', 'num_tokens': '1.121e+06', 'mean_token_accuracy': '0.8197', 'epoch': '0.2973'}
+{'loss': '1.915', 'grad_norm': '16.12', 'learning_rate': '7.027e-06', 'num_tokens': '1.122e+06', 'mean_token_accuracy': '0.5867', 'epoch': '0.2975'}
+{'loss': '1.17', 'grad_norm': '33.5', 'learning_rate': '7.025e-06', 'num_tokens': '1.122e+06', 'mean_token_accuracy': '0.6951', 'epoch': '0.2977'}
+{'loss': '2.024', 'grad_norm': '14.19', 'learning_rate': '7.023e-06', 'num_tokens': '1.122e+06', 'mean_token_accuracy': '0.5732', 'epoch': '0.2978'}
+{'loss': '1.853', 'grad_norm': '15', 'learning_rate': '7.022e-06', 'num_tokens': '1.123e+06', 'mean_token_accuracy': '0.6', 'epoch': '0.298'}
+{'loss': '0.9876', 'grad_norm': '19', 'learning_rate': '7.02e-06', 'num_tokens': '1.123e+06', 'mean_token_accuracy': '0.7832', 'epoch': '0.2982'}
+{'loss': '1.175', 'grad_norm': '12.56', 'learning_rate': '7.018e-06', 'num_tokens': '1.124e+06', 'mean_token_accuracy': '0.6704', 'epoch': '0.2984'}
+{'loss': '0.5258', 'grad_norm': '9.312', 'learning_rate': '7.016e-06', 'num_tokens': '1.125e+06', 'mean_token_accuracy': '0.8411', 'epoch': '0.2985'}
+{'loss': '1.08', 'grad_norm': '11.62', 'learning_rate': '7.015e-06', 'num_tokens': '1.126e+06', 'mean_token_accuracy': '0.6977', 'epoch': '0.2987'}
+
 30%|██▉       | 1721/5758 [09:49<21:08,  3.18it/s]
 30%|██▉       | 1722/5758 [09:49<22:39,  2.97it/s]
                                                   

 30%|██▉       | 1722/5758 [09:49<22:39,  2.97it/s]
 30%|██▉       | 1723/5758 [09:49<21:56,  3.07it/s]
                                                   

 30%|██▉       | 1723/5758 [09:49<21:56,  3.07it/s]
 30%|██▉       | 1724/5758 [09:50<22:47,  2.95it/s]
                                                   

 30%|██▉       | 1724/5758 [09:50<22:47,  2.95it/s]
 30%|██▉       | 1725/5758 [09:50<25:51,  2.60it/s]
                                                   

 30%|██▉       | 1725/5758 [09:50<25:51,  2.60it/s]
 30%|██▉       | 1726/5758 [09:50<24:37,  2.73it/s]
                                                   

 30%|██▉       | 1726/5758 [09:50<24:37,  2.73it/s]
 30%|██▉       | 1727/5758 [09:51<23:48,  2.82it/s]
                                                   

 30%|██▉       | 1727/5758 [09:51<23:48,  2.82it/s]
 30%|███       | 1728/5758 [09:51<23:02,  2.92it/s]
                                                   

 30%|███       | 1728/5758 [09:51<23:02,  2.92it/s]
 30%|███       | 1729/5758 [09:51<21:58,  3.06it/s]
                                                   

 30%|███       | 1729/5758 [09:51<21:58,  3.06it/s]
 30%|███       | 1730/5758 [09:52<21:14,  3.16it/s]
                                                   

 30%|███       | 1730/5758 [09:52<21:14,  3.16it/s]
 30%|███       | 1731/5758 [09:52<21:56,  3.06it/s]
                                                   

 30%|███       | 1731/5758 [09:52<21:56,  3.06it/s]
 30%|███       | 1732/5758 [09:52<21:45,  3.08it/s]
                                                   

 30%|███       | 1732/5758 [09:52<21:45,  3.08it/s]
 30%|███       | 1733/5758 [09:53<21:47,  3.08it/s]
                                                   

 30%|███       | 1733/5758 [09:53<21:47,  3.08it/s]
 30%|███       | 1734/5758 [09:53<21:07,  3.18it/s]
                                                   

 30%|███       | 1734/5758 [09:53<21:07,  3.18it/s]
 30%|███       | 1735/5758 [09:53<20:45,  3.23it/s]
                                                   

 30%|███       | 1735/5758 [09:53<20:45,  3.23it/s]
 30%|███       | 1736/5758 [09:54<21:53,  3.06it/s]
                                                   

 30%|███       | 1736/5758 [09:54<21:53,  3.06it/s]
 30%|███       | 1737/5758 [09:54<21:26,  3.13it/s]
                                                   

 30%|███       | 1737/5758 [09:54<21:26,  3.13it/s]
 30%|█���█       | 1738/5758 [09:54<21:29,  3.12it/s]
                                                   

 30%|███       | 1738/5758 [09:54<21:29,  3.12it/s]
 30%|███       | 1739/5758 [09:55<21:29,  3.12it/s]
                                                   

 30%|███       | 1739/5758 [09:55<21:29,  3.12it/s]
 30%|███       | 1740/5758 [09:55<21:48,  3.07it/s]
                                                   

 30%|███       | 1740/5758 [09:55<21:48,  3.07it/s]
 30%|███       | 1741/5758 [09:55<22:04,  3.03it/s]
                                                   

 30%|███       | 1741/5758 [09:55<22:04,  3.03it/s]
 30%|███       | 1742/5758 [09:56<21:47,  3.07it/s]
                                                   

 30%|███       | 1742/5758 [09:56<21:47,  3.07it/s]
 30%|███       | 1743/5758 [09:56<21:55,  3.05it/s]
                                                   

 30%|███       | 1743/5758 [09:56<21:55,  3.05it/s]
 30%|███       | 1744/5758 [09:56<21:13,  3.15it/s]
                                                   

 30%|███       | 1744/5758 [09:56<21:13,  3.15it/s]
 30%|███       | 1745/5758 [09:57<20:46,  3.22it/s]
                                                   

 30%|███       | 1745/5758 [09:57<20:46,  3.22it/s]
 30%|███       | 1746/5758 [09:57<21:14,  3.15it/s]
                                                   

 30%|███       | 1746/5758 [09:57<21:14,  3.15it/s]
 30%|███       | 1747/5758 [09:57<22:58,  2.91it/s]
                                                   

 30%|███       | 1747/5758 [09:57<22:58,  2.91it/s]
 30%|███       | 1748/5758 [09:58<22:50,  2.93it/s]
                                                   

 30%|███       | 1748/5758 [09:58<22:50,  2.93it/s]
 30%|███       | 1749/5758 [09:58<22:26,  2.98it/s]
                                                   

 30%|███       | 1749/5758 [09:58<22:26,  2.98it/s]
 30%|███       | 1750/5758 [09:58<21:59,  3.04it/s]
                                                   

 30%|███       | 1750/5758 [09:58<21:59,  3.04it/s]
 30%|███       | 1751/5758 [09:59<24:06,  2.77it/s]
                                                   

 30%|███       | 1751/5758 [09:59<24:06,  2.77it/s]
 30%|███       | 1752/5758 [09:59<22:57,  2.91it/s]
                                                   

 30%|███       | 1752/5758 [09:59<22:57,  2.91it/s]
 30%|███       | 1753/5758 [09:59<22:16,  3.00it/s]
                                                   

 30%|███       | 1753/5758 [09:59<22:16,  3.00it/s]
 30%|███       | 1754/5758 [10:00<22:42,  2.94it/s]
                                                   

 30%|███       | 1754/5758 [10:00<22:42,  2.94it/s]
 30%|███       | 1755/5758 [10:00<23:35,  2.83it/s]
                                                   

 30%|███       | 1755/5758 [10:00<23:35,  2.83it/s]
 30%|███       | 1756/5758 [10:00<23:30,  2.84it/s]
                                                   

 30%|███       | 1756/5758 [10:00<23:30,  2.84it/s]
 31%|███       | 1757/5758 [10:01<23:28,  2.84it/s]
                                                   

 31%|███       | 1757/5758 [10:01<23:28,  2.84it/s]
 31%|███       | 1758/5758 [10:01<23:57,  2.78it/s]
                                                   

 31%|███       | 1758/5758 [10:01<23:57,  2.78it/s]
 31%|███       | 1759/5758 [10:01<24:15,  2.75it/s]
                                                   

 31%|███       | 1759/5758 [10:01<24:15,  2.75it/s]
 31%|███       | 1760/5758 [10:02<23:21,  2.85it/s]
                                                   

 31%|███       | 1760/5758 [10:02<23:21,  2.85it/s]
 31%|███       | 1761/5758 [10:02<23:13,  2.87it/s]
                                                   

 31%|███       | 1761/5758 [10:02<23:13,  2.87it/s]
 31%|███       | 1762/5758 [10:02<22:11,  3.00it/s]
                                                   

 31%|███       | 1762/5758 [10:02<22:11,  3.00it/s]
 31%|███       | 1763/5758 [10:03<21:32,  3.09it/s]
                                                   

 31%|███       | 1763/5758 [10:03<21:32,  3.09it/s]
 31%|███       | 1764/5758 [10:03<21:37,  3.08it/s]
                                                   

 31%|███       | 1764/5758 [10:03<21:37,  3.08it/s]
 31%|███       | 1765/5758 [10:03<21:22,  3.11it/s]
                                                   

 31%|███       | 1765/5758 [10:03<21:22,  3.11it/s]
 31%|███       | 1766/5758 [10:04<21:20,  3.12it/s]
                                                   

 31%|███       | 1766/5758 [10:04<21:20,  3.12it/s]
 31%|███       | 1767/5758 [10:04<21:31,  3.09it/s]
                                                   

 31%|███       | 1767/5758 [10:04<21:31,  3.09it/s]
 31%|███       | 1768/5758 [10:04<22:04,  3.01it/s]
                                                   

 31%|███       | 1768/5758 [10:04<22:04,  3.01it/s]
 31%|███       | 1769/5758 [10:05<21:41,  3.06it/s]
                                                   

 31%|███       | 1769/5758 [10:05<21:41,  3.06it/s]
 31%|███       | 1770/5758 [10:05<21:01,  3.16it/s]
                                                   

 31%|███       | 1770/5758 [10:05<21:01,  3.16it/s]
 31%|███       | 1771/5758 [10:05<21:12,  3.13it/s]
                                                   

 31%|███       | 1771/5758 [10:05<21:12,  3.13it/s]
 31%|███       | 1772/5758 [10:06<21:21,  3.11it/s]
                                                   

 31%|███       | 1772/5758 [10:06<21:21,  3.11it/s]
 31%|███       | 1773/5758 [10:06<21:31,  3.09it/s]
                                                   

 31%|███       | 1773/5758 [10:06<21:31,  3.09it/s]
 31%|███       | 1774/5758 [10:06<22:03,  3.01it/s]
                                                   

 31%|███       | 1774/5758 [10:06<22:03,  3.01it/s]
 31%|███       | 1775/5758 [10:07<21:33,  3.08it/s]
                                                   

 31%|███       | 1775/5758 [10:07<21:33,  3.08it/s]
 31%|███       | 1776/5758 [10:07<21:18,  3.12it/s]
                                                   
{'loss': '1.164', 'grad_norm': '12.69', 'learning_rate': '7.013e-06', 'num_tokens': '1.127e+06', 'mean_token_accuracy': '0.6907', 'epoch': '0.2989'}
+{'loss': '1.132', 'grad_norm': '13.88', 'learning_rate': '7.011e-06', 'num_tokens': '1.127e+06', 'mean_token_accuracy': '0.7036', 'epoch': '0.2991'}
+{'loss': '1.728', 'grad_norm': '25.38', 'learning_rate': '7.009e-06', 'num_tokens': '1.127e+06', 'mean_token_accuracy': '0.6852', 'epoch': '0.2992'}
+{'loss': '1.055', 'grad_norm': '13.38', 'learning_rate': '7.008e-06', 'num_tokens': '1.128e+06', 'mean_token_accuracy': '0.6952', 'epoch': '0.2994'}
+{'loss': '1.017', 'grad_norm': '12.62', 'learning_rate': '7.006e-06', 'num_tokens': '1.128e+06', 'mean_token_accuracy': '0.7528', 'epoch': '0.2996'}
+{'loss': '1.099', 'grad_norm': '12.5', 'learning_rate': '7.004e-06', 'num_tokens': '1.129e+06', 'mean_token_accuracy': '0.6896', 'epoch': '0.2998'}
+{'loss': '1.289', 'grad_norm': '16.38', 'learning_rate': '7.002e-06', 'num_tokens': '1.13e+06', 'mean_token_accuracy': '0.7299', 'epoch': '0.2999'}
+{'loss': '1.133', 'grad_norm': '13.94', 'learning_rate': '7.001e-06', 'num_tokens': '1.13e+06', 'mean_token_accuracy': '0.6855', 'epoch': '0.3001'}
+{'loss': '0.9072', 'grad_norm': '11.38', 'learning_rate': '6.999e-06', 'num_tokens': '1.131e+06', 'mean_token_accuracy': '0.7864', 'epoch': '0.3003'}
+{'loss': '1.051', 'grad_norm': '17.5', 'learning_rate': '6.997e-06', 'num_tokens': '1.131e+06', 'mean_token_accuracy': '0.7608', 'epoch': '0.3005'}
+{'loss': '0.572', 'grad_norm': '6.094', 'learning_rate': '6.995e-06', 'num_tokens': '1.133e+06', 'mean_token_accuracy': '0.8153', 'epoch': '0.3006'}
+{'loss': '0.637', 'grad_norm': '13.56', 'learning_rate': '6.994e-06', 'num_tokens': '1.134e+06', 'mean_token_accuracy': '0.8386', 'epoch': '0.3008'}
+{'loss': '1.301', 'grad_norm': '56.5', 'learning_rate': '6.992e-06', 'num_tokens': '1.134e+06', 'mean_token_accuracy': '0.7283', 'epoch': '0.301'}
+{'loss': '1.096', 'grad_norm': '12.69', 'learning_rate': '6.99e-06', 'num_tokens': '1.134e+06', 'mean_token_accuracy': '0.6938', 'epoch': '0.3011'}
+{'loss': '1.766', 'grad_norm': '20.88', 'learning_rate': '6.989e-06', 'num_tokens': '1.135e+06', 'mean_token_accuracy': '0.6456', 'epoch': '0.3013'}
+{'loss': '1.312', 'grad_norm': '28.5', 'learning_rate': '6.987e-06', 'num_tokens': '1.135e+06', 'mean_token_accuracy': '0.725', 'epoch': '0.3015'}
+{'loss': '1.763', 'grad_norm': '11.81', 'learning_rate': '6.985e-06', 'num_tokens': '1.135e+06', 'mean_token_accuracy': '0.6362', 'epoch': '0.3017'}
+{'loss': '2.034', 'grad_norm': '19.38', 'learning_rate': '6.983e-06', 'num_tokens': '1.136e+06', 'mean_token_accuracy': '0.5736', 'epoch': '0.3018'}
+{'loss': '0.5288', 'grad_norm': '6.438', 'learning_rate': '6.982e-06', 'num_tokens': '1.137e+06', 'mean_token_accuracy': '0.8291', 'epoch': '0.302'}
+{'loss': '1.111', 'grad_norm': '5.688', 'learning_rate': '6.98e-06', 'num_tokens': '1.141e+06', 'mean_token_accuracy': '0.7373', 'epoch': '0.3022'}
+{'loss': '2.102', 'grad_norm': '28.12', 'learning_rate': '6.978e-06', 'num_tokens': '1.141e+06', 'mean_token_accuracy': '0.5614', 'epoch': '0.3024'}
+{'loss': '1.908', 'grad_norm': '27.88', 'learning_rate': '6.976e-06', 'num_tokens': '1.141e+06', 'mean_token_accuracy': '0.6085', 'epoch': '0.3025'}
+{'loss': '1.806', 'grad_norm': '8.625', 'learning_rate': '6.975e-06', 'num_tokens': '1.143e+06', 'mean_token_accuracy': '0.6381', 'epoch': '0.3027'}
+{'loss': '0.9999', 'grad_norm': '11.56', 'learning_rate': '6.973e-06', 'num_tokens': '1.144e+06', 'mean_token_accuracy': '0.7238', 'epoch': '0.3029'}
+{'loss': '1.019', 'grad_norm': '12.62', 'learning_rate': '6.971e-06', 'num_tokens': '1.144e+06', 'mean_token_accuracy': '0.6992', 'epoch': '0.3031'}
+{'loss': '0.6962', 'grad_norm': '7.812', 'learning_rate': '6.969e-06', 'num_tokens': '1.146e+06', 'mean_token_accuracy': '0.7959', 'epoch': '0.3032'}
+{'loss': '1.77', 'grad_norm': '6.75', 'learning_rate': '6.968e-06', 'num_tokens': '1.149e+06', 'mean_token_accuracy': '0.6276', 'epoch': '0.3034'}
+{'loss': '0.8853', 'grad_norm': '17.5', 'learning_rate': '6.966e-06', 'num_tokens': '1.15e+06', 'mean_token_accuracy': '0.7613', 'epoch': '0.3036'}
+{'loss': '1.443', 'grad_norm': '32.75', 'learning_rate': '6.964e-06', 'num_tokens': '1.15e+06', 'mean_token_accuracy': '0.6848', 'epoch': '0.3038'}
+{'loss': '1.202', 'grad_norm': '12.5', 'learning_rate': '6.962e-06', 'num_tokens': '1.15e+06', 'mean_token_accuracy': '0.6501', 'epoch': '0.3039'}
+{'loss': '1.163', 'grad_norm': '11.5', 'learning_rate': '6.961e-06', 'num_tokens': '1.151e+06', 'mean_token_accuracy': '0.6795', 'epoch': '0.3041'}
+{'loss': '1.923', 'grad_norm': '11.75', 'learning_rate': '6.959e-06', 'num_tokens': '1.152e+06', 'mean_token_accuracy': '0.6423', 'epoch': '0.3043'}
+{'loss': '1.357', 'grad_norm': '12.56', 'learning_rate': '6.957e-06', 'num_tokens': '1.153e+06', 'mean_token_accuracy': '0.6237', 'epoch': '0.3044'}
+{'loss': '1.88', 'grad_norm': '16', 'learning_rate': '6.956e-06', 'num_tokens': '1.153e+06', 'mean_token_accuracy': '0.6024', 'epoch': '0.3046'}
+{'loss': '1.999', 'grad_norm': '9.062', 'learning_rate': '6.954e-06', 'num_tokens': '1.154e+06', 'mean_token_accuracy': '0.5742', 'epoch': '0.3048'}
+{'loss': '0.7034', 'grad_norm': '32', 'learning_rate': '6.952e-06', 'num_tokens': '1.154e+06', 'mean_token_accuracy': '0.8519', 'epoch': '0.305'}
+{'loss': '1.02', 'grad_norm': '9.75', 'learning_rate': '6.95e-06', 'num_tokens': '1.155e+06', 'mean_token_accuracy': '0.7789', 'epoch': '0.3051'}
+{'loss': '0.743', 'grad_norm': '10.31', 'learning_rate': '6.949e-06', 'num_tokens': '1.157e+06', 'mean_token_accuracy': '0.7636', 'epoch': '0.3053'}
+{'loss': '1.741', 'grad_norm': '20.88', 'learning_rate': '6.947e-06', 'num_tokens': '1.157e+06', 'mean_token_accuracy': '0.5986', 'epoch': '0.3055'}
+{'loss': '1.122', 'grad_norm': '11.88', 'learning_rate': '6.945e-06', 'num_tokens': '1.158e+06', 'mean_token_accuracy': '0.6805', 'epoch': '0.3057'}
+{'loss': '0.834', 'grad_norm': '39.25', 'learning_rate': '6.943e-06', 'num_tokens': '1.158e+06', 'mean_token_accuracy': '0.7818', 'epoch': '0.3058'}
+{'loss': '1.064', 'grad_norm': '12.75', 'learning_rate': '6.942e-06', 'num_tokens': '1.158e+06', 'mean_token_accuracy': '0.7023', 'epoch': '0.306'}
+{'loss': '1.218', 'grad_norm': '13.31', 'learning_rate': '6.94e-06', 'num_tokens': '1.159e+06', 'mean_token_accuracy': '0.6653', 'epoch': '0.3062'}
+{'loss': '0.4847', 'grad_norm': '6.219', 'learning_rate': '6.938e-06', 'num_tokens': '1.161e+06', 'mean_token_accuracy': '0.854', 'epoch': '0.3064'}
+{'loss': '1.357', 'grad_norm': '12.31', 'learning_rate': '6.936e-06', 'num_tokens': '1.162e+06', 'mean_token_accuracy': '0.6526', 'epoch': '0.3065'}
+{'loss': '1.814', 'grad_norm': '9.812', 'learning_rate': '6.935e-06', 'num_tokens': '1.162e+06', 'mean_token_accuracy': '0.6361', 'epoch': '0.3067'}
+{'loss': '1.31', 'grad_norm': '23.5', 'learning_rate': '6.933e-06', 'num_tokens': '1.163e+06', 'mean_token_accuracy': '0.6916', 'epoch': '0.3069'}
+{'loss': '1.771', 'grad_norm': '15.44', 'learning_rate': '6.931e-06', 'num_tokens': '1.163e+06', 'mean_token_accuracy': '0.6559', 'epoch': '0.3071'}
+{'loss': '1.294', 'grad_norm': '12.31', 'learning_rate': '6.929e-06', 'num_tokens': '1.164e+06', 'mean_token_accuracy': '0.6388', 'epoch': '0.3072'}
+{'loss': '2.084', 'grad_norm': '13.38', 'learning_rate': '6.928e-06', 'num_tokens': '1.165e+06', 'mean_token_accuracy': '0.5701', 'epoch': '0.3074'}
+{'loss': '1.362', 'grad_norm': '12.81', 'learning_rate': '6.926e-06', 'num_tokens': '1.165e+06', 'mean_token_accuracy': '0.6479', 'epoch': '0.3076'}
+{'loss': '1.314', 'grad_norm': '12.62', 'learning_rate': '6.924e-06', 'num_tokens': '1.166e+06', 'mean_token_accuracy': '0.6387', 'epoch': '0.3077'}
+{'loss': '2.136', 'grad_norm': '26.62', 'learning_rate': '6.923e-06', 'num_tokens': '1.166e+06', 'mean_token_accuracy': '0.599', 'epoch': '0.3079'}
+{'loss': '1.95', 'grad_norm': '14.19', 'learning_rate': '6.921e-06', 'num_tokens': '1.167e+06', 'mean_token_accuracy': '0.6192', 'epoch': '0.3081'}
+{'loss': '1.271', 'grad_norm': '12.94', 'learning_rate': '6.919e-06', 'num_tokens': '1.168e+06', 'mean_token_accuracy': '0.6549', 'epoch': '0.3083'}
+
 31%|███       | 1776/5758 [10:07<21:18,  3.12it/s]
 31%|███       | 1777/5758 [10:07<21:24,  3.10it/s]
                                                   

 31%|███       | 1777/5758 [10:07<21:24,  3.10it/s]
 31%|███       | 1778/5758 [10:08<25:08,  2.64it/s]
                                                   

 31%|███       | 1778/5758 [10:08<25:08,  2.64it/s]
 31%|███       | 1779/5758 [10:08<23:31,  2.82it/s]
                                                   

 31%|███       | 1779/5758 [10:08<23:31,  2.82it/s]
 31%|███       | 1780/5758 [10:08<22:13,  2.98it/s]
                                                   

 31%|███       | 1780/5758 [10:08<22:13,  2.98it/s]
 31%|███       | 1781/5758 [10:09<22:25,  2.95it/s]
                                                   

 31%|███       | 1781/5758 [10:09<22:25,  2.95it/s]
 31%|███       | 1782/5758 [10:09<21:30,  3.08it/s]
                                                   

 31%|███       | 1782/5758 [10:09<21:30,  3.08it/s]
 31%|███       | 1783/5758 [10:09<20:55,  3.17it/s]
                                                   

 31%|███       | 1783/5758 [10:09<20:55,  3.17it/s]
 31%|███       | 1784/5758 [10:10<21:21,  3.10it/s]
                                                   

 31%|███       | 1784/5758 [10:10<21:21,  3.10it/s]
 31%|███       | 1785/5758 [10:10<20:45,  3.19it/s]
                                                   

 31%|███       | 1785/5758 [10:10<20:45,  3.19it/s]
 31%|███       | 1786/5758 [10:10<20:47,  3.18it/s]
                                                   

 31%|███       | 1786/5758 [10:10<20:47,  3.18it/s]
 31%|███       | 1787/5758 [10:11<20:55,  3.16it/s]
                                                   

 31%|███       | 1787/5758 [10:11<20:55,  3.16it/s]
 31%|███       | 1788/5758 [10:11<21:10,  3.13it/s]
                                                   

 31%|███       | 1788/5758 [10:11<21:10,  3.13it/s]
 31%|███       | 1789/5758 [10:11<21:05,  3.14it/s]
                                                   

 31%|███       | 1789/5758 [10:11<21:05,  3.14it/s]
 31%|███       | 1790/5758 [10:12<22:23,  2.95it/s]
                                                   

 31%|███       | 1790/5758 [10:12<22:23,  2.95it/s]
 31%|███       | 1791/5758 [10:12<21:34,  3.06it/s]
                                                   

 31%|███       | 1791/5758 [10:12<21:34,  3.06it/s]
 31%|███       | 1792/5758 [10:12<21:12,  3.12it/s]
                                                   

 31%|███       | 1792/5758 [10:12<21:12,  3.12it/s]
 31%|███       | 1793/5758 [10:13<22:13,  2.97it/s]
                                                   

 31%|███       | 1793/5758 [10:13<22:13,  2.97it/s]
 31%|███       | 1794/5758 [10:13<21:50,  3.03it/s]
                                                   

 31%|███       | 1794/5758 [10:13<21:50,  3.03it/s]
 31%|███       | 1795/5758 [10:13<21:00,  3.14it/s]
                                                   

 31%|███       | 1795/5758 [10:13<21:00,  3.14it/s]
 31%|███       | 1796/5758 [10:14<22:18,  2.96it/s]
                                                   

 31%|███       | 1796/5758 [10:14<22:18,  2.96it/s]
 31%|███       | 1797/5758 [10:14<21:57,  3.01it/s]
                                                   

 31%|███       | 1797/5758 [10:14<21:57,  3.01it/s]
 31%|███       | 1798/5758 [10:14<21:45,  3.03it/s]
                                                   

 31%|███       | 1798/5758 [10:14<21:45,  3.03it/s]
 31%|███       | 1799/5758 [10:15<22:09,  2.98it/s]
                                                   

 31%|███       | 1799/5758 [10:15<22:09,  2.98it/s]
 31%|███▏      | 1800/5758 [10:15<21:18,  3.10it/s]
                                                   

 31%|███▏      | 1800/5758 [10:15<21:18,  3.10it/s]
 31%|███▏      | 1801/5758 [10:15<21:35,  3.05it/s]
                                                   

 31%|███▏      | 1801/5758 [10:15<21:35,  3.05it/s]
 31%|███▏      | 1802/5758 [10:15<21:19,  3.09it/s]
                                                   

 31%|███▏      | 1802/5758 [10:15<21:19,  3.09it/s]
 31%|███▏      | 1803/5758 [10:16<21:32,  3.06it/s]
                                                   

 31%|███▏      | 1803/5758 [10:16<21:32,  3.06it/s]
 31%|███▏      | 1804/5758 [10:16<21:16,  3.10it/s]
                                                   

 31%|███▏      | 1804/5758 [10:16<21:16,  3.10it/s]
 31%|███▏      | 1805/5758 [10:16<20:41,  3.18it/s]
                                                   

 31%|███▏      | 1805/5758 [10:16<20:41,  3.18it/s]
 31%|███▏      | 1806/5758 [10:17<20:46,  3.17it/s]
                                                   

 31%|███▏      | 1806/5758 [10:17<20:46,  3.17it/s]
 31%|███▏      | 1807/5758 [10:17<23:19,  2.82it/s]
                                                   

 31%|███▏      | 1807/5758 [10:17<23:19,  2.82it/s]
 31%|███▏      | 1808/5758 [10:18<22:29,  2.93it/s]
                                                   

 31%|███▏      | 1808/5758 [10:18<22:29,  2.93it/s]
 31%|███▏      | 1809/5758 [10:18<22:14,  2.96it/s]
                                                   

 31%|███▏      | 1809/5758 [10:18<22:14,  2.96it/s]
 31%|███▏      | 1810/5758 [10:18<20:42,  3.18it/s]
                                                   

 31%|███▏      | 1810/5758 [10:18<20:42,  3.18it/s]
 31%|███▏      | 1811/5758 [10:18<21:32,  3.05it/s]
                                                   

 31%|███▏      | 1811/5758 [10:18<21:32,  3.05it/s]
 31%|███▏      | 1812/5758 [10:19<20:51,  3.15it/s]
                                                   

 31%|███▏      | 1812/5758 [10:19<20:51,  3.15it/s]
 31%|███▏      | 1813/5758 [10:19<20:38,  3.19it/s]
                                                   

 31%|███▏      | 1813/5758 [10:19<20:38,  3.19it/s]
 32%|███▏      | 1814/5758 [10:19<21:08,  3.11it/s]
                                                   

 32%|███▏      | 1814/5758 [10:19<21:08,  3.11it/s]
 32%|███▏      | 1815/5758 [10:20<21:07,  3.11it/s]
                                                   

 32%|███▏      | 1815/5758 [10:20<21:07,  3.11it/s]
 32%|███▏      | 1816/5758 [10:20<21:34,  3.05it/s]
                                                   

 32%|███▏      | 1816/5758 [10:20<21:34,  3.05it/s]
 32%|███▏      | 1817/5758 [10:20<22:37,  2.90it/s]
                                                   

 32%|███▏      | 1817/5758 [10:20<22:37,  2.90it/s]
 32%|███▏      | 1818/5758 [10:21<21:57,  2.99it/s]
                                                   

 32%|███▏      | 1818/5758 [10:21<21:57,  2.99it/s]
 32%|███▏      | 1819/5758 [10:21<22:22,  2.93it/s]
                                                   

 32%|███▏      | 1819/5758 [10:21<22:22,  2.93it/s]
 32%|███▏      | 1820/5758 [10:21<21:42,  3.02it/s]
                                                   

 32%|███▏      | 1820/5758 [10:21<21:42,  3.02it/s]
 32%|███▏      | 1821/5758 [10:22<22:04,  2.97it/s]
                                                   

 32%|███▏      | 1821/5758 [10:22<22:04,  2.97it/s]
 32%|███▏      | 1822/5758 [10:22<22:27,  2.92it/s]
                                                   

 32%|███▏      | 1822/5758 [10:22<22:27,  2.92it/s]
 32%|███▏      | 1823/5758 [10:22<22:48,  2.88it/s]
                                                   

 32%|███▏      | 1823/5758 [10:22<22:48,  2.88it/s]
 32%|███▏      | 1824/5758 [10:23<23:23,  2.80it/s]
                                                   

 32%|███▏      | 1824/5758 [10:23<23:23,  2.80it/s]
 32%|███▏      | 1825/5758 [10:23<22:16,  2.94it/s]
                                                   

 32%|███▏      | 1825/5758 [10:23<22:16,  2.94it/s]
 32%|███▏      | 1826/5758 [10:24<22:38,  2.89it/s]
                                                   

 32%|███▏      | 1826/5758 [10:24<22:38,  2.89it/s]
 32%|███▏      | 1827/5758 [10:24<22:17,  2.94it/s]
                                                   

 32%|███▏      | 1827/5758 [10:24<22:17,  2.94it/s]
 32%|███▏      | 1828/5758 [10:24<21:37,  3.03it/s]
                                                   

 32%|███▏      | 1828/5758 [10:24<21:37,  3.03it/s]
 32%|███▏      | 1829/5758 [10:24<20:48,  3.15it/s]
                                                   

 32%|███▏      | 1829/5758 [10:24<20:48,  3.15it/s]
 32%|███▏      | 1830/5758 [10:25<20:11,  3.24it/s]
                                                   

 32%|███▏      | 1830/5758 [10:25<20:11,  3.24it/s]
 32%|███▏      | 1831/5758 [10:25<20:09,  3.25it/s]
                                                   
{'loss': '1.147', 'grad_norm': '12', 'learning_rate': '6.917e-06', 'num_tokens': '1.168e+06', 'mean_token_accuracy': '0.6969', 'epoch': '0.3084'}
+{'loss': '1.761', 'grad_norm': '11.25', 'learning_rate': '6.916e-06', 'num_tokens': '1.169e+06', 'mean_token_accuracy': '0.6307', 'epoch': '0.3086'}
+{'loss': '1.678', 'grad_norm': '15.56', 'learning_rate': '6.914e-06', 'num_tokens': '1.17e+06', 'mean_token_accuracy': '0.6433', 'epoch': '0.3088'}
+{'loss': '0.6142', 'grad_norm': '7.875', 'learning_rate': '6.912e-06', 'num_tokens': '1.171e+06', 'mean_token_accuracy': '0.8013', 'epoch': '0.309'}
+{'loss': '1.347', 'grad_norm': '13.5', 'learning_rate': '6.91e-06', 'num_tokens': '1.172e+06', 'mean_token_accuracy': '0.6316', 'epoch': '0.3091'}
+{'loss': '0.4758', 'grad_norm': '16.88', 'learning_rate': '6.909e-06', 'num_tokens': '1.172e+06', 'mean_token_accuracy': '0.8777', 'epoch': '0.3093'}
+{'loss': '1.887', 'grad_norm': '9.125', 'learning_rate': '6.907e-06', 'num_tokens': '1.173e+06', 'mean_token_accuracy': '0.6077', 'epoch': '0.3095'}
+{'loss': '1.966', 'grad_norm': '12.19', 'learning_rate': '6.905e-06', 'num_tokens': '1.174e+06', 'mean_token_accuracy': '0.6179', 'epoch': '0.3097'}
+{'loss': '1.943', 'grad_norm': '15.88', 'learning_rate': '6.903e-06', 'num_tokens': '1.174e+06', 'mean_token_accuracy': '0.593', 'epoch': '0.3098'}
+{'loss': '2.256', 'grad_norm': '13.88', 'learning_rate': '6.902e-06', 'num_tokens': '1.175e+06', 'mean_token_accuracy': '0.549', 'epoch': '0.31'}
+{'loss': '1.456', 'grad_norm': '14.5', 'learning_rate': '6.9e-06', 'num_tokens': '1.175e+06', 'mean_token_accuracy': '0.6929', 'epoch': '0.3102'}
+{'loss': '1.456', 'grad_norm': '13', 'learning_rate': '6.898e-06', 'num_tokens': '1.176e+06', 'mean_token_accuracy': '0.6297', 'epoch': '0.3104'}
+{'loss': '0.6518', 'grad_norm': '7.969', 'learning_rate': '6.896e-06', 'num_tokens': '1.178e+06', 'mean_token_accuracy': '0.7963', 'epoch': '0.3105'}
+{'loss': '1.377', 'grad_norm': '12.19', 'learning_rate': '6.895e-06', 'num_tokens': '1.179e+06', 'mean_token_accuracy': '0.645', 'epoch': '0.3107'}
+{'loss': '1.958', 'grad_norm': '19.75', 'learning_rate': '6.893e-06', 'num_tokens': '1.179e+06', 'mean_token_accuracy': '0.6034', 'epoch': '0.3109'}
+{'loss': '2.123', 'grad_norm': '29.38', 'learning_rate': '6.891e-06', 'num_tokens': '1.179e+06', 'mean_token_accuracy': '0.6', 'epoch': '0.311'}
+{'loss': '0.9999', 'grad_norm': '11.81', 'learning_rate': '6.89e-06', 'num_tokens': '1.18e+06', 'mean_token_accuracy': '0.7123', 'epoch': '0.3112'}
+{'loss': '1.556', 'grad_norm': '28.75', 'learning_rate': '6.888e-06', 'num_tokens': '1.18e+06', 'mean_token_accuracy': '0.7355', 'epoch': '0.3114'}
+{'loss': '1.644', 'grad_norm': '17.25', 'learning_rate': '6.886e-06', 'num_tokens': '1.18e+06', 'mean_token_accuracy': '0.6532', 'epoch': '0.3116'}
+{'loss': '1.359', 'grad_norm': '15.38', 'learning_rate': '6.884e-06', 'num_tokens': '1.181e+06', 'mean_token_accuracy': '0.6782', 'epoch': '0.3117'}
+{'loss': '1.184', 'grad_norm': '9.812', 'learning_rate': '6.883e-06', 'num_tokens': '1.182e+06', 'mean_token_accuracy': '0.723', 'epoch': '0.3119'}
+{'loss': '1.442', 'grad_norm': '12.19', 'learning_rate': '6.881e-06', 'num_tokens': '1.183e+06', 'mean_token_accuracy': '0.6723', 'epoch': '0.3121'}
+{'loss': '1.274', 'grad_norm': '21.38', 'learning_rate': '6.879e-06', 'num_tokens': '1.183e+06', 'mean_token_accuracy': '0.7196', 'epoch': '0.3123'}
+{'loss': '0.9243', 'grad_norm': '17.38', 'learning_rate': '6.877e-06', 'num_tokens': '1.184e+06', 'mean_token_accuracy': '0.794', 'epoch': '0.3124'}
+{'loss': '0.9745', 'grad_norm': '12.94', 'learning_rate': '6.876e-06', 'num_tokens': '1.184e+06', 'mean_token_accuracy': '0.7297', 'epoch': '0.3126'}
+{'loss': '1.198', 'grad_norm': '7.125', 'learning_rate': '6.874e-06', 'num_tokens': '1.186e+06', 'mean_token_accuracy': '0.6868', 'epoch': '0.3128'}
+{'loss': '1.11', 'grad_norm': '11.44', 'learning_rate': '6.872e-06', 'num_tokens': '1.187e+06', 'mean_token_accuracy': '0.6872', 'epoch': '0.313'}
+{'loss': '1.458', 'grad_norm': '16.38', 'learning_rate': '6.87e-06', 'num_tokens': '1.187e+06', 'mean_token_accuracy': '0.669', 'epoch': '0.3131'}
+{'loss': '1.373', 'grad_norm': '12.06', 'learning_rate': '6.869e-06', 'num_tokens': '1.188e+06', 'mean_token_accuracy': '0.6471', 'epoch': '0.3133'}
+{'loss': '0.8598', 'grad_norm': '12.75', 'learning_rate': '6.867e-06', 'num_tokens': '1.189e+06', 'mean_token_accuracy': '0.7441', 'epoch': '0.3135'}
+{'loss': '0.9581', 'grad_norm': '17.12', 'learning_rate': '6.865e-06', 'num_tokens': '1.189e+06', 'mean_token_accuracy': '0.7893', 'epoch': '0.3137'}
+{'loss': '1.143', 'grad_norm': '29.38', 'learning_rate': '6.863e-06', 'num_tokens': '1.189e+06', 'mean_token_accuracy': '0.7579', 'epoch': '0.3138'}
+{'loss': '0.9946', 'grad_norm': '24.62', 'learning_rate': '6.862e-06', 'num_tokens': '1.189e+06', 'mean_token_accuracy': '0.7624', 'epoch': '0.314'}
+{'loss': '1.057', 'grad_norm': '39.25', 'learning_rate': '6.86e-06', 'num_tokens': '1.189e+06', 'mean_token_accuracy': '0.7766', 'epoch': '0.3142'}
+{'loss': '0.6578', 'grad_norm': '52.5', 'learning_rate': '6.858e-06', 'num_tokens': '1.189e+06', 'mean_token_accuracy': '0.8261', 'epoch': '0.3143'}
+{'loss': '1.989', 'grad_norm': '18.12', 'learning_rate': '6.857e-06', 'num_tokens': '1.19e+06', 'mean_token_accuracy': '0.5992', 'epoch': '0.3145'}
+{'loss': '1.352', 'grad_norm': '14.62', 'learning_rate': '6.855e-06', 'num_tokens': '1.19e+06', 'mean_token_accuracy': '0.6283', 'epoch': '0.3147'}
+{'loss': '1.271', 'grad_norm': '12.12', 'learning_rate': '6.853e-06', 'num_tokens': '1.191e+06', 'mean_token_accuracy': '0.6588', 'epoch': '0.3149'}
+{'loss': '1.789', 'grad_norm': '34', 'learning_rate': '6.851e-06', 'num_tokens': '1.191e+06', 'mean_token_accuracy': '0.6327', 'epoch': '0.315'}
+{'loss': '0.7559', 'grad_norm': '23.25', 'learning_rate': '6.85e-06', 'num_tokens': '1.191e+06', 'mean_token_accuracy': '0.8333', 'epoch': '0.3152'}
+{'loss': '2.217', 'grad_norm': '13.38', 'learning_rate': '6.848e-06', 'num_tokens': '1.192e+06', 'mean_token_accuracy': '0.5328', 'epoch': '0.3154'}
+{'loss': '0.9977', 'grad_norm': '18.25', 'learning_rate': '6.846e-06', 'num_tokens': '1.193e+06', 'mean_token_accuracy': '0.8063', 'epoch': '0.3156'}
+{'loss': '1.073', 'grad_norm': '17.38', 'learning_rate': '6.844e-06', 'num_tokens': '1.193e+06', 'mean_token_accuracy': '0.759', 'epoch': '0.3157'}
+{'loss': '1.397', 'grad_norm': '20.75', 'learning_rate': '6.843e-06', 'num_tokens': '1.193e+06', 'mean_token_accuracy': '0.6911', 'epoch': '0.3159'}
+{'loss': '1.308', 'grad_norm': '13.31', 'learning_rate': '6.841e-06', 'num_tokens': '1.194e+06', 'mean_token_accuracy': '0.6415', 'epoch': '0.3161'}
+{'loss': '1.837', 'grad_norm': '9.688', 'learning_rate': '6.839e-06', 'num_tokens': '1.195e+06', 'mean_token_accuracy': '0.662', 'epoch': '0.3163'}
+{'loss': '1.617', 'grad_norm': '19.88', 'learning_rate': '6.837e-06', 'num_tokens': '1.195e+06', 'mean_token_accuracy': '0.6747', 'epoch': '0.3164'}
+{'loss': '1.882', 'grad_norm': '33.75', 'learning_rate': '6.836e-06', 'num_tokens': '1.195e+06', 'mean_token_accuracy': '0.5478', 'epoch': '0.3166'}
+{'loss': '1.775', 'grad_norm': '24.38', 'learning_rate': '6.834e-06', 'num_tokens': '1.196e+06', 'mean_token_accuracy': '0.6214', 'epoch': '0.3168'}
+{'loss': '1.005', 'grad_norm': '12.25', 'learning_rate': '6.832e-06', 'num_tokens': '1.196e+06', 'mean_token_accuracy': '0.7167', 'epoch': '0.317'}
+{'loss': '1.078', 'grad_norm': '9.875', 'learning_rate': '6.83e-06', 'num_tokens': '1.197e+06', 'mean_token_accuracy': '0.7462', 'epoch': '0.3171'}
+{'loss': '1.035', 'grad_norm': '10.31', 'learning_rate': '6.829e-06', 'num_tokens': '1.198e+06', 'mean_token_accuracy': '0.7332', 'epoch': '0.3173'}
+{'loss': '1.683', 'grad_norm': '11.25', 'learning_rate': '6.827e-06', 'num_tokens': '1.199e+06', 'mean_token_accuracy': '0.6602', 'epoch': '0.3175'}
+{'loss': '1.361', 'grad_norm': '14.25', 'learning_rate': '6.825e-06', 'num_tokens': '1.199e+06', 'mean_token_accuracy': '0.6472', 'epoch': '0.3176'}
+{'loss': '1.221', 'grad_norm': '13.62', 'learning_rate': '6.824e-06', 'num_tokens': '1.2e+06', 'mean_token_accuracy': '0.6661', 'epoch': '0.3178'}
+
 32%|███▏      | 1831/5758 [10:25<20:09,  3.25it/s]
 32%|███▏      | 1832/5758 [10:25<23:09,  2.83it/s]
                                                   

 32%|███▏      | 1832/5758 [10:25<23:09,  2.83it/s]
 32%|███▏      | 1833/5758 [10:26<22:50,  2.86it/s]
                                                   

 32%|███▏      | 1833/5758 [10:26<22:50,  2.86it/s]
 32%|███▏      | 1834/5758 [10:26<23:14,  2.81it/s]
                                                   

 32%|███▏      | 1834/5758 [10:26<23:14,  2.81it/s]
 32%|███▏      | 1835/5758 [10:27<23:06,  2.83it/s]
                                                   

 32%|███▏      | 1835/5758 [10:27<23:06,  2.83it/s]
 32%|███▏      | 1836/5758 [10:27<22:50,  2.86it/s]
                                                   

 32%|███▏      | 1836/5758 [10:27<22:50,  2.86it/s]
 32%|███▏      | 1837/5758 [10:27<23:20,  2.80it/s]
                                                   

 32%|███▏      | 1837/5758 [10:27<23:20,  2.80it/s]
 32%|███▏      | 1838/5758 [10:28<23:30,  2.78it/s]
                                                   

 32%|███▏      | 1838/5758 [10:28<23:30,  2.78it/s]
 32%|███▏      | 1839/5758 [10:28<22:33,  2.90it/s]
                                                   

 32%|███▏      | 1839/5758 [10:28<22:33,  2.90it/s]
 32%|███▏      | 1840/5758 [10:28<21:34,  3.03it/s]
                                                   

 32%|███▏      | 1840/5758 [10:28<21:34,  3.03it/s]
 32%|███▏      | 1841/5758 [10:29<21:40,  3.01it/s]
                                                   

 32%|███▏      | 1841/5758 [10:29<21:40,  3.01it/s]
 32%|███▏      | 1842/5758 [10:29<22:40,  2.88it/s]
                                                   

 32%|███▏      | 1842/5758 [10:29<22:40,  2.88it/s]
 32%|███▏      | 1843/5758 [10:29<21:30,  3.03it/s]
                                                   

 32%|███▏      | 1843/5758 [10:29<21:30,  3.03it/s]
 32%|███▏      | 1844/5758 [10:30<22:15,  2.93it/s]
                                                   

 32%|███▏      | 1844/5758 [10:30<22:15,  2.93it/s]
 32%|███▏      | 1845/5758 [10:30<21:41,  3.01it/s]
                                                   

 32%|███▏      | 1845/5758 [10:30<21:41,  3.01it/s]
 32%|███▏      | 1846/5758 [10:30<21:07,  3.09it/s]
                                                   

 32%|███▏      | 1846/5758 [10:30<21:07,  3.09it/s]
 32%|███▏      | 1847/5758 [10:31<21:43,  3.00it/s]
                                                   

 32%|███▏      | 1847/5758 [10:31<21:43,  3.00it/s]
 32%|███▏      | 1848/5758 [10:31<21:11,  3.07it/s]
                                                   

 32%|███▏      | 1848/5758 [10:31<21:11,  3.07it/s]
 32%|███▏      | 1849/5758 [10:31<20:27,  3.19it/s]
                                                   

 32%|███▏      | 1849/5758 [10:31<20:27,  3.19it/s]
 32%|███▏      | 1850/5758 [10:31<20:33,  3.17it/s]
                                                   

 32%|███▏      | 1850/5758 [10:31<20:33,  3.17it/s]
 32%|███▏      | 1851/5758 [10:32<21:49,  2.98it/s]
                                                   

 32%|███▏      | 1851/5758 [10:32<21:49,  2.98it/s]
 32%|███▏      | 1852/5758 [10:32<21:28,  3.03it/s]
                                                   

 32%|███▏      | 1852/5758 [10:32<21:28,  3.03it/s]
 32%|███▏      | 1853/5758 [10:32<20:45,  3.14it/s]
                                                   

 32%|███▏      | 1853/5758 [10:32<20:45,  3.14it/s]
 32%|███▏      | 1854/5758 [10:33<21:34,  3.02it/s]
                                                   

 32%|███▏      | 1854/5758 [10:33<21:34,  3.02it/s]
 32%|███▏      | 1855/5758 [10:33<21:15,  3.06it/s]
                                                   

 32%|███▏      | 1855/5758 [10:33<21:15,  3.06it/s]
 32%|███▏      | 1856/5758 [10:33<21:04,  3.09it/s]
                                                   

 32%|███▏      | 1856/5758 [10:33<21:04,  3.09it/s]
 32%|███▏      | 1857/5758 [10:34<20:59,  3.10it/s]
                                                   

 32%|███▏      | 1857/5758 [10:34<20:59,  3.10it/s]
 32%|███▏      | 1858/5758 [10:34<23:22,  2.78it/s]
                                                   

 32%|███▏      | 1858/5758 [10:34<23:22,  2.78it/s]
 32%|███▏      | 1859/5758 [10:35<22:13,  2.92it/s]
                                                   

 32%|███▏      | 1859/5758 [10:35<22:13,  2.92it/s]
 32%|███▏      | 1860/5758 [10:35<21:41,  2.99it/s]
                                                   

 32%|███▏      | 1860/5758 [10:35<21:41,  2.99it/s]
 32%|███▏      | 1861/5758 [10:35<21:47,  2.98it/s]
                                                   

 32%|███▏      | 1861/5758 [10:35<21:47,  2.98it/s]
 32%|███▏      | 1862/5758 [10:36<21:51,  2.97it/s]
                                                   

 32%|███▏      | 1862/5758 [10:36<21:51,  2.97it/s]
 32%|███▏      | 1863/5758 [10:36<22:51,  2.84it/s]
                                                   

 32%|███▏      | 1863/5758 [10:36<22:51,  2.84it/s]
 32%|███▏      | 1864/5758 [10:36<21:44,  2.98it/s]
                                                   

 32%|███▏      | 1864/5758 [10:36<21:44,  2.98it/s]
 32%|███▏      | 1865/5758 [10:37<22:16,  2.91it/s]
                                                   

 32%|███▏      | 1865/5758 [10:37<22:16,  2.91it/s]
 32%|███▏      | 1866/5758 [10:37<22:45,  2.85it/s]
                                                   

 32%|███▏      | 1866/5758 [10:37<22:45,  2.85it/s]
 32%|███▏      | 1867/5758 [10:37<22:00,  2.95it/s]
                                                   

 32%|███▏      | 1867/5758 [10:37<22:00,  2.95it/s]
 32%|███▏      | 1868/5758 [10:38<21:15,  3.05it/s]
                                                   

 32%|███▏      | 1868/5758 [10:38<21:15,  3.05it/s]
 32%|███▏      | 1869/5758 [10:38<21:42,  2.99it/s]
                                                   

 32%|███▏      | 1869/5758 [10:38<21:42,  2.99it/s]
 32%|███▏      | 1870/5758 [10:38<22:34,  2.87it/s]
                                                   

 32%|███▏      | 1870/5758 [10:38<22:34,  2.87it/s]
 32%|███▏      | 1871/5758 [10:39<21:46,  2.97it/s]
                                                   

 32%|███▏      | 1871/5758 [10:39<21:46,  2.97it/s]
 33%|███▎      | 1872/5758 [10:39<21:09,  3.06it/s]
                                                   

 33%|███▎      | 1872/5758 [10:39<21:09,  3.06it/s]
 33%|███▎      | 1873/5758 [10:39<20:39,  3.13it/s]
                                                   

 33%|███▎      | 1873/5758 [10:39<20:39,  3.13it/s]
 33%|███▎      | 1874/5758 [10:40<21:36,  3.00it/s]
                                                   

 33%|███▎      | 1874/5758 [10:40<21:36,  3.00it/s]
 33%|███▎      | 1875/5758 [10:40<22:33,  2.87it/s]
                                                   

 33%|███▎      | 1875/5758 [10:40<22:33,  2.87it/s]
 33%|███▎      | 1876/5758 [10:40<21:42,  2.98it/s]
                                                   

 33%|███▎      | 1876/5758 [10:40<21:42,  2.98it/s]
 33%|███▎      | 1877/5758 [10:41<21:30,  3.01it/s]
                                                   

 33%|███▎      | 1877/5758 [10:41<21:30,  3.01it/s]
 33%|███▎      | 1878/5758 [10:41<22:16,  2.90it/s]
                                                   

 33%|███▎      | 1878/5758 [10:41<22:16,  2.90it/s]
 33%|███▎      | 1879/5758 [10:41<21:22,  3.02it/s]
                                                   

 33%|███▎      | 1879/5758 [10:41<21:22,  3.02it/s]
 33%|███▎      | 1880/5758 [10:42<21:06,  3.06it/s]
                                                   

 33%|███▎      | 1880/5758 [10:42<21:06,  3.06it/s]
 33%|███▎      | 1881/5758 [10:42<21:29,  3.01it/s]
                                                   

 33%|███▎      | 1881/5758 [10:42<21:29,  3.01it/s]
 33%|███▎      | 1882/5758 [10:42<21:08,  3.06it/s]
                                                   

 33%|███▎      | 1882/5758 [10:42<21:08,  3.06it/s]
 33%|███▎      | 1883/5758 [10:43<22:21,  2.89it/s]
                                                   

 33%|███▎      | 1883/5758 [10:43<22:21,  2.89it/s]
 33%|███▎      | 1884/5758 [10:43<21:50,  2.96it/s]
                                                   

 33%|███▎      | 1884/5758 [10:43<21:50,  2.96it/s]
 33%|███▎      | 1885/5758 [10:43<24:28,  2.64it/s]
                                                   

 33%|███▎      | 1885/5758 [10:43<24:28,  2.64it/s]
 33%|███▎      | 1886/5758 [10:44<24:46,  2.60it/s]
                                                   
{'loss': '1.557', 'grad_norm': '13.25', 'learning_rate': '6.822e-06', 'num_tokens': '1.201e+06', 'mean_token_accuracy': '0.6171', 'epoch': '0.318'}
+{'loss': '1.871', 'grad_norm': '11.31', 'learning_rate': '6.82e-06', 'num_tokens': '1.202e+06', 'mean_token_accuracy': '0.5932', 'epoch': '0.3182'}
+{'loss': '1.132', 'grad_norm': '10.19', 'learning_rate': '6.818e-06', 'num_tokens': '1.202e+06', 'mean_token_accuracy': '0.735', 'epoch': '0.3183'}
+{'loss': '2.927', 'grad_norm': '47.25', 'learning_rate': '6.817e-06', 'num_tokens': '1.203e+06', 'mean_token_accuracy': '0.475', 'epoch': '0.3185'}
+{'loss': '1.799', 'grad_norm': '16.62', 'learning_rate': '6.815e-06', 'num_tokens': '1.203e+06', 'mean_token_accuracy': '0.6186', 'epoch': '0.3187'}
+{'loss': '0.9041', 'grad_norm': '22.62', 'learning_rate': '6.813e-06', 'num_tokens': '1.203e+06', 'mean_token_accuracy': '0.7822', 'epoch': '0.3189'}
+{'loss': '1.322', 'grad_norm': '26.5', 'learning_rate': '6.811e-06', 'num_tokens': '1.203e+06', 'mean_token_accuracy': '0.7177', 'epoch': '0.319'}
+{'loss': '1.472', 'grad_norm': '15.69', 'learning_rate': '6.81e-06', 'num_tokens': '1.204e+06', 'mean_token_accuracy': '0.5817', 'epoch': '0.3192'}
+{'loss': '1.483', 'grad_norm': '15.94', 'learning_rate': '6.808e-06', 'num_tokens': '1.204e+06', 'mean_token_accuracy': '0.6872', 'epoch': '0.3194'}
+{'loss': '1.306', 'grad_norm': '13.19', 'learning_rate': '6.806e-06', 'num_tokens': '1.205e+06', 'mean_token_accuracy': '0.6475', 'epoch': '0.3196'}
+{'loss': '0.7345', 'grad_norm': '7.906', 'learning_rate': '6.804e-06', 'num_tokens': '1.207e+06', 'mean_token_accuracy': '0.781', 'epoch': '0.3197'}
+{'loss': '2.19', 'grad_norm': '22', 'learning_rate': '6.803e-06', 'num_tokens': '1.207e+06', 'mean_token_accuracy': '0.5863', 'epoch': '0.3199'}
+{'loss': '2.195', 'grad_norm': '17.5', 'learning_rate': '6.801e-06', 'num_tokens': '1.207e+06', 'mean_token_accuracy': '0.5418', 'epoch': '0.3201'}
+{'loss': '0.3665', 'grad_norm': '19.75', 'learning_rate': '6.799e-06', 'num_tokens': '1.208e+06', 'mean_token_accuracy': '0.9083', 'epoch': '0.3203'}
+{'loss': '1.148', 'grad_norm': '11.81', 'learning_rate': '6.797e-06', 'num_tokens': '1.208e+06', 'mean_token_accuracy': '0.6939', 'epoch': '0.3204'}
+{'loss': '1.421', 'grad_norm': '11.38', 'learning_rate': '6.796e-06', 'num_tokens': '1.209e+06', 'mean_token_accuracy': '0.7115', 'epoch': '0.3206'}
+{'loss': '1.896', 'grad_norm': '19', 'learning_rate': '6.794e-06', 'num_tokens': '1.209e+06', 'mean_token_accuracy': '0.6245', 'epoch': '0.3208'}
+{'loss': '1.02', 'grad_norm': '45.25', 'learning_rate': '6.792e-06', 'num_tokens': '1.21e+06', 'mean_token_accuracy': '0.7875', 'epoch': '0.3209'}
+{'loss': '0.6019', 'grad_norm': '14.75', 'learning_rate': '6.791e-06', 'num_tokens': '1.21e+06', 'mean_token_accuracy': '0.8414', 'epoch': '0.3211'}
+{'loss': '1.352', 'grad_norm': '21.12', 'learning_rate': '6.789e-06', 'num_tokens': '1.21e+06', 'mean_token_accuracy': '0.6943', 'epoch': '0.3213'}
+{'loss': '2.321', 'grad_norm': '36.75', 'learning_rate': '6.787e-06', 'num_tokens': '1.21e+06', 'mean_token_accuracy': '0.5615', 'epoch': '0.3215'}
+{'loss': '1.016', 'grad_norm': '19.88', 'learning_rate': '6.785e-06', 'num_tokens': '1.21e+06', 'mean_token_accuracy': '0.7892', 'epoch': '0.3216'}
+{'loss': '0.775', 'grad_norm': '17.12', 'learning_rate': '6.784e-06', 'num_tokens': '1.211e+06', 'mean_token_accuracy': '0.7968', 'epoch': '0.3218'}
+{'loss': '2.031', 'grad_norm': '10.38', 'learning_rate': '6.782e-06', 'num_tokens': '1.212e+06', 'mean_token_accuracy': '0.5914', 'epoch': '0.322'}
+{'loss': '0.5427', 'grad_norm': '6.875', 'learning_rate': '6.78e-06', 'num_tokens': '1.213e+06', 'mean_token_accuracy': '0.8341', 'epoch': '0.3222'}
+{'loss': '1.246', 'grad_norm': '12.12', 'learning_rate': '6.778e-06', 'num_tokens': '1.214e+06', 'mean_token_accuracy': '0.6591', 'epoch': '0.3223'}
+{'loss': '1.201', 'grad_norm': '11.94', 'learning_rate': '6.777e-06', 'num_tokens': '1.215e+06', 'mean_token_accuracy': '0.6748', 'epoch': '0.3225'}
+{'loss': '2.356', 'grad_norm': '15.5', 'learning_rate': '6.775e-06', 'num_tokens': '1.216e+06', 'mean_token_accuracy': '0.5493', 'epoch': '0.3227'}
+{'loss': '1.128', 'grad_norm': '26.62', 'learning_rate': '6.773e-06', 'num_tokens': '1.216e+06', 'mean_token_accuracy': '0.7707', 'epoch': '0.3229'}
+{'loss': '1.138', 'grad_norm': '12.56', 'learning_rate': '6.771e-06', 'num_tokens': '1.216e+06', 'mean_token_accuracy': '0.6882', 'epoch': '0.323'}
+{'loss': '1.743', 'grad_norm': '8.812', 'learning_rate': '6.77e-06', 'num_tokens': '1.218e+06', 'mean_token_accuracy': '0.6267', 'epoch': '0.3232'}
+{'loss': '2.184', 'grad_norm': '16.88', 'learning_rate': '6.768e-06', 'num_tokens': '1.219e+06', 'mean_token_accuracy': '0.5762', 'epoch': '0.3234'}
+{'loss': '1.064', 'grad_norm': '15.25', 'learning_rate': '6.766e-06', 'num_tokens': '1.219e+06', 'mean_token_accuracy': '0.6817', 'epoch': '0.3235'}
+{'loss': '0.8528', 'grad_norm': '14.62', 'learning_rate': '6.765e-06', 'num_tokens': '1.22e+06', 'mean_token_accuracy': '0.7886', 'epoch': '0.3237'}
+{'loss': '1.432', 'grad_norm': '4.906', 'learning_rate': '6.763e-06', 'num_tokens': '1.224e+06', 'mean_token_accuracy': '0.6549', 'epoch': '0.3239'}
+{'loss': '1.201', 'grad_norm': '17.5', 'learning_rate': '6.761e-06', 'num_tokens': '1.224e+06', 'mean_token_accuracy': '0.7597', 'epoch': '0.3241'}
+{'loss': '1.025', 'grad_norm': '11.56', 'learning_rate': '6.759e-06', 'num_tokens': '1.225e+06', 'mean_token_accuracy': '0.6974', 'epoch': '0.3242'}
+{'loss': '0.8931', 'grad_norm': '17.5', 'learning_rate': '6.758e-06', 'num_tokens': '1.225e+06', 'mean_token_accuracy': '0.8108', 'epoch': '0.3244'}
+{'loss': '1.42', 'grad_norm': '11.12', 'learning_rate': '6.756e-06', 'num_tokens': '1.226e+06', 'mean_token_accuracy': '0.62', 'epoch': '0.3246'}
+{'loss': '0.4804', 'grad_norm': '5.312', 'learning_rate': '6.754e-06', 'num_tokens': '1.228e+06', 'mean_token_accuracy': '0.8597', 'epoch': '0.3248'}
+{'loss': '0.6247', 'grad_norm': '8', 'learning_rate': '6.752e-06', 'num_tokens': '1.23e+06', 'mean_token_accuracy': '0.8127', 'epoch': '0.3249'}
+{'loss': '1.774', 'grad_norm': '20.38', 'learning_rate': '6.751e-06', 'num_tokens': '1.23e+06', 'mean_token_accuracy': '0.6375', 'epoch': '0.3251'}
+{'loss': '1.282', 'grad_norm': '14.94', 'learning_rate': '6.749e-06', 'num_tokens': '1.231e+06', 'mean_token_accuracy': '0.6471', 'epoch': '0.3253'}
+{'loss': '1.927', 'grad_norm': '13.12', 'learning_rate': '6.747e-06', 'num_tokens': '1.231e+06', 'mean_token_accuracy': '0.5655', 'epoch': '0.3255'}
+{'loss': '1.253', 'grad_norm': '14.56', 'learning_rate': '6.745e-06', 'num_tokens': '1.232e+06', 'mean_token_accuracy': '0.6413', 'epoch': '0.3256'}
+{'loss': '1.075', 'grad_norm': '11.69', 'learning_rate': '6.744e-06', 'num_tokens': '1.232e+06', 'mean_token_accuracy': '0.6975', 'epoch': '0.3258'}
+{'loss': '1.594', 'grad_norm': '28.75', 'learning_rate': '6.742e-06', 'num_tokens': '1.233e+06', 'mean_token_accuracy': '0.6633', 'epoch': '0.326'}
+{'loss': '1.138', 'grad_norm': '14.94', 'learning_rate': '6.74e-06', 'num_tokens': '1.233e+06', 'mean_token_accuracy': '0.7251', 'epoch': '0.3262'}
+{'loss': '1.292', 'grad_norm': '12.75', 'learning_rate': '6.738e-06', 'num_tokens': '1.234e+06', 'mean_token_accuracy': '0.6572', 'epoch': '0.3263'}
+{'loss': '1.602', 'grad_norm': '16.38', 'learning_rate': '6.737e-06', 'num_tokens': '1.234e+06', 'mean_token_accuracy': '0.6639', 'epoch': '0.3265'}
+{'loss': '1.695', 'grad_norm': '13.19', 'learning_rate': '6.735e-06', 'num_tokens': '1.235e+06', 'mean_token_accuracy': '0.6338', 'epoch': '0.3267'}
+{'loss': '1.434', 'grad_norm': '44.75', 'learning_rate': '6.733e-06', 'num_tokens': '1.235e+06', 'mean_token_accuracy': '0.6744', 'epoch': '0.3268'}
+{'loss': '0.9258', 'grad_norm': '24.38', 'learning_rate': '6.732e-06', 'num_tokens': '1.235e+06', 'mean_token_accuracy': '0.7863', 'epoch': '0.327'}
+{'loss': '1.466', 'grad_norm': '14.5', 'learning_rate': '6.73e-06', 'num_tokens': '1.235e+06', 'mean_token_accuracy': '0.664', 'epoch': '0.3272'}
+{'loss': '1.656', 'grad_norm': '21.25', 'learning_rate': '6.728e-06', 'num_tokens': '1.236e+06', 'mean_token_accuracy': '0.6128', 'epoch': '0.3274'}
+
 33%|███▎      | 1886/5758 [10:44<24:46,  2.60it/s]
 33%|███▎      | 1887/5758 [10:44<23:00,  2.80it/s]
                                                   

 33%|███▎      | 1887/5758 [10:44<23:00,  2.80it/s]
 33%|███▎      | 1888/5758 [10:44<23:09,  2.79it/s]
                                                   

 33%|███▎      | 1888/5758 [10:44<23:09,  2.79it/s]
 33%|███▎      | 1889/5758 [10:45<22:01,  2.93it/s]
                                                   

 33%|███▎      | 1889/5758 [10:45<22:01,  2.93it/s]
 33%|███▎      | 1890/5758 [10:45<21:38,  2.98it/s]
                                                   

 33%|███▎      | 1890/5758 [10:45<21:38,  2.98it/s]
 33%|███▎      | 1891/5758 [10:45<21:15,  3.03it/s]
                                                   

 33%|███▎      | 1891/5758 [10:45<21:15,  3.03it/s]
 33%|███▎      | 1892/5758 [10:46<20:06,  3.20it/s]
                                                   

 33%|███▎      | 1892/5758 [10:46<20:06,  3.20it/s]
 33%|███▎      | 1893/5758 [10:46<21:04,  3.06it/s]
                                                   

 33%|███▎      | 1893/5758 [10:46<21:04,  3.06it/s]
 33%|███▎      | 1894/5758 [10:46<20:33,  3.13it/s]
                                                   

 33%|███▎      | 1894/5758 [10:46<20:33,  3.13it/s]
 33%|███▎      | 1895/5758 [10:47<20:13,  3.18it/s]
                                                   

 33%|███▎      | 1895/5758 [10:47<20:13,  3.18it/s]
 33%|███▎      | 1896/5758 [10:47<21:44,  2.96it/s]
                                                   

 33%|███▎      | 1896/5758 [10:47<21:44,  2.96it/s]
 33%|███▎      | 1897/5758 [10:47<21:49,  2.95it/s]
                                                   

 33%|███▎      | 1897/5758 [10:47<21:49,  2.95it/s]
 33%|███▎      | 1898/5758 [10:48<21:56,  2.93it/s]
                                                   

 33%|███▎      | 1898/5758 [10:48<21:56,  2.93it/s]
 33%|███▎      | 1899/5758 [10:48<21:36,  2.98it/s]
                                                   

 33%|███▎      | 1899/5758 [10:48<21:36,  2.98it/s]
 33%|███▎      | 1900/5758 [10:48<20:44,  3.10it/s]
                                                   

 33%|███▎      | 1900/5758 [10:48<20:44,  3.10it/s]
 33%|███▎      | 1901/5758 [10:49<20:27,  3.14it/s]
                                                   

 33%|███▎      | 1901/5758 [10:49<20:27,  3.14it/s]
 33%|███▎      | 1902/5758 [10:49<20:33,  3.13it/s]
                                                   

 33%|███▎      | 1902/5758 [10:49<20:33,  3.13it/s]
 33%|███▎      | 1903/5758 [10:49<20:38,  3.11it/s]
                                                   

 33%|███▎      | 1903/5758 [10:49<20:38,  3.11it/s]
 33%|███▎      | 1904/5758 [10:50<20:55,  3.07it/s]
                                                   

 33%|███▎      | 1904/5758 [10:50<20:55,  3.07it/s]
 33%|███▎      | 1905/5758 [10:50<20:38,  3.11it/s]
                                                   

 33%|███▎      | 1905/5758 [10:50<20:38,  3.11it/s]
 33%|███▎      | 1906/5758 [10:50<21:22,  3.00it/s]
                                                   

 33%|███▎      | 1906/5758 [10:50<21:22,  3.00it/s]
 33%|███▎      | 1907/5758 [10:51<22:10,  2.89it/s]
                                                   

 33%|███▎      | 1907/5758 [10:51<22:10,  2.89it/s]
 33%|███▎      | 1908/5758 [10:51<22:34,  2.84it/s]
                                                   

 33%|███▎      | 1908/5758 [10:51<22:34,  2.84it/s]
 33%|███▎      | 1909/5758 [10:51<21:43,  2.95it/s]
                                                   

 33%|███▎      | 1909/5758 [10:51<21:43,  2.95it/s]
 33%|███▎      | 1910/5758 [10:52<21:35,  2.97it/s]
                                                   

 33%|███▎      | 1910/5758 [10:52<21:35,  2.97it/s]
 33%|███▎      | 1911/5758 [10:52<20:49,  3.08it/s]
                                                   

 33%|███▎      | 1911/5758 [10:52<20:49,  3.08it/s]
 33%|███▎      | 1912/5758 [10:52<20:17,  3.16it/s]
                                                   

 33%|███▎      | 1912/5758 [10:52<20:17,  3.16it/s]
 33%|███▎      | 1913/5758 [10:53<21:28,  2.98it/s]
                                                   

 33%|███▎      | 1913/5758 [10:53<21:28,  2.98it/s]
 33%|███▎      | 1914/5758 [10:53<24:39,  2.60it/s]
                                                   

 33%|███▎      | 1914/5758 [10:53<24:39,  2.60it/s]
 33%|███▎      | 1915/5758 [10:54<24:46,  2.59it/s]
                                                   

 33%|███▎      | 1915/5758 [10:54<24:46,  2.59it/s]
 33%|███▎      | 1916/5758 [10:54<23:42,  2.70it/s]
                                                   

 33%|███▎      | 1916/5758 [10:54<23:42,  2.70it/s]
 33%|███▎      | 1917/5758 [10:54<23:06,  2.77it/s]
                                                   

 33%|███▎      | 1917/5758 [10:54<23:06,  2.77it/s]
 33%|███▎      | 1918/5758 [10:55<22:03,  2.90it/s]
                                                   

 33%|███▎      | 1918/5758 [10:55<22:03,  2.90it/s]
 33%|███▎      | 1919/5758 [10:55<21:42,  2.95it/s]
                                                   

 33%|███▎      | 1919/5758 [10:55<21:42,  2.95it/s]
 33%|███▎      | 1920/5758 [10:55<21:10,  3.02it/s]
                                                   

 33%|███▎      | 1920/5758 [10:55<21:10,  3.02it/s]
 33%|███▎      | 1921/5758 [10:55<20:33,  3.11it/s]
                                                   

 33%|███▎      | 1921/5758 [10:55<20:33,  3.11it/s]
 33%|███▎      | 1922/5758 [10:56<20:58,  3.05it/s]
                                                   

 33%|███▎      | 1922/5758 [10:56<20:58,  3.05it/s]
 33%|███▎      | 1923/5758 [10:56<21:40,  2.95it/s]
                                                   

 33%|███▎      | 1923/5758 [10:56<21:40,  2.95it/s]
 33%|███▎      | 1924/5758 [10:56<21:05,  3.03it/s]
                                                   

 33%|███▎      | 1924/5758 [10:57<21:05,  3.03it/s]
 33%|███▎      | 1925/5758 [10:57<21:24,  2.98it/s]
                                                   

 33%|███▎      | 1925/5758 [10:57<21:24,  2.98it/s]
 33%|███▎      | 1926/5758 [10:57<20:32,  3.11it/s]
                                                   

 33%|███▎      | 1926/5758 [10:57<20:32,  3.11it/s]
 33%|███▎      | 1927/5758 [10:58<21:31,  2.97it/s]
                                                   

 33%|███▎      | 1927/5758 [10:58<21:31,  2.97it/s]
 33%|███▎      | 1928/5758 [10:58<22:09,  2.88it/s]
                                                   

 33%|███▎      | 1928/5758 [10:58<22:09,  2.88it/s]
 34%|███▎      | 1929/5758 [10:58<21:53,  2.92it/s]
                                                   

 34%|███▎      | 1929/5758 [10:58<21:53,  2.92it/s]
 34%|███▎      | 1930/5758 [10:59<22:33,  2.83it/s]
                                                   

 34%|███▎      | 1930/5758 [10:59<22:33,  2.83it/s]
 34%|███▎      | 1931/5758 [10:59<22:15,  2.87it/s]
                                                   

 34%|███▎      | 1931/5758 [10:59<22:15,  2.87it/s]
 34%|███▎      | 1932/5758 [10:59<22:05,  2.89it/s]
                                                   

 34%|███▎      | 1932/5758 [10:59<22:05,  2.89it/s]
 34%|███▎      | 1933/5758 [11:00<21:25,  2.97it/s]
                                                   

 34%|███▎      | 1933/5758 [11:00<21:25,  2.97it/s]
 34%|███▎      | 1934/5758 [11:00<20:39,  3.08it/s]
                                                   

 34%|███▎      | 1934/5758 [11:00<20:39,  3.08it/s]
 34%|███▎      | 1935/5758 [11:00<19:59,  3.19it/s]
                                                   

 34%|███▎      | 1935/5758 [11:00<19:59,  3.19it/s]
 34%|███▎      | 1936/5758 [11:00<20:09,  3.16it/s]
                                                   

 34%|███▎      | 1936/5758 [11:00<20:09,  3.16it/s]
 34%|███▎      | 1937/5758 [11:01<21:11,  3.00it/s]
                                                   

 34%|███▎      | 1937/5758 [11:01<21:11,  3.00it/s]
 34%|███▎      | 1938/5758 [11:01<21:18,  2.99it/s]
                                                   

 34%|███▎      | 1938/5758 [11:01<21:18,  2.99it/s]
 34%|███▎      | 1939/5758 [11:02<21:34,  2.95it/s]
                                                   

 34%|███▎      | 1939/5758 [11:02<21:34,  2.95it/s]
 34%|███▎      | 1940/5758 [11:02<23:42,  2.68it/s]
                                                   

 34%|███▎      | 1940/5758 [11:02<23:42,  2.68it/s]
 34%|███▎      | 1941/5758 [11:02<22:58,  2.77it/s]
                                                   
{'loss': '1.804', 'grad_norm': '13', 'learning_rate': '6.726e-06', 'num_tokens': '1.236e+06', 'mean_token_accuracy': '0.6064', 'epoch': '0.3275'}
+{'loss': '1.045', 'grad_norm': '12.56', 'learning_rate': '6.725e-06', 'num_tokens': '1.237e+06', 'mean_token_accuracy': '0.7283', 'epoch': '0.3277'}
+{'loss': '1.534', 'grad_norm': '36', 'learning_rate': '6.723e-06', 'num_tokens': '1.237e+06', 'mean_token_accuracy': '0.6435', 'epoch': '0.3279'}
+{'loss': '1.229', 'grad_norm': '13.38', 'learning_rate': '6.721e-06', 'num_tokens': '1.238e+06', 'mean_token_accuracy': '0.6782', 'epoch': '0.3281'}
+{'loss': '1.315', 'grad_norm': '12.69', 'learning_rate': '6.719e-06', 'num_tokens': '1.238e+06', 'mean_token_accuracy': '0.6388', 'epoch': '0.3282'}
+{'loss': '1.202', 'grad_norm': '11.69', 'learning_rate': '6.718e-06', 'num_tokens': '1.239e+06', 'mean_token_accuracy': '0.6793', 'epoch': '0.3284'}
+{'loss': '0.9618', 'grad_norm': '41', 'learning_rate': '6.716e-06', 'num_tokens': '1.239e+06', 'mean_token_accuracy': '0.8246', 'epoch': '0.3286'}
+{'loss': '2.231', 'grad_norm': '20', 'learning_rate': '6.714e-06', 'num_tokens': '1.24e+06', 'mean_token_accuracy': '0.5768', 'epoch': '0.3288'}
+{'loss': '2.081', 'grad_norm': '21.25', 'learning_rate': '6.712e-06', 'num_tokens': '1.24e+06', 'mean_token_accuracy': '0.5576', 'epoch': '0.3289'}
+{'loss': '1.894', 'grad_norm': '11.44', 'learning_rate': '6.711e-06', 'num_tokens': '1.241e+06', 'mean_token_accuracy': '0.6121', 'epoch': '0.3291'}
+{'loss': '1.125', 'grad_norm': '17.62', 'learning_rate': '6.709e-06', 'num_tokens': '1.241e+06', 'mean_token_accuracy': '0.7475', 'epoch': '0.3293'}
+{'loss': '2.149', 'grad_norm': '16.5', 'learning_rate': '6.707e-06', 'num_tokens': '1.241e+06', 'mean_token_accuracy': '0.559', 'epoch': '0.3295'}
+{'loss': '1.554', 'grad_norm': '19.62', 'learning_rate': '6.705e-06', 'num_tokens': '1.242e+06', 'mean_token_accuracy': '0.6502', 'epoch': '0.3296'}
+{'loss': '1.188', 'grad_norm': '12.81', 'learning_rate': '6.704e-06', 'num_tokens': '1.243e+06', 'mean_token_accuracy': '0.6659', 'epoch': '0.3298'}
+{'loss': '2.347', 'grad_norm': '14.44', 'learning_rate': '6.702e-06', 'num_tokens': '1.243e+06', 'mean_token_accuracy': '0.5739', 'epoch': '0.33'}
+{'loss': '1.02', 'grad_norm': '11.12', 'learning_rate': '6.7e-06', 'num_tokens': '1.244e+06', 'mean_token_accuracy': '0.7178', 'epoch': '0.3301'}
+{'loss': '1.102', 'grad_norm': '32.75', 'learning_rate': '6.699e-06', 'num_tokens': '1.244e+06', 'mean_token_accuracy': '0.7667', 'epoch': '0.3303'}
+{'loss': '1.153', 'grad_norm': '13.75', 'learning_rate': '6.697e-06', 'num_tokens': '1.245e+06', 'mean_token_accuracy': '0.6806', 'epoch': '0.3305'}
+{'loss': '1.872', 'grad_norm': '14.06', 'learning_rate': '6.695e-06', 'num_tokens': '1.246e+06', 'mean_token_accuracy': '0.5849', 'epoch': '0.3307'}
+{'loss': '1.55', 'grad_norm': '17.12', 'learning_rate': '6.693e-06', 'num_tokens': '1.246e+06', 'mean_token_accuracy': '0.6657', 'epoch': '0.3308'}
+{'loss': '1.216', 'grad_norm': '14.69', 'learning_rate': '6.692e-06', 'num_tokens': '1.247e+06', 'mean_token_accuracy': '0.7083', 'epoch': '0.331'}
+{'loss': '1.904', 'grad_norm': '21.12', 'learning_rate': '6.69e-06', 'num_tokens': '1.247e+06', 'mean_token_accuracy': '0.6044', 'epoch': '0.3312'}
+{'loss': '2.452', 'grad_norm': '7.594', 'learning_rate': '6.688e-06', 'num_tokens': '1.249e+06', 'mean_token_accuracy': '0.5476', 'epoch': '0.3314'}
+{'loss': '1.237', 'grad_norm': '17', 'learning_rate': '6.686e-06', 'num_tokens': '1.249e+06', 'mean_token_accuracy': '0.7175', 'epoch': '0.3315'}
+{'loss': '1.922', 'grad_norm': '13.25', 'learning_rate': '6.685e-06', 'num_tokens': '1.25e+06', 'mean_token_accuracy': '0.6253', 'epoch': '0.3317'}
+{'loss': '1.176', 'grad_norm': '13.44', 'learning_rate': '6.683e-06', 'num_tokens': '1.25e+06', 'mean_token_accuracy': '0.6682', 'epoch': '0.3319'}
+{'loss': '1.469', 'grad_norm': '13.44', 'learning_rate': '6.681e-06', 'num_tokens': '1.251e+06', 'mean_token_accuracy': '0.6051', 'epoch': '0.3321'}
+{'loss': '1.119', 'grad_norm': '13.75', 'learning_rate': '6.679e-06', 'num_tokens': '1.251e+06', 'mean_token_accuracy': '0.692', 'epoch': '0.3322'}
+{'loss': '0.8309', 'grad_norm': '12.88', 'learning_rate': '6.678e-06', 'num_tokens': '1.252e+06', 'mean_token_accuracy': '0.7315', 'epoch': '0.3324'}
+{'loss': '0.713', 'grad_norm': '30.38', 'learning_rate': '6.676e-06', 'num_tokens': '1.252e+06', 'mean_token_accuracy': '0.8271', 'epoch': '0.3326'}
+{'loss': '2.397', 'grad_norm': '14.81', 'learning_rate': '6.674e-06', 'num_tokens': '1.253e+06', 'mean_token_accuracy': '0.5109', 'epoch': '0.3328'}
+{'loss': '1.957', 'grad_norm': '17.88', 'learning_rate': '6.672e-06', 'num_tokens': '1.253e+06', 'mean_token_accuracy': '0.5697', 'epoch': '0.3329'}
+{'loss': '1.614', 'grad_norm': '14.56', 'learning_rate': '6.671e-06', 'num_tokens': '1.254e+06', 'mean_token_accuracy': '0.5962', 'epoch': '0.3331'}
+{'loss': '1.464', 'grad_norm': '12.88', 'learning_rate': '6.669e-06', 'num_tokens': '1.254e+06', 'mean_token_accuracy': '0.6157', 'epoch': '0.3333'}
+{'loss': '1.153', 'grad_norm': '11.44', 'learning_rate': '6.667e-06', 'num_tokens': '1.255e+06', 'mean_token_accuracy': '0.6543', 'epoch': '0.3334'}
+{'loss': '2.032', 'grad_norm': '13', 'learning_rate': '6.666e-06', 'num_tokens': '1.256e+06', 'mean_token_accuracy': '0.5725', 'epoch': '0.3336'}
+{'loss': '0.9432', 'grad_norm': '9.625', 'learning_rate': '6.664e-06', 'num_tokens': '1.257e+06', 'mean_token_accuracy': '0.7651', 'epoch': '0.3338'}
+{'loss': '0.649', 'grad_norm': '9.562', 'learning_rate': '6.662e-06', 'num_tokens': '1.258e+06', 'mean_token_accuracy': '0.8079', 'epoch': '0.334'}
+{'loss': '0.8073', 'grad_norm': '17', 'learning_rate': '6.66e-06', 'num_tokens': '1.258e+06', 'mean_token_accuracy': '0.7715', 'epoch': '0.3341'}
+{'loss': '1.585', 'grad_norm': '11.25', 'learning_rate': '6.659e-06', 'num_tokens': '1.259e+06', 'mean_token_accuracy': '0.629', 'epoch': '0.3343'}
+{'loss': '1.265', 'grad_norm': '13.38', 'learning_rate': '6.657e-06', 'num_tokens': '1.26e+06', 'mean_token_accuracy': '0.6485', 'epoch': '0.3345'}
+{'loss': '1.097', 'grad_norm': '13.38', 'learning_rate': '6.655e-06', 'num_tokens': '1.26e+06', 'mean_token_accuracy': '0.7505', 'epoch': '0.3347'}
+{'loss': '1.778', 'grad_norm': '16.25', 'learning_rate': '6.653e-06', 'num_tokens': '1.261e+06', 'mean_token_accuracy': '0.6352', 'epoch': '0.3348'}
+{'loss': '1.079', 'grad_norm': '11.06', 'learning_rate': '6.652e-06', 'num_tokens': '1.262e+06', 'mean_token_accuracy': '0.6835', 'epoch': '0.335'}
+{'loss': '1.092', 'grad_norm': '13.56', 'learning_rate': '6.65e-06', 'num_tokens': '1.262e+06', 'mean_token_accuracy': '0.6982', 'epoch': '0.3352'}
+{'loss': '1.713', 'grad_norm': '14.44', 'learning_rate': '6.648e-06', 'num_tokens': '1.263e+06', 'mean_token_accuracy': '0.6584', 'epoch': '0.3354'}
+{'loss': '1.414', 'grad_norm': '16.12', 'learning_rate': '6.646e-06', 'num_tokens': '1.263e+06', 'mean_token_accuracy': '0.6615', 'epoch': '0.3355'}
+{'loss': '1.145', 'grad_norm': '12.56', 'learning_rate': '6.645e-06', 'num_tokens': '1.264e+06', 'mean_token_accuracy': '0.6644', 'epoch': '0.3357'}
+{'loss': '1.085', 'grad_norm': '12.19', 'learning_rate': '6.643e-06', 'num_tokens': '1.265e+06', 'mean_token_accuracy': '0.7093', 'epoch': '0.3359'}
+{'loss': '1.187', 'grad_norm': '14.56', 'learning_rate': '6.641e-06', 'num_tokens': '1.265e+06', 'mean_token_accuracy': '0.6709', 'epoch': '0.3361'}
+{'loss': '1.721', 'grad_norm': '18.75', 'learning_rate': '6.639e-06', 'num_tokens': '1.266e+06', 'mean_token_accuracy': '0.6309', 'epoch': '0.3362'}
+{'loss': '0.7603', 'grad_norm': '11.75', 'learning_rate': '6.638e-06', 'num_tokens': '1.266e+06', 'mean_token_accuracy': '0.8233', 'epoch': '0.3364'}
+{'loss': '0.5293', 'grad_norm': '6.094', 'learning_rate': '6.636e-06', 'num_tokens': '1.268e+06', 'mean_token_accuracy': '0.8457', 'epoch': '0.3366'}
+{'loss': '1.011', 'grad_norm': '8.125', 'learning_rate': '6.634e-06', 'num_tokens': '1.269e+06', 'mean_token_accuracy': '0.7518', 'epoch': '0.3367'}
+{'loss': '0.8263', 'grad_norm': '20.88', 'learning_rate': '6.633e-06', 'num_tokens': '1.269e+06', 'mean_token_accuracy': '0.7928', 'epoch': '0.3369'}
+
 34%|███▎      | 1941/5758 [11:02<22:58,  2.77it/s]
 34%|███▎      | 1942/5758 [11:03<22:07,  2.88it/s]
                                                   

 34%|███▎      | 1942/5758 [11:03<22:07,  2.88it/s]
 34%|███▎      | 1943/5758 [11:03<22:10,  2.87it/s]
                                                   

 34%|███▎      | 1943/5758 [11:03<22:10,  2.87it/s]
 34%|███▍      | 1944/5758 [11:03<21:07,  3.01it/s]
                                                   

 34%|███▍      | 1944/5758 [11:03<21:07,  3.01it/s]
 34%|███▍      | 1945/5758 [11:04<20:35,  3.09it/s]
                                                   

 34%|███▍      | 1945/5758 [11:04<20:35,  3.09it/s]
 34%|███▍      | 1946/5758 [11:04<21:38,  2.94it/s]
                                                   

 34%|███▍      | 1946/5758 [11:04<21:38,  2.94it/s]
 34%|███▍      | 1947/5758 [11:04<21:16,  2.99it/s]
                                                   

 34%|███▍      | 1947/5758 [11:04<21:16,  2.99it/s]
 34%|███▍      | 1948/5758 [11:05<20:25,  3.11it/s]
                                                   

 34%|███▍      | 1948/5758 [11:05<20:25,  3.11it/s]
 34%|███▍      | 1949/5758 [11:05<21:30,  2.95it/s]
                                                   

 34%|███▍      | 1949/5758 [11:05<21:30,  2.95it/s]
 34%|███▍      | 1950/5758 [11:05<20:53,  3.04it/s]
                                                   

 34%|███▍      | 1950/5758 [11:05<20:53,  3.04it/s]
 34%|███▍      | 1951/5758 [11:06<21:43,  2.92it/s]
                                                   

 34%|███▍      | 1951/5758 [11:06<21:43,  2.92it/s]
 34%|███▍      | 1952/5758 [11:06<20:41,  3.07it/s]
                                                   

 34%|███▍      | 1952/5758 [11:06<20:41,  3.07it/s]
 34%|███▍      | 1953/5758 [11:06<21:01,  3.02it/s]
                                                   

 34%|███▍      | 1953/5758 [11:06<21:01,  3.02it/s]
 34%|███▍      | 1954/5758 [11:07<20:14,  3.13it/s]
                                                   

 34%|███▍      | 1954/5758 [11:07<20:14,  3.13it/s]
 34%|███▍      | 1955/5758 [11:07<20:58,  3.02it/s]
                                                   

 34%|███▍      | 1955/5758 [11:07<20:58,  3.02it/s]
 34%|███▍      | 1956/5758 [11:07<21:32,  2.94it/s]
                                                   

 34%|███▍      | 1956/5758 [11:07<21:32,  2.94it/s]
 34%|███▍      | 1957/5758 [11:08<21:28,  2.95it/s]
                                                   

 34%|███▍      | 1957/5758 [11:08<21:28,  2.95it/s]
 34%|███▍      | 1958/5758 [11:08<20:34,  3.08it/s]
                                                   

 34%|███▍      | 1958/5758 [11:08<20:34,  3.08it/s]
 34%|███▍      | 1959/5758 [11:08<20:31,  3.08it/s]
                                                   

 34%|███▍      | 1959/5758 [11:08<20:31,  3.08it/s]
 34%|███▍      | 1960/5758 [11:09<20:43,  3.05it/s]
                                                   

 34%|███▍      | 1960/5758 [11:09<20:43,  3.05it/s]
 34%|███▍      | 1961/5758 [11:09<20:22,  3.11it/s]
                                                   

 34%|███▍      | 1961/5758 [11:09<20:22,  3.11it/s]
 34%|███▍      | 1962/5758 [11:09<20:20,  3.11it/s]
                                                   

 34%|███▍      | 1962/5758 [11:09<20:20,  3.11it/s]
 34%|███▍      | 1963/5758 [11:10<20:09,  3.14it/s]
                                                   

 34%|███▍      | 1963/5758 [11:10<20:09,  3.14it/s]
 34%|███▍      | 1964/5758 [11:10<20:28,  3.09it/s]
                                                   

 34%|███▍      | 1964/5758 [11:10<20:28,  3.09it/s]
 34%|███▍      | 1965/5758 [11:10<20:13,  3.12it/s]
                                                   

 34%|███▍      | 1965/5758 [11:10<20:13,  3.12it/s]
 34%|███▍      | 1966/5758 [11:11<22:39,  2.79it/s]
                                                   

 34%|███▍      | 1966/5758 [11:11<22:39,  2.79it/s]
 34%|███▍      | 1967/5758 [11:11<23:19,  2.71it/s]
                                                   

 34%|███▍      | 1967/5758 [11:11<23:19,  2.71it/s]
 34%|███▍      | 1968/5758 [11:11<22:25,  2.82it/s]
                                                   

 34%|███▍      | 1968/5758 [11:11<22:25,  2.82it/s]
 34%|███▍      | 1969/5758 [11:12<21:54,  2.88it/s]
                                                   

 34%|███▍      | 1969/5758 [11:12<21:54,  2.88it/s]
 34%|███▍      | 1970/5758 [11:12<22:02,  2.86it/s]
                                                   

 34%|███▍      | 1970/5758 [11:12<22:02,  2.86it/s]
 34%|███▍      | 1971/5758 [11:12<22:32,  2.80it/s]
                                                   

 34%|███▍      | 1971/5758 [11:12<22:32,  2.80it/s]
 34%|███▍      | 1972/5758 [11:13<22:07,  2.85it/s]
                                                   

 34%|███▍      | 1972/5758 [11:13<22:07,  2.85it/s]
 34%|███▍      | 1973/5758 [11:13<21:30,  2.93it/s]
                                                   

 34%|███▍      | 1973/5758 [11:13<21:30,  2.93it/s]
 34%|███▍      | 1974/5758 [11:13<21:34,  2.92it/s]
                                                   

 34%|███▍      | 1974/5758 [11:13<21:34,  2.92it/s]
 34%|███▍      | 1975/5758 [11:14<22:06,  2.85it/s]
                                                   

 34%|███▍      | 1975/5758 [11:14<22:06,  2.85it/s]
 34%|███▍      | 1976/5758 [11:14<21:44,  2.90it/s]
                                                   

 34%|███▍      | 1976/5758 [11:14<21:44,  2.90it/s]
 34%|███▍      | 1977/5758 [11:14<20:47,  3.03it/s]
                                                   

 34%|███▍      | 1977/5758 [11:14<20:47,  3.03it/s]
 34%|███▍      | 1978/5758 [11:15<21:48,  2.89it/s]
                                                   

 34%|███▍      | 1978/5758 [11:15<21:48,  2.89it/s]
 34%|███▍      | 1979/5758 [11:15<21:16,  2.96it/s]
                                                   

 34%|███▍      | 1979/5758 [11:15<21:16,  2.96it/s]
 34%|███▍      | 1980/5758 [11:15<20:37,  3.05it/s]
                                                   

 34%|███▍      | 1980/5758 [11:15<20:37,  3.05it/s]
 34%|███▍      | 1981/5758 [11:16<21:47,  2.89it/s]
                                                   

 34%|███▍      | 1981/5758 [11:16<21:47,  2.89it/s]
 34%|███▍      | 1982/5758 [11:16<22:44,  2.77it/s]
                                                   

 34%|███▍      | 1982/5758 [11:16<22:44,  2.77it/s]
 34%|███▍      | 1983/5758 [11:17<22:36,  2.78it/s]
                                                   

 34%|███▍      | 1983/5758 [11:17<22:36,  2.78it/s]
 34%|███▍      | 1984/5758 [11:17<21:30,  2.92it/s]
                                                   

 34%|███▍      | 1984/5758 [11:17<21:30,  2.92it/s]
 34%|███▍      | 1985/5758 [11:17<21:57,  2.86it/s]
                                                   

 34%|███▍      | 1985/5758 [11:17<21:57,  2.86it/s]
 34%|███▍      | 1986/5758 [11:18<22:08,  2.84it/s]
                                                   

 34%|███▍      | 1986/5758 [11:18<22:08,  2.84it/s]
 35%|███▍      | 1987/5758 [11:18<21:00,  2.99it/s]
                                                   

 35%|███▍      | 1987/5758 [11:18<21:00,  2.99it/s]
 35%|███▍      | 1988/5758 [11:18<21:00,  2.99it/s]
                                                   

 35%|███▍      | 1988/5758 [11:18<21:00,  2.99it/s]
 35%|███▍      | 1989/5758 [11:19<22:06,  2.84it/s]
                                                   

 35%|███▍      | 1989/5758 [11:19<22:06,  2.84it/s]
 35%|███▍      | 1990/5758 [11:19<21:15,  2.95it/s]
                                                   

 35%|███▍      | 1990/5758 [11:19<21:15,  2.95it/s]
 35%|███▍      | 1991/5758 [11:19<22:05,  2.84it/s]
                                                   

 35%|███▍      | 1991/5758 [11:19<22:05,  2.84it/s]
 35%|███▍      | 1992/5758 [11:20<23:58,  2.62it/s]
                                                   

 35%|███▍      | 1992/5758 [11:20<23:58,  2.62it/s]
 35%|███▍      | 1993/5758 [11:20<22:13,  2.82it/s]
                                                   

 35%|███▍      | 1993/5758 [11:20<22:13,  2.82it/s]
 35%|███▍      | 1994/5758 [11:20<23:07,  2.71it/s]
                                                   

 35%|███▍      | 1994/5758 [11:20<23:07,  2.71it/s]
 35%|███▍      | 1995/5758 [11:21<21:47,  2.88it/s]
                                                   

 35%|███▍      | 1995/5758 [11:21<21:47,  2.88it/s]
 35%|███▍      | 1996/5758 [11:21<22:18,  2.81it/s]
                                                   
{'loss': '0.6119', 'grad_norm': '6.75', 'learning_rate': '6.631e-06', 'num_tokens': '1.271e+06', 'mean_token_accuracy': '0.8249', 'epoch': '0.3371'}
+{'loss': '0.8748', 'grad_norm': '9.25', 'learning_rate': '6.629e-06', 'num_tokens': '1.272e+06', 'mean_token_accuracy': '0.7878', 'epoch': '0.3373'}
+{'loss': '1.726', 'grad_norm': '17.88', 'learning_rate': '6.627e-06', 'num_tokens': '1.272e+06', 'mean_token_accuracy': '0.6442', 'epoch': '0.3374'}
+{'loss': '1.331', 'grad_norm': '34.25', 'learning_rate': '6.626e-06', 'num_tokens': '1.273e+06', 'mean_token_accuracy': '0.7067', 'epoch': '0.3376'}
+{'loss': '1.038', 'grad_norm': '11.94', 'learning_rate': '6.624e-06', 'num_tokens': '1.273e+06', 'mean_token_accuracy': '0.7173', 'epoch': '0.3378'}
+{'loss': '0.6933', 'grad_norm': '24.38', 'learning_rate': '6.622e-06', 'num_tokens': '1.273e+06', 'mean_token_accuracy': '0.8615', 'epoch': '0.338'}
+{'loss': '2.036', 'grad_norm': '24.75', 'learning_rate': '6.62e-06', 'num_tokens': '1.274e+06', 'mean_token_accuracy': '0.5931', 'epoch': '0.3381'}
+{'loss': '1.579', 'grad_norm': '13.56', 'learning_rate': '6.619e-06', 'num_tokens': '1.274e+06', 'mean_token_accuracy': '0.601', 'epoch': '0.3383'}
+{'loss': '1.065', 'grad_norm': '14.12', 'learning_rate': '6.617e-06', 'num_tokens': '1.275e+06', 'mean_token_accuracy': '0.6826', 'epoch': '0.3385'}
+{'loss': '1.331', 'grad_norm': '12.69', 'learning_rate': '6.615e-06', 'num_tokens': '1.276e+06', 'mean_token_accuracy': '0.6546', 'epoch': '0.3387'}
+{'loss': '1.102', 'grad_norm': '10.06', 'learning_rate': '6.613e-06', 'num_tokens': '1.277e+06', 'mean_token_accuracy': '0.6782', 'epoch': '0.3388'}
+{'loss': '0.9436', 'grad_norm': '31.25', 'learning_rate': '6.612e-06', 'num_tokens': '1.277e+06', 'mean_token_accuracy': '0.7682', 'epoch': '0.339'}
+{'loss': '1.316', 'grad_norm': '19.62', 'learning_rate': '6.61e-06', 'num_tokens': '1.277e+06', 'mean_token_accuracy': '0.704', 'epoch': '0.3392'}
+{'loss': '1.71', 'grad_norm': '14.62', 'learning_rate': '6.608e-06', 'num_tokens': '1.278e+06', 'mean_token_accuracy': '0.6355', 'epoch': '0.3394'}
+{'loss': '1.195', 'grad_norm': '17.75', 'learning_rate': '6.606e-06', 'num_tokens': '1.278e+06', 'mean_token_accuracy': '0.7398', 'epoch': '0.3395'}
+{'loss': '1.136', 'grad_norm': '9.562', 'learning_rate': '6.605e-06', 'num_tokens': '1.279e+06', 'mean_token_accuracy': '0.7375', 'epoch': '0.3397'}
+{'loss': '0.6496', 'grad_norm': '6.812', 'learning_rate': '6.603e-06', 'num_tokens': '1.281e+06', 'mean_token_accuracy': '0.8117', 'epoch': '0.3399'}
+{'loss': '1.644', 'grad_norm': '12', 'learning_rate': '6.601e-06', 'num_tokens': '1.281e+06', 'mean_token_accuracy': '0.6624', 'epoch': '0.34'}
+{'loss': '2.074', 'grad_norm': '18', 'learning_rate': '6.6e-06', 'num_tokens': '1.282e+06', 'mean_token_accuracy': '0.5377', 'epoch': '0.3402'}
+{'loss': '1.304', 'grad_norm': '12.12', 'learning_rate': '6.598e-06', 'num_tokens': '1.282e+06', 'mean_token_accuracy': '0.65', 'epoch': '0.3404'}
+{'loss': '1.086', 'grad_norm': '30.62', 'learning_rate': '6.596e-06', 'num_tokens': '1.283e+06', 'mean_token_accuracy': '0.7135', 'epoch': '0.3406'}
+{'loss': '1.656', 'grad_norm': '15.81', 'learning_rate': '6.594e-06', 'num_tokens': '1.283e+06', 'mean_token_accuracy': '0.6361', 'epoch': '0.3407'}
+{'loss': '1.141', 'grad_norm': '12.44', 'learning_rate': '6.593e-06', 'num_tokens': '1.284e+06', 'mean_token_accuracy': '0.6813', 'epoch': '0.3409'}
+{'loss': '0.5498', 'grad_norm': '6.844', 'learning_rate': '6.591e-06', 'num_tokens': '1.285e+06', 'mean_token_accuracy': '0.838', 'epoch': '0.3411'}
+{'loss': '1.397', 'grad_norm': '12.12', 'learning_rate': '6.589e-06', 'num_tokens': '1.286e+06', 'mean_token_accuracy': '0.6313', 'epoch': '0.3413'}
+{'loss': '2.002', 'grad_norm': '16.25', 'learning_rate': '6.587e-06', 'num_tokens': '1.287e+06', 'mean_token_accuracy': '0.625', 'epoch': '0.3414'}
+{'loss': '1.937', 'grad_norm': '31', 'learning_rate': '6.586e-06', 'num_tokens': '1.287e+06', 'mean_token_accuracy': '0.6336', 'epoch': '0.3416'}
+{'loss': '1.775', 'grad_norm': '41.75', 'learning_rate': '6.584e-06', 'num_tokens': '1.287e+06', 'mean_token_accuracy': '0.6222', 'epoch': '0.3418'}
+{'loss': '0.6304', 'grad_norm': '6.75', 'learning_rate': '6.582e-06', 'num_tokens': '1.288e+06', 'mean_token_accuracy': '0.82', 'epoch': '0.342'}
+{'loss': '1.047', 'grad_norm': '33', 'learning_rate': '6.58e-06', 'num_tokens': '1.288e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.3421'}
+{'loss': '1.644', 'grad_norm': '19', 'learning_rate': '6.579e-06', 'num_tokens': '1.289e+06', 'mean_token_accuracy': '0.6231', 'epoch': '0.3423'}
+{'loss': '1.835', 'grad_norm': '23.88', 'learning_rate': '6.577e-06', 'num_tokens': '1.289e+06', 'mean_token_accuracy': '0.5981', 'epoch': '0.3425'}
+{'loss': '0.9364', 'grad_norm': '17.25', 'learning_rate': '6.575e-06', 'num_tokens': '1.289e+06', 'mean_token_accuracy': '0.7623', 'epoch': '0.3427'}
+{'loss': '2.064', 'grad_norm': '19.75', 'learning_rate': '6.573e-06', 'num_tokens': '1.29e+06', 'mean_token_accuracy': '0.6112', 'epoch': '0.3428'}
+{'loss': '1.765', 'grad_norm': '13.38', 'learning_rate': '6.572e-06', 'num_tokens': '1.29e+06', 'mean_token_accuracy': '0.6103', 'epoch': '0.343'}
+{'loss': '1.542', 'grad_norm': '20.12', 'learning_rate': '6.57e-06', 'num_tokens': '1.291e+06', 'mean_token_accuracy': '0.6792', 'epoch': '0.3432'}
+{'loss': '1.227', 'grad_norm': '15.12', 'learning_rate': '6.568e-06', 'num_tokens': '1.291e+06', 'mean_token_accuracy': '0.6474', 'epoch': '0.3433'}
+{'loss': '1.233', 'grad_norm': '13.62', 'learning_rate': '6.567e-06', 'num_tokens': '1.292e+06', 'mean_token_accuracy': '0.6713', 'epoch': '0.3435'}
+{'loss': '1.187', 'grad_norm': '12.44', 'learning_rate': '6.565e-06', 'num_tokens': '1.293e+06', 'mean_token_accuracy': '0.6757', 'epoch': '0.3437'}
+{'loss': '0.7947', 'grad_norm': '32.5', 'learning_rate': '6.563e-06', 'num_tokens': '1.293e+06', 'mean_token_accuracy': '0.8571', 'epoch': '0.3439'}
+{'loss': '2.527', 'grad_norm': '18.38', 'learning_rate': '6.561e-06', 'num_tokens': '1.293e+06', 'mean_token_accuracy': '0.5085', 'epoch': '0.344'}
+{'loss': '0.4186', 'grad_norm': '21.5', 'learning_rate': '6.56e-06', 'num_tokens': '1.293e+06', 'mean_token_accuracy': '0.8993', 'epoch': '0.3442'}
+{'loss': '1.398', 'grad_norm': '13.88', 'learning_rate': '6.558e-06', 'num_tokens': '1.294e+06', 'mean_token_accuracy': '0.6953', 'epoch': '0.3444'}
+{'loss': '0.9329', 'grad_norm': '15.75', 'learning_rate': '6.556e-06', 'num_tokens': '1.295e+06', 'mean_token_accuracy': '0.7464', 'epoch': '0.3446'}
+{'loss': '0.7027', 'grad_norm': '23', 'learning_rate': '6.554e-06', 'num_tokens': '1.295e+06', 'mean_token_accuracy': '0.8235', 'epoch': '0.3447'}
+{'loss': '2.036', 'grad_norm': '13.88', 'learning_rate': '6.553e-06', 'num_tokens': '1.295e+06', 'mean_token_accuracy': '0.581', 'epoch': '0.3449'}
+{'loss': '1.094', 'grad_norm': '13.25', 'learning_rate': '6.551e-06', 'num_tokens': '1.296e+06', 'mean_token_accuracy': '0.6879', 'epoch': '0.3451'}
+{'loss': '0.6664', 'grad_norm': '7.188', 'learning_rate': '6.549e-06', 'num_tokens': '1.298e+06', 'mean_token_accuracy': '0.8014', 'epoch': '0.3453'}
+{'loss': '1.304', 'grad_norm': '20.62', 'learning_rate': '6.547e-06', 'num_tokens': '1.298e+06', 'mean_token_accuracy': '0.6997', 'epoch': '0.3454'}
+{'loss': '2.055', 'grad_norm': '16.5', 'learning_rate': '6.546e-06', 'num_tokens': '1.299e+06', 'mean_token_accuracy': '0.584', 'epoch': '0.3456'}
+{'loss': '1.028', 'grad_norm': '18.38', 'learning_rate': '6.544e-06', 'num_tokens': '1.299e+06', 'mean_token_accuracy': '0.7923', 'epoch': '0.3458'}
+{'loss': '2.009', 'grad_norm': '24.88', 'learning_rate': '6.542e-06', 'num_tokens': '1.299e+06', 'mean_token_accuracy': '0.5613', 'epoch': '0.346'}
+{'loss': '1.403', 'grad_norm': '25', 'learning_rate': '6.54e-06', 'num_tokens': '1.299e+06', 'mean_token_accuracy': '0.7089', 'epoch': '0.3461'}
+{'loss': '1.147', 'grad_norm': '26.75', 'learning_rate': '6.539e-06', 'num_tokens': '1.299e+06', 'mean_token_accuracy': '0.7324', 'epoch': '0.3463'}
+{'loss': '1.961', 'grad_norm': '35.5', 'learning_rate': '6.537e-06', 'num_tokens': '1.3e+06', 'mean_token_accuracy': '0.622', 'epoch': '0.3465'}
+
 35%|███▍      | 1996/5758 [11:21<22:18,  2.81it/s]
 35%|███▍      | 1997/5758 [11:21<21:45,  2.88it/s]
                                                   

 35%|███▍      | 1997/5758 [11:21<21:45,  2.88it/s]
 35%|███▍      | 1998/5758 [11:22<22:30,  2.78it/s]
                                                   

 35%|███▍      | 1998/5758 [11:22<22:30,  2.78it/s]
 35%|███▍      | 1999/5758 [11:22<23:31,  2.66it/s]
                                                   

 35%|███▍      | 1999/5758 [11:22<23:31,  2.66it/s]
 35%|███▍      | 2000/5758 [11:23<22:04,  2.84it/s]
                                                   

 35%|███▍      | 2000/5758 [11:23<22:04,  2.84it/s]{'loss': '0.9571', 'grad_norm': '29.75', 'learning_rate': '6.535e-06', 'num_tokens': '1.3e+06', 'mean_token_accuracy': '0.7823', 'epoch': '0.3466'}
+{'loss': '1.38', 'grad_norm': '12.38', 'learning_rate': '6.534e-06', 'num_tokens': '1.301e+06', 'mean_token_accuracy': '0.6793', 'epoch': '0.3468'}
+{'loss': '1.424', 'grad_norm': '14.25', 'learning_rate': '6.532e-06', 'num_tokens': '1.301e+06', 'mean_token_accuracy': '0.6952', 'epoch': '0.347'}
+{'loss': '0.8802', 'grad_norm': '24.75', 'learning_rate': '6.53e-06', 'num_tokens': '1.301e+06', 'mean_token_accuracy': '0.7808', 'epoch': '0.3472'}
+{'loss': '1.869', 'grad_norm': '8.625', 'learning_rate': '6.528e-06', 'num_tokens': '1.303e+06', 'mean_token_accuracy': '0.6019', 'epoch': '0.3473'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.57it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.57it/s]
+
 35%|███▍      | 2001/5758 [11:25<55:20,  1.13it/s]
                                                   

 35%|███▍      | 2001/5758 [11:25<55:20,  1.13it/s]
 35%|███▍      | 2002/5758 [11:25<44:28,  1.41it/s]
                                                   

 35%|███▍      | 2002/5758 [11:25<44:28,  1.41it/s]
 35%|███▍      | 2003/5758 [11:25<37:03,  1.69it/s]
                                                   

 35%|███▍      | 2003/5758 [11:25<37:03,  1.69it/s]
 35%|███▍      | 2004/5758 [11:26<32:43,  1.91it/s]
                                                   

 35%|███▍      | 2004/5758 [11:26<32:43,  1.91it/s]
 35%|███▍      | 2005/5758 [11:26<29:21,  2.13it/s]
                                                   

 35%|███▍      | 2005/5758 [11:26<29:21,  2.13it/s]
 35%|███▍      | 2006/5758 [11:26<26:51,  2.33it/s]
                                                   

 35%|███▍      | 2006/5758 [11:26<26:51,  2.33it/s]
 35%|███▍      | 2007/5758 [11:27<26:10,  2.39it/s]
                                                   

 35%|███▍      | 2007/5758 [11:27<26:10,  2.39it/s]
 35%|███▍      | 2008/5758 [11:27<24:57,  2.50it/s]
                                                   

 35%|███▍      | 2008/5758 [11:27<24:57,  2.50it/s]
 35%|███▍      | 2009/5758 [11:27<23:58,  2.61it/s]
                                                   

 35%|███▍      | 2009/5758 [11:27<23:58,  2.61it/s]
 35%|███▍      | 2010/5758 [11:28<23:51,  2.62it/s]
                                                   

 35%|███▍      | 2010/5758 [11:28<23:51,  2.62it/s]
 35%|███▍      | 2011/5758 [11:28<22:29,  2.78it/s]
                                                   

 35%|███▍      | 2011/5758 [11:28<22:29,  2.78it/s]
 35%|███▍      | 2012/5758 [11:28<22:04,  2.83it/s]
                                                   

 35%|███▍      | 2012/5758 [11:28<22:04,  2.83it/s]
 35%|███▍      | 2013/5758 [11:29<21:55,  2.85it/s]
                                                   

 35%|███▍      | 2013/5758 [11:29<21:55,  2.85it/s]
 35%|███▍      | 2014/5758 [11:29<21:36,  2.89it/s]
                                                   

 35%|███▍      | 2014/5758 [11:29<21:36,  2.89it/s]
 35%|███▍      | 2015/5758 [11:29<20:35,  3.03it/s]
                                                   

 35%|███▍      | 2015/5758 [11:29<20:35,  3.03it/s]
 35%|███▌      | 2016/5758 [11:30<20:34,  3.03it/s]
                                                   

 35%|███▌      | 2016/5758 [11:30<20:34,  3.03it/s]
 35%|███▌      | 2017/5758 [11:30<21:06,  2.95it/s]
                                                   

 35%|███▌      | 2017/5758 [11:30<21:06,  2.95it/s]
 35%|███▌      | 2018/5758 [11:30<20:29,  3.04it/s]
                                                   

 35%|███▌      | 2018/5758 [11:30<20:29,  3.04it/s]
 35%|███▌      | 2019/5758 [11:31<23:29,  2.65it/s]
                                                   

 35%|███▌      | 2019/5758 [11:31<23:29,  2.65it/s]
 35%|███▌      | 2020/5758 [11:31<22:53,  2.72it/s]
                                                   

 35%|███▌      | 2020/5758 [11:31<22:53,  2.72it/s]
 35%|███▌      | 2021/5758 [11:32<22:28,  2.77it/s]
                                                   

 35%|███▌      | 2021/5758 [11:32<22:28,  2.77it/s]
 35%|███▌      | 2022/5758 [11:32<22:45,  2.74it/s]
                                                   

 35%|███▌      | 2022/5758 [11:32<22:45,  2.74it/s]
 35%|███▌      | 2023/5758 [11:32<21:25,  2.90it/s]
                                                   

 35%|███▌      | 2023/5758 [11:32<21:25,  2.90it/s]
 35%|███▌      | 2024/5758 [11:33<21:10,  2.94it/s]
                                                   

 35%|███▌      | 2024/5758 [11:33<21:10,  2.94it/s]
 35%|███▌      | 2025/5758 [11:33<22:05,  2.82it/s]
                                                   

 35%|███▌      | 2025/5758 [11:33<22:05,  2.82it/s]
 35%|███▌      | 2026/5758 [11:33<21:51,  2.85it/s]
                                                   

 35%|███▌      | 2026/5758 [11:33<21:51,  2.85it/s]
 35%|███▌      | 2027/5758 [11:34<22:13,  2.80it/s]
                                                   

 35%|███▌      | 2027/5758 [11:34<22:13,  2.80it/s]
 35%|███▌      | 2028/5758 [11:34<21:18,  2.92it/s]
                                                   

 35%|███▌      | 2028/5758 [11:34<21:18,  2.92it/s]
 35%|███���      | 2029/5758 [11:34<20:22,  3.05it/s]
                                                   

 35%|███▌      | 2029/5758 [11:34<20:22,  3.05it/s]
 35%|███▌      | 2030/5758 [11:35<20:14,  3.07it/s]
                                                   

 35%|███▌      | 2030/5758 [11:35<20:14,  3.07it/s]
 35%|███▌      | 2031/5758 [11:35<20:09,  3.08it/s]
                                                   

 35%|███▌      | 2031/5758 [11:35<20:09,  3.08it/s]
 35%|███▌      | 2032/5758 [11:35<20:28,  3.03it/s]
                                                   

 35%|███▌      | 2032/5758 [11:35<20:28,  3.03it/s]
 35%|███▌      | 2033/5758 [11:36<19:53,  3.12it/s]
                                                   

 35%|███▌      | 2033/5758 [11:36<19:53,  3.12it/s]
 35%|███▌      | 2034/5758 [11:36<20:41,  3.00it/s]
                                                   

 35%|███▌      | 2034/5758 [11:36<20:41,  3.00it/s]
 35%|███▌      | 2035/5758 [11:36<20:01,  3.10it/s]
                                                   

 35%|███▌      | 2035/5758 [11:36<20:01,  3.10it/s]
 35%|███▌      | 2036/5758 [11:37<20:47,  2.98it/s]
                                                   

 35%|███▌      | 2036/5758 [11:37<20:47,  2.98it/s]
 35%|███▌      | 2037/5758 [11:37<20:54,  2.97it/s]
                                                   

 35%|███▌      | 2037/5758 [11:37<20:54,  2.97it/s]
 35%|███▌      | 2038/5758 [11:37<20:28,  3.03it/s]
                                                   

 35%|███▌      | 2038/5758 [11:37<20:28,  3.03it/s]
 35%|███▌      | 2039/5758 [11:38<19:56,  3.11it/s]
                                                   

 35%|███▌      | 2039/5758 [11:38<19:56,  3.11it/s]
 35%|███▌      | 2040/5758 [11:38<20:32,  3.02it/s]
                                                   

 35%|███▌      | 2040/5758 [11:38<20:32,  3.02it/s]
 35%|███▌      | 2041/5758 [11:38<21:12,  2.92it/s]
                                                   

 35%|███▌      | 2041/5758 [11:38<21:12,  2.92it/s]
 35%|███▌      | 2042/5758 [11:39<20:29,  3.02it/s]
                                                   

 35%|███▌      | 2042/5758 [11:39<20:29,  3.02it/s]
 35%|███▌      | 2043/5758 [11:39<20:04,  3.09it/s]
                                                   

 35%|███▌      | 2043/5758 [11:39<20:04,  3.09it/s]
 35%|███▌      | 2044/5758 [11:39<19:50,  3.12it/s]
                                                   

 35%|███▌      | 2044/5758 [11:39<19:50,  3.12it/s]
 36%|███▌      | 2045/5758 [11:40<20:05,  3.08it/s]
                                                   

 36%|███▌      | 2045/5758 [11:40<20:05,  3.08it/s]
 36%|███▌      | 2046/5758 [11:40<19:57,  3.10it/s]
                                                   

 36%|███▌      | 2046/5758 [11:40<19:57,  3.10it/s]
 36%|███▌      | 2047/5758 [11:40<19:34,  3.16it/s]
                                                   

 36%|███▌      | 2047/5758 [11:40<19:34,  3.16it/s]
 36%|███▌      | 2048/5758 [11:41<21:32,  2.87it/s]
                                                   

 36%|███▌      | 2048/5758 [11:41<21:32,  2.87it/s]
 36%|███▌      | 2049/5758 [11:41<22:12,  2.78it/s]
                                                   

 36%|███▌      | 2049/5758 [11:41<22:12,  2.78it/s]
 36%|███▌      | 2050/5758 [11:41<22:12,  2.78it/s]
                                                   

 36%|███▌      | 2050/5758 [11:41<22:12,  2.78it/s]
 36%|███▌      | 2051/5758 [11:42<21:25,  2.88it/s]
                                                   

 36%|███▌      | 2051/5758 [11:42<21:25,  2.88it/s]
 36%|███▌      | 2052/5758 [11:42<21:11,  2.91it/s]
                                                   

 36%|███▌      | 2052/5758 [11:42<21:11,  2.91it/s]
 36%|███▌      | 2053/5758 [11:42<20:19,  3.04it/s]
                                                   

 36%|███▌      | 2053/5758 [11:42<20:19,  3.04it/s]
 36%|███▌      | 2054/5758 [11:43<19:51,  3.11it/s]
                                                   

 36%|███▌      | 2054/5758 [11:43<19:51,  3.11it/s]
 36%|███▌      | 2055/5758 [11:43<20:04,  3.08it/s]
                                                   

 36%|███▌      | 2055/5758 [11:43<20:04,  3.08it/s]
 36%|███▌      | 2056/5758 [11:43<19:25,  3.18it/s]
                                                   
{'loss': '0.8049', 'grad_norm': '14.19', 'learning_rate': '6.527e-06', 'num_tokens': '1.303e+06', 'mean_token_accuracy': '0.8123', 'epoch': '0.3475'}
+{'loss': '1.881', 'grad_norm': '13.06', 'learning_rate': '6.525e-06', 'num_tokens': '1.304e+06', 'mean_token_accuracy': '0.6527', 'epoch': '0.3477'}
+{'loss': '1.148', 'grad_norm': '12.19', 'learning_rate': '6.523e-06', 'num_tokens': '1.304e+06', 'mean_token_accuracy': '0.7462', 'epoch': '0.3479'}
+{'loss': '1.098', 'grad_norm': '5.094', 'learning_rate': '6.521e-06', 'num_tokens': '1.309e+06', 'mean_token_accuracy': '0.7267', 'epoch': '0.348'}
+{'loss': '1.174', 'grad_norm': '11.25', 'learning_rate': '6.52e-06', 'num_tokens': '1.309e+06', 'mean_token_accuracy': '0.6685', 'epoch': '0.3482'}
+{'loss': '1.036', 'grad_norm': '15.62', 'learning_rate': '6.518e-06', 'num_tokens': '1.31e+06', 'mean_token_accuracy': '0.7512', 'epoch': '0.3484'}
+{'loss': '1.844', 'grad_norm': '28.75', 'learning_rate': '6.516e-06', 'num_tokens': '1.31e+06', 'mean_token_accuracy': '0.6222', 'epoch': '0.3486'}
+{'loss': '1.048', 'grad_norm': '7.375', 'learning_rate': '6.514e-06', 'num_tokens': '1.312e+06', 'mean_token_accuracy': '0.7523', 'epoch': '0.3487'}
+{'loss': '1.963', 'grad_norm': '10', 'learning_rate': '6.513e-06', 'num_tokens': '1.313e+06', 'mean_token_accuracy': '0.5857', 'epoch': '0.3489'}
+{'loss': '2.074', 'grad_norm': '14.19', 'learning_rate': '6.511e-06', 'num_tokens': '1.313e+06', 'mean_token_accuracy': '0.581', 'epoch': '0.3491'}
+{'loss': '1.735', 'grad_norm': '10.81', 'learning_rate': '6.509e-06', 'num_tokens': '1.314e+06', 'mean_token_accuracy': '0.6023', 'epoch': '0.3493'}
+{'loss': '0.6822', 'grad_norm': '19.5', 'learning_rate': '6.507e-06', 'num_tokens': '1.314e+06', 'mean_token_accuracy': '0.8122', 'epoch': '0.3494'}
+{'loss': '1.653', 'grad_norm': '38.75', 'learning_rate': '6.506e-06', 'num_tokens': '1.314e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.3496'}
+{'loss': '1.456', 'grad_norm': '21.88', 'learning_rate': '6.504e-06', 'num_tokens': '1.315e+06', 'mean_token_accuracy': '0.713', 'epoch': '0.3498'}
+{'loss': '0.6673', 'grad_norm': '14', 'learning_rate': '6.502e-06', 'num_tokens': '1.315e+06', 'mean_token_accuracy': '0.8694', 'epoch': '0.3499'}
+{'loss': '0.801', 'grad_norm': '27.5', 'learning_rate': '6.501e-06', 'num_tokens': '1.315e+06', 'mean_token_accuracy': '0.7917', 'epoch': '0.3501'}
+{'loss': '1.159', 'grad_norm': '14.81', 'learning_rate': '6.499e-06', 'num_tokens': '1.315e+06', 'mean_token_accuracy': '0.6614', 'epoch': '0.3503'}
+{'loss': '0.5201', 'grad_norm': '6.75', 'learning_rate': '6.497e-06', 'num_tokens': '1.317e+06', 'mean_token_accuracy': '0.8402', 'epoch': '0.3505'}
+{'loss': '1.803', 'grad_norm': '17.5', 'learning_rate': '6.495e-06', 'num_tokens': '1.317e+06', 'mean_token_accuracy': '0.6386', 'epoch': '0.3506'}
+{'loss': '2.125', 'grad_norm': '19.38', 'learning_rate': '6.494e-06', 'num_tokens': '1.318e+06', 'mean_token_accuracy': '0.5724', 'epoch': '0.3508'}
+{'loss': '0.9068', 'grad_norm': '15', 'learning_rate': '6.492e-06', 'num_tokens': '1.318e+06', 'mean_token_accuracy': '0.7807', 'epoch': '0.351'}
+{'loss': '0.9885', 'grad_norm': '14.94', 'learning_rate': '6.49e-06', 'num_tokens': '1.319e+06', 'mean_token_accuracy': '0.7344', 'epoch': '0.3512'}
+{'loss': '1.139', 'grad_norm': '13.5', 'learning_rate': '6.488e-06', 'num_tokens': '1.319e+06', 'mean_token_accuracy': '0.6926', 'epoch': '0.3513'}
+{'loss': '0.958', 'grad_norm': '13.56', 'learning_rate': '6.487e-06', 'num_tokens': '1.32e+06', 'mean_token_accuracy': '0.7193', 'epoch': '0.3515'}
+{'loss': '1.283', 'grad_norm': '11.69', 'learning_rate': '6.485e-06', 'num_tokens': '1.321e+06', 'mean_token_accuracy': '0.6505', 'epoch': '0.3517'}
+{'loss': '0.7326', 'grad_norm': '8.062', 'learning_rate': '6.483e-06', 'num_tokens': '1.322e+06', 'mean_token_accuracy': '0.7834', 'epoch': '0.3519'}
+{'loss': '1.078', 'grad_norm': '13.56', 'learning_rate': '6.481e-06', 'num_tokens': '1.323e+06', 'mean_token_accuracy': '0.7448', 'epoch': '0.352'}
+{'loss': '1.438', 'grad_norm': '21.75', 'learning_rate': '6.48e-06', 'num_tokens': '1.323e+06', 'mean_token_accuracy': '0.6677', 'epoch': '0.3522'}
+{'loss': '1.229', 'grad_norm': '55', 'learning_rate': '6.478e-06', 'num_tokens': '1.323e+06', 'mean_token_accuracy': '0.8305', 'epoch': '0.3524'}
+{'loss': '1.329', 'grad_norm': '5.906', 'learning_rate': '6.476e-06', 'num_tokens': '1.326e+06', 'mean_token_accuracy': '0.6828', 'epoch': '0.3526'}
+{'loss': '1.669', 'grad_norm': '11.44', 'learning_rate': '6.474e-06', 'num_tokens': '1.327e+06', 'mean_token_accuracy': '0.6756', 'epoch': '0.3527'}
+{'loss': '1.793', 'grad_norm': '11.5', 'learning_rate': '6.473e-06', 'num_tokens': '1.328e+06', 'mean_token_accuracy': '0.6142', 'epoch': '0.3529'}
+{'loss': '1.082', 'grad_norm': '12.56', 'learning_rate': '6.471e-06', 'num_tokens': '1.329e+06', 'mean_token_accuracy': '0.7123', 'epoch': '0.3531'}
+{'loss': '0.9087', 'grad_norm': '14.88', 'learning_rate': '6.469e-06', 'num_tokens': '1.329e+06', 'mean_token_accuracy': '0.7096', 'epoch': '0.3532'}
+{'loss': '1.716', 'grad_norm': '9.25', 'learning_rate': '6.468e-06', 'num_tokens': '1.33e+06', 'mean_token_accuracy': '0.621', 'epoch': '0.3534'}
+{'loss': '1.761', 'grad_norm': '20.38', 'learning_rate': '6.466e-06', 'num_tokens': '1.331e+06', 'mean_token_accuracy': '0.6538', 'epoch': '0.3536'}
+{'loss': '1.63', 'grad_norm': '16.25', 'learning_rate': '6.464e-06', 'num_tokens': '1.331e+06', 'mean_token_accuracy': '0.6096', 'epoch': '0.3538'}
+{'loss': '1.216', 'grad_norm': '12.44', 'learning_rate': '6.462e-06', 'num_tokens': '1.332e+06', 'mean_token_accuracy': '0.6635', 'epoch': '0.3539'}
+{'loss': '1.308', 'grad_norm': '13.56', 'learning_rate': '6.461e-06', 'num_tokens': '1.333e+06', 'mean_token_accuracy': '0.6555', 'epoch': '0.3541'}
+{'loss': '1.932', 'grad_norm': '18.25', 'learning_rate': '6.459e-06', 'num_tokens': '1.333e+06', 'mean_token_accuracy': '0.5778', 'epoch': '0.3543'}
+{'loss': '2.123', 'grad_norm': '18.75', 'learning_rate': '6.457e-06', 'num_tokens': '1.334e+06', 'mean_token_accuracy': '0.5577', 'epoch': '0.3545'}
+{'loss': '1.114', 'grad_norm': '15.75', 'learning_rate': '6.455e-06', 'num_tokens': '1.334e+06', 'mean_token_accuracy': '0.7309', 'epoch': '0.3546'}
+{'loss': '1.292', 'grad_norm': '12.94', 'learning_rate': '6.454e-06', 'num_tokens': '1.335e+06', 'mean_token_accuracy': '0.6559', 'epoch': '0.3548'}
+{'loss': '1.351', 'grad_norm': '13.44', 'learning_rate': '6.452e-06', 'num_tokens': '1.336e+06', 'mean_token_accuracy': '0.6366', 'epoch': '0.355'}
+{'loss': '1.303', 'grad_norm': '10.94', 'learning_rate': '6.45e-06', 'num_tokens': '1.337e+06', 'mean_token_accuracy': '0.6478', 'epoch': '0.3552'}
+{'loss': '1.154', 'grad_norm': '21.88', 'learning_rate': '6.448e-06', 'num_tokens': '1.337e+06', 'mean_token_accuracy': '0.7252', 'epoch': '0.3553'}
+{'loss': '1.144', 'grad_norm': '13', 'learning_rate': '6.447e-06', 'num_tokens': '1.338e+06', 'mean_token_accuracy': '0.6948', 'epoch': '0.3555'}
+{'loss': '1.276', 'grad_norm': '13.56', 'learning_rate': '6.445e-06', 'num_tokens': '1.338e+06', 'mean_token_accuracy': '0.6466', 'epoch': '0.3557'}
+{'loss': '2.246', 'grad_norm': '14.81', 'learning_rate': '6.443e-06', 'num_tokens': '1.339e+06', 'mean_token_accuracy': '0.5658', 'epoch': '0.3559'}
+{'loss': '1.327', 'grad_norm': '22.62', 'learning_rate': '6.441e-06', 'num_tokens': '1.339e+06', 'mean_token_accuracy': '0.722', 'epoch': '0.356'}
+{'loss': '0.8632', 'grad_norm': '14.38', 'learning_rate': '6.44e-06', 'num_tokens': '1.34e+06', 'mean_token_accuracy': '0.811', 'epoch': '0.3562'}
+{'loss': '0.6558', 'grad_norm': '17.62', 'learning_rate': '6.438e-06', 'num_tokens': '1.34e+06', 'mean_token_accuracy': '0.8545', 'epoch': '0.3564'}
+{'loss': '1.111', 'grad_norm': '13.12', 'learning_rate': '6.436e-06', 'num_tokens': '1.34e+06', 'mean_token_accuracy': '0.6983', 'epoch': '0.3565'}
+{'loss': '2.009', 'grad_norm': '11.19', 'learning_rate': '6.435e-06', 'num_tokens': '1.341e+06', 'mean_token_accuracy': '0.5655', 'epoch': '0.3567'}
+{'loss': '1.858', 'grad_norm': '9.688', 'learning_rate': '6.433e-06', 'num_tokens': '1.342e+06', 'mean_token_accuracy': '0.6362', 'epoch': '0.3569'}
+
 36%|███▌      | 2056/5758 [11:43<19:25,  3.18it/s]
 36%|███▌      | 2057/5758 [11:44<19:54,  3.10it/s]
                                                   

 36%|███▌      | 2057/5758 [11:44<19:54,  3.10it/s]
 36%|███▌      | 2058/5758 [11:44<19:27,  3.17it/s]
                                                   

 36%|███▌      | 2058/5758 [11:44<19:27,  3.17it/s]
 36%|███▌      | 2059/5758 [11:44<19:12,  3.21it/s]
                                                   

 36%|███▌      | 2059/5758 [11:44<19:12,  3.21it/s]
 36%|███▌      | 2060/5758 [11:44<20:09,  3.06it/s]
                                                   

 36%|███▌      | 2060/5758 [11:44<20:09,  3.06it/s]
 36%|███▌      | 2061/5758 [11:45<20:20,  3.03it/s]
                                                   

 36%|███▌      | 2061/5758 [11:45<20:20,  3.03it/s]
 36%|███▌      | 2062/5758 [11:45<19:51,  3.10it/s]
                                                   

 36%|███▌      | 2062/5758 [11:45<19:51,  3.10it/s]
 36%|███▌      | 2063/5758 [11:45<20:26,  3.01it/s]
                                                   

 36%|███▌      | 2063/5758 [11:45<20:26,  3.01it/s]
 36%|███▌      | 2064/5758 [11:46<20:27,  3.01it/s]
                                                   

 36%|███▌      | 2064/5758 [11:46<20:27,  3.01it/s]
 36%|███▌      | 2065/5758 [11:46<20:30,  3.00it/s]
                                                   

 36%|███▌      | 2065/5758 [11:46<20:30,  3.00it/s]
 36%|███▌      | 2066/5758 [11:47<21:08,  2.91it/s]
                                                   

 36%|███▌      | 2066/5758 [11:47<21:08,  2.91it/s]
 36%|███▌      | 2067/5758 [11:47<20:19,  3.03it/s]
                                                   

 36%|███▌      | 2067/5758 [11:47<20:19,  3.03it/s]
 36%|███▌      | 2068/5758 [11:47<21:27,  2.87it/s]
                                                   

 36%|███▌      | 2068/5758 [11:47<21:27,  2.87it/s]
 36%|███▌      | 2069/5758 [11:48<20:54,  2.94it/s]
                                                   

 36%|███▌      | 2069/5758 [11:48<20:54,  2.94it/s]
 36%|███▌      | 2070/5758 [11:48<20:34,  2.99it/s]
                                                   

 36%|███▌      | 2070/5758 [11:48<20:34,  2.99it/s]
 36%|███▌      | 2071/5758 [11:48<20:20,  3.02it/s]
                                                   

 36%|███▌      | 2071/5758 [11:48<20:20,  3.02it/s]
 36%|███▌      | 2072/5758 [11:49<21:28,  2.86it/s]
                                                   

 36%|███▌      | 2072/5758 [11:49<21:28,  2.86it/s]
 36%|███▌      | 2073/5758 [11:49<22:33,  2.72it/s]
                                                   

 36%|███▌      | 2073/5758 [11:49<22:33,  2.72it/s]
 36%|███▌      | 2074/5758 [11:49<22:54,  2.68it/s]
                                                   

 36%|███▌      | 2074/5758 [11:49<22:54,  2.68it/s]
 36%|███▌      | 2075/5758 [11:50<21:34,  2.85it/s]
                                                   

 36%|███▌      | 2075/5758 [11:50<21:34,  2.85it/s]
 36%|███▌      | 2076/5758 [11:50<21:01,  2.92it/s]
                                                   

 36%|███▌      | 2076/5758 [11:50<21:01,  2.92it/s]
 36%|███▌      | 2077/5758 [11:50<20:45,  2.95it/s]
                                                   

 36%|███▌      | 2077/5758 [11:50<20:45,  2.95it/s]
 36%|███▌      | 2078/5758 [11:51<20:42,  2.96it/s]
                                                   

 36%|███▌      | 2078/5758 [11:51<20:42,  2.96it/s]
 36%|███▌      | 2079/5758 [11:51<20:58,  2.92it/s]
                                                   

 36%|███▌      | 2079/5758 [11:51<20:58,  2.92it/s]
 36%|███▌      | 2080/5758 [11:51<20:04,  3.05it/s]
                                                   

 36%|███▌      | 2080/5758 [11:51<20:04,  3.05it/s]
 36%|███▌      | 2081/5758 [11:52<19:45,  3.10it/s]
                                                   

 36%|███▌      | 2081/5758 [11:52<19:45,  3.10it/s]
 36%|███▌      | 2082/5758 [11:52<19:58,  3.07it/s]
                                                   

 36%|███▌      | 2082/5758 [11:52<19:58,  3.07it/s]
 36%|███▌      | 2083/5758 [11:52<20:03,  3.05it/s]
                                                   

 36%|███▌      | 2083/5758 [11:52<20:03,  3.05it/s]
 36%|███▌      | 2084/5758 [11:53<19:46,  3.10it/s]
                                                   

 36%|███▌      | 2084/5758 [11:53<19:46,  3.10it/s]
 36%|███▌      | 2085/5758 [11:53<19:27,  3.15it/s]
                                                   

 36%|███▌      | 2085/5758 [11:53<19:27,  3.15it/s]
 36%|███▌      | 2086/5758 [11:53<20:16,  3.02it/s]
                                                   

 36%|███▌      | 2086/5758 [11:53<20:16,  3.02it/s]
 36%|███▌      | 2087/5758 [11:54<19:41,  3.11it/s]
                                                   

 36%|███▌      | 2087/5758 [11:54<19:41,  3.11it/s]
 36%|███▋      | 2088/5758 [11:54<20:13,  3.02it/s]
                                                   

 36%|███▋      | 2088/5758 [11:54<20:13,  3.02it/s]
 36%|███▋      | 2089/5758 [11:54<21:02,  2.91it/s]
                                                   

 36%|███▋      | 2089/5758 [11:54<21:02,  2.91it/s]
 36%|███▋      | 2090/5758 [11:55<21:39,  2.82it/s]
                                                   

 36%|███▋      | 2090/5758 [11:55<21:39,  2.82it/s]
 36%|███▋      | 2091/5758 [11:55<20:59,  2.91it/s]
                                                   

 36%|███▋      | 2091/5758 [11:55<20:59,  2.91it/s]
 36%|███▋      | 2092/5758 [11:55<20:06,  3.04it/s]
                                                   

 36%|███▋      | 2092/5758 [11:55<20:06,  3.04it/s]
 36%|███▋      | 2093/5758 [11:56<20:23,  3.00it/s]
                                                   

 36%|███▋      | 2093/5758 [11:56<20:23,  3.00it/s]
 36%|███▋      | 2094/5758 [11:56<19:53,  3.07it/s]
                                                   

 36%|███▋      | 2094/5758 [11:56<19:53,  3.07it/s]
 36%|███▋      | 2095/5758 [11:56<20:37,  2.96it/s]
                                                   

 36%|███▋      | 2095/5758 [11:56<20:37,  2.96it/s]
 36%|███▋      | 2096/5758 [11:57<20:45,  2.94it/s]
                                                   

 36%|███▋      | 2096/5758 [11:57<20:45,  2.94it/s]
 36%|███▋      | 2097/5758 [11:57<19:51,  3.07it/s]
                                                   

 36%|███▋      | 2097/5758 [11:57<19:51,  3.07it/s]
 36%|███▋      | 2098/5758 [11:57<20:53,  2.92it/s]
                                                   

 36%|███▋      | 2098/5758 [11:57<20:53,  2.92it/s]
 36%|███▋      | 2099/5758 [11:58<22:48,  2.67it/s]
                                                   

 36%|███▋      | 2099/5758 [11:58<22:48,  2.67it/s]
 36%|███▋      | 2100/5758 [11:58<21:27,  2.84it/s]
                                                   

 36%|███▋      | 2100/5758 [11:58<21:27,  2.84it/s]
 36%|███▋      | 2101/5758 [11:58<20:55,  2.91it/s]
                                                   

 36%|███▋      | 2101/5758 [11:58<20:55,  2.91it/s]
 37%|███▋      | 2102/5758 [11:59<19:59,  3.05it/s]
                                                   

 37%|███▋      | 2102/5758 [11:59<19:59,  3.05it/s]
 37%|███▋      | 2103/5758 [11:59<19:52,  3.06it/s]
                                                   

 37%|███▋      | 2103/5758 [11:59<19:52,  3.06it/s]
 37%|███▋      | 2104/5758 [11:59<19:15,  3.16it/s]
                                                   

 37%|███▋      | 2104/5758 [11:59<19:15,  3.16it/s]
 37%|███▋      | 2105/5758 [12:00<19:06,  3.19it/s]
                                                   

 37%|███▋      | 2105/5758 [12:00<19:06,  3.19it/s]
 37%|███▋      | 2106/5758 [12:00<18:47,  3.24it/s]
                                                   

 37%|███▋      | 2106/5758 [12:00<18:47,  3.24it/s]
 37%|███▋      | 2107/5758 [12:00<19:49,  3.07it/s]
                                                   

 37%|███▋      | 2107/5758 [12:00<19:49,  3.07it/s]
 37%|███▋      | 2108/5758 [12:01<19:25,  3.13it/s]
                                                   

 37%|███▋      | 2108/5758 [12:01<19:25,  3.13it/s]
 37%|███▋      | 2109/5758 [12:01<20:31,  2.96it/s]
                                                   

 37%|███▋      | 2109/5758 [12:01<20:31,  2.96it/s]
 37%|███▋      | 2110/5758 [12:01<20:56,  2.90it/s]
                                                   

 37%|███▋      | 2110/5758 [12:01<20:56,  2.90it/s]
 37%|███▋      | 2111/5758 [12:02<21:35,  2.81it/s]
                                                   
{'loss': '1.868', 'grad_norm': '12.56', 'learning_rate': '6.431e-06', 'num_tokens': '1.343e+06', 'mean_token_accuracy': '0.5954', 'epoch': '0.3571'}
+{'loss': '1.264', 'grad_norm': '19', 'learning_rate': '6.429e-06', 'num_tokens': '1.343e+06', 'mean_token_accuracy': '0.7636', 'epoch': '0.3572'}
+{'loss': '1.116', 'grad_norm': '5.812', 'learning_rate': '6.428e-06', 'num_tokens': '1.345e+06', 'mean_token_accuracy': '0.7372', 'epoch': '0.3574'}
+{'loss': '1.147', 'grad_norm': '21.38', 'learning_rate': '6.426e-06', 'num_tokens': '1.345e+06', 'mean_token_accuracy': '0.7688', 'epoch': '0.3576'}
+{'loss': '0.4995', 'grad_norm': '16.62', 'learning_rate': '6.424e-06', 'num_tokens': '1.346e+06', 'mean_token_accuracy': '0.8876', 'epoch': '0.3578'}
+{'loss': '1.241', 'grad_norm': '11.88', 'learning_rate': '6.422e-06', 'num_tokens': '1.347e+06', 'mean_token_accuracy': '0.6477', 'epoch': '0.3579'}
+{'loss': '1.125', 'grad_norm': '12.44', 'learning_rate': '6.421e-06', 'num_tokens': '1.347e+06', 'mean_token_accuracy': '0.6752', 'epoch': '0.3581'}
+{'loss': '1.748', 'grad_norm': '17.75', 'learning_rate': '6.419e-06', 'num_tokens': '1.348e+06', 'mean_token_accuracy': '0.6202', 'epoch': '0.3583'}
+{'loss': '1.318', 'grad_norm': '35.5', 'learning_rate': '6.417e-06', 'num_tokens': '1.348e+06', 'mean_token_accuracy': '0.7053', 'epoch': '0.3585'}
+{'loss': '2.012', 'grad_norm': '16.75', 'learning_rate': '6.415e-06', 'num_tokens': '1.348e+06', 'mean_token_accuracy': '0.5829', 'epoch': '0.3586'}
+{'loss': '1.544', 'grad_norm': '15.69', 'learning_rate': '6.414e-06', 'num_tokens': '1.349e+06', 'mean_token_accuracy': '0.6602', 'epoch': '0.3588'}
+{'loss': '1.7', 'grad_norm': '29.75', 'learning_rate': '6.412e-06', 'num_tokens': '1.349e+06', 'mean_token_accuracy': '0.6766', 'epoch': '0.359'}
+{'loss': '0.7713', 'grad_norm': '22.62', 'learning_rate': '6.41e-06', 'num_tokens': '1.349e+06', 'mean_token_accuracy': '0.8346', 'epoch': '0.3592'}
+{'loss': '1.275', 'grad_norm': '25.38', 'learning_rate': '6.408e-06', 'num_tokens': '1.349e+06', 'mean_token_accuracy': '0.7435', 'epoch': '0.3593'}
+{'loss': '0.7634', 'grad_norm': '13.94', 'learning_rate': '6.407e-06', 'num_tokens': '1.35e+06', 'mean_token_accuracy': '0.8179', 'epoch': '0.3595'}
+{'loss': '1.39', 'grad_norm': '10.31', 'learning_rate': '6.405e-06', 'num_tokens': '1.351e+06', 'mean_token_accuracy': '0.7113', 'epoch': '0.3597'}
+{'loss': '0.7515', 'grad_norm': '22.75', 'learning_rate': '6.403e-06', 'num_tokens': '1.351e+06', 'mean_token_accuracy': '0.8731', 'epoch': '0.3598'}
+{'loss': '1.393', 'grad_norm': '18.12', 'learning_rate': '6.402e-06', 'num_tokens': '1.351e+06', 'mean_token_accuracy': '0.6903', 'epoch': '0.36'}
+{'loss': '1.597', 'grad_norm': '13.62', 'learning_rate': '6.4e-06', 'num_tokens': '1.352e+06', 'mean_token_accuracy': '0.6862', 'epoch': '0.3602'}
+{'loss': '1.462', 'grad_norm': '11.25', 'learning_rate': '6.398e-06', 'num_tokens': '1.352e+06', 'mean_token_accuracy': '0.6882', 'epoch': '0.3604'}
+{'loss': '0.5491', 'grad_norm': '41', 'learning_rate': '6.396e-06', 'num_tokens': '1.352e+06', 'mean_token_accuracy': '0.8295', 'epoch': '0.3605'}
+{'loss': '0.5474', 'grad_norm': '26.5', 'learning_rate': '6.395e-06', 'num_tokens': '1.352e+06', 'mean_token_accuracy': '0.9022', 'epoch': '0.3607'}
+{'loss': '0.6877', 'grad_norm': '7.062', 'learning_rate': '6.393e-06', 'num_tokens': '1.354e+06', 'mean_token_accuracy': '0.7934', 'epoch': '0.3609'}
+{'loss': '1.08', 'grad_norm': '20.12', 'learning_rate': '6.391e-06', 'num_tokens': '1.354e+06', 'mean_token_accuracy': '0.7436', 'epoch': '0.3611'}
+{'loss': '1.101', 'grad_norm': '8.812', 'learning_rate': '6.389e-06', 'num_tokens': '1.356e+06', 'mean_token_accuracy': '0.7508', 'epoch': '0.3612'}
+{'loss': '1.77', 'grad_norm': '39.5', 'learning_rate': '6.388e-06', 'num_tokens': '1.356e+06', 'mean_token_accuracy': '0.6588', 'epoch': '0.3614'}
+{'loss': '1.654', 'grad_norm': '39', 'learning_rate': '6.386e-06', 'num_tokens': '1.356e+06', 'mean_token_accuracy': '0.6979', 'epoch': '0.3616'}
+{'loss': '1.555', 'grad_norm': '14.38', 'learning_rate': '6.384e-06', 'num_tokens': '1.357e+06', 'mean_token_accuracy': '0.5852', 'epoch': '0.3618'}
+{'loss': '1.867', 'grad_norm': '7.969', 'learning_rate': '6.382e-06', 'num_tokens': '1.358e+06', 'mean_token_accuracy': '0.6165', 'epoch': '0.3619'}
+{'loss': '0.9393', 'grad_norm': '12.56', 'learning_rate': '6.381e-06', 'num_tokens': '1.359e+06', 'mean_token_accuracy': '0.7445', 'epoch': '0.3621'}
+{'loss': '1.389', 'grad_norm': '15.38', 'learning_rate': '6.379e-06', 'num_tokens': '1.359e+06', 'mean_token_accuracy': '0.7055', 'epoch': '0.3623'}
+{'loss': '0.9164', 'grad_norm': '12.31', 'learning_rate': '6.377e-06', 'num_tokens': '1.36e+06', 'mean_token_accuracy': '0.7753', 'epoch': '0.3625'}
+{'loss': '1.253', 'grad_norm': '10.44', 'learning_rate': '6.375e-06', 'num_tokens': '1.361e+06', 'mean_token_accuracy': '0.6507', 'epoch': '0.3626'}
+{'loss': '1.239', 'grad_norm': '14.25', 'learning_rate': '6.374e-06', 'num_tokens': '1.361e+06', 'mean_token_accuracy': '0.6746', 'epoch': '0.3628'}
+{'loss': '2.325', 'grad_norm': '18', 'learning_rate': '6.372e-06', 'num_tokens': '1.362e+06', 'mean_token_accuracy': '0.5589', 'epoch': '0.363'}
+{'loss': '1.177', 'grad_norm': '5.438', 'learning_rate': '6.37e-06', 'num_tokens': '1.365e+06', 'mean_token_accuracy': '0.7182', 'epoch': '0.3631'}
+{'loss': '1.705', 'grad_norm': '8.125', 'learning_rate': '6.369e-06', 'num_tokens': '1.366e+06', 'mean_token_accuracy': '0.6213', 'epoch': '0.3633'}
+{'loss': '0.736', 'grad_norm': '10.12', 'learning_rate': '6.367e-06', 'num_tokens': '1.367e+06', 'mean_token_accuracy': '0.7974', 'epoch': '0.3635'}
+{'loss': '1.48', 'grad_norm': '13.38', 'learning_rate': '6.365e-06', 'num_tokens': '1.368e+06', 'mean_token_accuracy': '0.6366', 'epoch': '0.3637'}
+{'loss': '2.116', 'grad_norm': '13.94', 'learning_rate': '6.363e-06', 'num_tokens': '1.369e+06', 'mean_token_accuracy': '0.5868', 'epoch': '0.3638'}
+{'loss': '2.092', 'grad_norm': '21.62', 'learning_rate': '6.362e-06', 'num_tokens': '1.369e+06', 'mean_token_accuracy': '0.5456', 'epoch': '0.364'}
+{'loss': '1.543', 'grad_norm': '8.375', 'learning_rate': '6.36e-06', 'num_tokens': '1.37e+06', 'mean_token_accuracy': '0.6612', 'epoch': '0.3642'}
+{'loss': '1.351', 'grad_norm': '11.19', 'learning_rate': '6.358e-06', 'num_tokens': '1.371e+06', 'mean_token_accuracy': '0.6976', 'epoch': '0.3644'}
+{'loss': '1.351', 'grad_norm': '11.81', 'learning_rate': '6.356e-06', 'num_tokens': '1.372e+06', 'mean_token_accuracy': '0.651', 'epoch': '0.3645'}
+{'loss': '1.034', 'grad_norm': '11.38', 'learning_rate': '6.355e-06', 'num_tokens': '1.373e+06', 'mean_token_accuracy': '0.7151', 'epoch': '0.3647'}
+{'loss': '1.838', 'grad_norm': '11.19', 'learning_rate': '6.353e-06', 'num_tokens': '1.373e+06', 'mean_token_accuracy': '0.6327', 'epoch': '0.3649'}
+{'loss': '1.213', 'grad_norm': '8.812', 'learning_rate': '6.351e-06', 'num_tokens': '1.375e+06', 'mean_token_accuracy': '0.728', 'epoch': '0.3651'}
+{'loss': '1.197', 'grad_norm': '7.438', 'learning_rate': '6.349e-06', 'num_tokens': '1.376e+06', 'mean_token_accuracy': '0.729', 'epoch': '0.3652'}
+{'loss': '1.277', 'grad_norm': '19.5', 'learning_rate': '6.348e-06', 'num_tokens': '1.377e+06', 'mean_token_accuracy': '0.6997', 'epoch': '0.3654'}
+{'loss': '1.082', 'grad_norm': '14.88', 'learning_rate': '6.346e-06', 'num_tokens': '1.377e+06', 'mean_token_accuracy': '0.7216', 'epoch': '0.3656'}
+{'loss': '1.027', 'grad_norm': '13.38', 'learning_rate': '6.344e-06', 'num_tokens': '1.378e+06', 'mean_token_accuracy': '0.6972', 'epoch': '0.3658'}
+{'loss': '1.566', 'grad_norm': '12.81', 'learning_rate': '6.342e-06', 'num_tokens': '1.378e+06', 'mean_token_accuracy': '0.6534', 'epoch': '0.3659'}
+{'loss': '0.5314', 'grad_norm': '6.656', 'learning_rate': '6.341e-06', 'num_tokens': '1.38e+06', 'mean_token_accuracy': '0.8343', 'epoch': '0.3661'}
+{'loss': '1.569', 'grad_norm': '18.12', 'learning_rate': '6.339e-06', 'num_tokens': '1.38e+06', 'mean_token_accuracy': '0.6436', 'epoch': '0.3663'}
+{'loss': '1.131', 'grad_norm': '9.375', 'learning_rate': '6.337e-06', 'num_tokens': '1.381e+06', 'mean_token_accuracy': '0.752', 'epoch': '0.3664'}
+
 37%|███▋      | 2111/5758 [12:02<21:35,  2.81it/s]
 37%|███▋      | 2112/5758 [12:02<21:25,  2.84it/s]
                                                   

 37%|███▋      | 2112/5758 [12:02<21:25,  2.84it/s]
 37%|███▋      | 2113/5758 [12:02<21:46,  2.79it/s]
                                                   

 37%|███▋      | 2113/5758 [12:02<21:46,  2.79it/s]
 37%|███▋      | 2114/5758 [12:03<22:14,  2.73it/s]
                                                   

 37%|███▋      | 2114/5758 [12:03<22:14,  2.73it/s]
 37%|███▋      | 2115/5758 [12:03<20:54,  2.90it/s]
                                                   

 37%|███▋      | 2115/5758 [12:03<20:54,  2.90it/s]
 37%|███▋      | 2116/5758 [12:03<20:41,  2.93it/s]
                                                   

 37%|███▋      | 2116/5758 [12:03<20:41,  2.93it/s]
 37%|███▋      | 2117/5758 [12:04<19:58,  3.04it/s]
                                                   

 37%|███▋      | 2117/5758 [12:04<19:58,  3.04it/s]
 37%|███▋      | 2118/5758 [12:04<20:12,  3.00it/s]
                                                   

 37%|███▋      | 2118/5758 [12:04<20:12,  3.00it/s]
 37%|███▋      | 2119/5758 [12:04<19:21,  3.13it/s]
                                                   

 37%|███▋      | 2119/5758 [12:04<19:21,  3.13it/s]
 37%|███▋      | 2120/5758 [12:05<18:57,  3.20it/s]
                                                   

 37%|███▋      | 2120/5758 [12:05<18:57,  3.20it/s]
 37%|███▋      | 2121/5758 [12:05<19:15,  3.15it/s]
                                                   

 37%|███▋      | 2121/5758 [12:05<19:15,  3.15it/s]
 37%|███▋      | 2122/5758 [12:05<19:47,  3.06it/s]
                                                   

 37%|███▋      | 2122/5758 [12:05<19:47,  3.06it/s]
 37%|███▋      | 2123/5758 [12:06<19:06,  3.17it/s]
                                                   

 37%|███▋      | 2123/5758 [12:06<19:06,  3.17it/s]
 37%|███▋      | 2124/5758 [12:06<19:09,  3.16it/s]
                                                   

 37%|███▋      | 2124/5758 [12:06<19:09,  3.16it/s]
 37%|███▋      | 2125/5758 [12:06<21:35,  2.80it/s]
                                                   

 37%|███▋      | 2125/5758 [12:06<21:35,  2.80it/s]
 37%|███▋      | 2126/5758 [12:07<21:29,  2.82it/s]
                                                   

 37%|███▋      | 2126/5758 [12:07<21:29,  2.82it/s]
 37%|███▋      | 2127/5758 [12:07<21:03,  2.87it/s]
                                                   

 37%|███▋      | 2127/5758 [12:07<21:03,  2.87it/s]
 37%|███▋      | 2128/5758 [12:07<20:08,  3.00it/s]
                                                   

 37%|███▋      | 2128/5758 [12:07<20:08,  3.00it/s]
 37%|███▋      | 2129/5758 [12:08<19:27,  3.11it/s]
                                                   

 37%|███▋      | 2129/5758 [12:08<19:27,  3.11it/s]
 37%|███▋      | 2130/5758 [12:08<19:39,  3.08it/s]
                                                   

 37%|███▋      | 2130/5758 [12:08<19:39,  3.08it/s]
 37%|███▋      | 2131/5758 [12:08<20:20,  2.97it/s]
                                                   

 37%|███▋      | 2131/5758 [12:08<20:20,  2.97it/s]
 37%|███▋      | 2132/5758 [12:09<19:36,  3.08it/s]
                                                   

 37%|███▋      | 2132/5758 [12:09<19:36,  3.08it/s]
 37%|███▋      | 2133/5758 [12:09<20:06,  3.00it/s]
                                                   

 37%|███▋      | 2133/5758 [12:09<20:06,  3.00it/s]
 37%|███▋      | 2134/5758 [12:09<20:19,  2.97it/s]
                                                   

 37%|███▋      | 2134/5758 [12:09<20:19,  2.97it/s]
 37%|███▋      | 2135/5758 [12:10<21:16,  2.84it/s]
                                                   

 37%|███▋      | 2135/5758 [12:10<21:16,  2.84it/s]
 37%|███▋      | 2136/5758 [12:10<20:49,  2.90it/s]
                                                   

 37%|███▋      | 2136/5758 [12:10<20:49,  2.90it/s]
 37%|███▋      | 2137/5758 [12:10<21:10,  2.85it/s]
                                                   

 37%|███▋      | 2137/5758 [12:10<21:10,  2.85it/s]
 37%|███▋      | 2138/5758 [12:11<20:00,  3.02it/s]
                                                   

 37%|███▋      | 2138/5758 [12:11<20:00,  3.02it/s]
 37%|███▋      | 2139/5758 [12:11<19:54,  3.03it/s]
                                                   

 37%|███▋      | 2139/5758 [12:11<19:54,  3.03it/s]
 37%|███▋      | 2140/5758 [12:11<21:02,  2.87it/s]
                                                   

 37%|███▋      | 2140/5758 [12:11<21:02,  2.87it/s]
 37%|███▋      | 2141/5758 [12:12<20:01,  3.01it/s]
                                                   

 37%|███▋      | 2141/5758 [12:12<20:01,  3.01it/s]
 37%|███▋      | 2142/5758 [12:12<19:29,  3.09it/s]
                                                   

 37%|███▋      | 2142/5758 [12:12<19:29,  3.09it/s]
 37%|███▋      | 2143/5758 [12:12<19:27,  3.10it/s]
                                                   

 37%|███▋      | 2143/5758 [12:12<19:27,  3.10it/s]
 37%|███▋      | 2144/5758 [12:13<19:16,  3.12it/s]
                                                   

 37%|███▋      | 2144/5758 [12:13<19:16,  3.12it/s]
 37%|███▋      | 2145/5758 [12:13<19:12,  3.13it/s]
                                                   

 37%|███▋      | 2145/5758 [12:13<19:12,  3.13it/s]
 37%|███▋      | 2146/5758 [12:13<19:07,  3.15it/s]
                                                   

 37%|███▋      | 2146/5758 [12:13<19:07,  3.15it/s]
 37%|███▋      | 2147/5758 [12:14<19:10,  3.14it/s]
                                                   

 37%|███▋      | 2147/5758 [12:14<19:10,  3.14it/s]
 37%|███▋      | 2148/5758 [12:14<18:45,  3.21it/s]
                                                   

 37%|███▋      | 2148/5758 [12:14<18:45,  3.21it/s]
 37%|███▋      | 2149/5758 [12:14<19:00,  3.16it/s]
                                                   

 37%|███▋      | 2149/5758 [12:14<19:00,  3.16it/s]
 37%|███▋      | 2150/5758 [12:15<18:41,  3.22it/s]
                                                   

 37%|███▋      | 2150/5758 [12:15<18:41,  3.22it/s]
 37%|███▋      | 2151/5758 [12:15<21:28,  2.80it/s]
                                                   

 37%|███▋      | 2151/5758 [12:15<21:28,  2.80it/s]
 37%|███▋      | 2152/5758 [12:15<20:31,  2.93it/s]
                                                   

 37%|███▋      | 2152/5758 [12:15<20:31,  2.93it/s]
 37%|███▋      | 2153/5758 [12:16<20:54,  2.87it/s]
                                                   

 37%|███▋      | 2153/5758 [12:16<20:54,  2.87it/s]
 37%|███▋      | 2154/5758 [12:16<20:43,  2.90it/s]
                                                   

 37%|███▋      | 2154/5758 [12:16<20:43,  2.90it/s]
 37%|███▋      | 2155/5758 [12:16<20:13,  2.97it/s]
                                                   

 37%|███▋      | 2155/5758 [12:16<20:13,  2.97it/s]
 37%|███▋      | 2156/5758 [12:17<20:19,  2.95it/s]
                                                   

 37%|███▋      | 2156/5758 [12:17<20:19,  2.95it/s]
 37%|███▋      | 2157/5758 [12:17<20:20,  2.95it/s]
                                                   

 37%|███▋      | 2157/5758 [12:17<20:20,  2.95it/s]
 37%|███▋      | 2158/5758 [12:17<20:57,  2.86it/s]
                                                   

 37%|███▋      | 2158/5758 [12:17<20:57,  2.86it/s]
 37%|███▋      | 2159/5758 [12:18<20:10,  2.97it/s]
                                                   

 37%|███▋      | 2159/5758 [12:18<20:10,  2.97it/s]
 38%|███▊      | 2160/5758 [12:18<20:08,  2.98it/s]
                                                   

 38%|███▊      | 2160/5758 [12:18<20:08,  2.98it/s]
 38%|███▊      | 2161/5758 [12:18<21:10,  2.83it/s]
                                                   

 38%|███▊      | 2161/5758 [12:18<21:10,  2.83it/s]
 38%|███▊      | 2162/5758 [12:19<21:27,  2.79it/s]
                                                   

 38%|███▊      | 2162/5758 [12:19<21:27,  2.79it/s]
 38%|███▊      | 2163/5758 [12:19<21:45,  2.75it/s]
                                                   

 38%|███▊      | 2163/5758 [12:19<21:45,  2.75it/s]
 38%|███▊      | 2164/5758 [12:20<22:05,  2.71it/s]
                                                   

 38%|███▊      | 2164/5758 [12:20<22:05,  2.71it/s]
 38%|███▊      | 2165/5758 [12:20<21:16,  2.82it/s]
                                                   

 38%|███▊      | 2165/5758 [12:20<21:16,  2.82it/s]
 38%|███▊      | 2166/5758 [12:20<20:30,  2.92it/s]
                                                   
{'loss': '1.994', 'grad_norm': '22.88', 'learning_rate': '6.336e-06', 'num_tokens': '1.381e+06', 'mean_token_accuracy': '0.6078', 'epoch': '0.3666'}
+{'loss': '0.5613', 'grad_norm': '5.875', 'learning_rate': '6.334e-06', 'num_tokens': '1.383e+06', 'mean_token_accuracy': '0.8344', 'epoch': '0.3668'}
+{'loss': '1.504', 'grad_norm': '11.06', 'learning_rate': '6.332e-06', 'num_tokens': '1.384e+06', 'mean_token_accuracy': '0.6799', 'epoch': '0.367'}
+{'loss': '0.8052', 'grad_norm': '32.25', 'learning_rate': '6.33e-06', 'num_tokens': '1.385e+06', 'mean_token_accuracy': '0.8333', 'epoch': '0.3671'}
+{'loss': '2.024', 'grad_norm': '16.38', 'learning_rate': '6.329e-06', 'num_tokens': '1.385e+06', 'mean_token_accuracy': '0.5895', 'epoch': '0.3673'}
+{'loss': '0.6449', 'grad_norm': '7.125', 'learning_rate': '6.327e-06', 'num_tokens': '1.387e+06', 'mean_token_accuracy': '0.812', 'epoch': '0.3675'}
+{'loss': '1.124', 'grad_norm': '12.88', 'learning_rate': '6.325e-06', 'num_tokens': '1.388e+06', 'mean_token_accuracy': '0.6952', 'epoch': '0.3677'}
+{'loss': '0.9152', 'grad_norm': '17.62', 'learning_rate': '6.323e-06', 'num_tokens': '1.388e+06', 'mean_token_accuracy': '0.7787', 'epoch': '0.3678'}
+{'loss': '0.9222', 'grad_norm': '18.12', 'learning_rate': '6.322e-06', 'num_tokens': '1.388e+06', 'mean_token_accuracy': '0.8247', 'epoch': '0.368'}
+{'loss': '0.8301', 'grad_norm': '12.5', 'learning_rate': '6.32e-06', 'num_tokens': '1.389e+06', 'mean_token_accuracy': '0.7493', 'epoch': '0.3682'}
+{'loss': '1.215', 'grad_norm': '11.31', 'learning_rate': '6.318e-06', 'num_tokens': '1.39e+06', 'mean_token_accuracy': '0.6636', 'epoch': '0.3684'}
+{'loss': '1.259', 'grad_norm': '10.75', 'learning_rate': '6.316e-06', 'num_tokens': '1.391e+06', 'mean_token_accuracy': '0.6519', 'epoch': '0.3685'}
+{'loss': '1.015', 'grad_norm': '13.38', 'learning_rate': '6.315e-06', 'num_tokens': '1.391e+06', 'mean_token_accuracy': '0.6971', 'epoch': '0.3687'}
+{'loss': '1.241', 'grad_norm': '11.31', 'learning_rate': '6.313e-06', 'num_tokens': '1.392e+06', 'mean_token_accuracy': '0.6728', 'epoch': '0.3689'}
+{'loss': '0.5513', 'grad_norm': '28.75', 'learning_rate': '6.311e-06', 'num_tokens': '1.392e+06', 'mean_token_accuracy': '0.8614', 'epoch': '0.3691'}
+{'loss': '0.5766', 'grad_norm': '5.906', 'learning_rate': '6.309e-06', 'num_tokens': '1.394e+06', 'mean_token_accuracy': '0.8396', 'epoch': '0.3692'}
+{'loss': '0.5966', 'grad_norm': '6.812', 'learning_rate': '6.308e-06', 'num_tokens': '1.396e+06', 'mean_token_accuracy': '0.8232', 'epoch': '0.3694'}
+{'loss': '0.6998', 'grad_norm': '14.88', 'learning_rate': '6.306e-06', 'num_tokens': '1.396e+06', 'mean_token_accuracy': '0.8435', 'epoch': '0.3696'}
+{'loss': '1.821', 'grad_norm': '19.5', 'learning_rate': '6.304e-06', 'num_tokens': '1.396e+06', 'mean_token_accuracy': '0.6372', 'epoch': '0.3697'}
+{'loss': '1.789', 'grad_norm': '9.938', 'learning_rate': '6.303e-06', 'num_tokens': '1.397e+06', 'mean_token_accuracy': '0.6265', 'epoch': '0.3699'}
+{'loss': '1.451', 'grad_norm': '13.38', 'learning_rate': '6.301e-06', 'num_tokens': '1.398e+06', 'mean_token_accuracy': '0.6654', 'epoch': '0.3701'}
+{'loss': '1.222', 'grad_norm': '26', 'learning_rate': '6.299e-06', 'num_tokens': '1.398e+06', 'mean_token_accuracy': '0.7055', 'epoch': '0.3703'}
+{'loss': '1.241', 'grad_norm': '11.25', 'learning_rate': '6.297e-06', 'num_tokens': '1.399e+06', 'mean_token_accuracy': '0.7168', 'epoch': '0.3704'}
+{'loss': '1.008', 'grad_norm': '6.812', 'learning_rate': '6.296e-06', 'num_tokens': '1.401e+06', 'mean_token_accuracy': '0.7302', 'epoch': '0.3706'}
+{'loss': '1.702', 'grad_norm': '6.594', 'learning_rate': '6.294e-06', 'num_tokens': '1.403e+06', 'mean_token_accuracy': '0.6336', 'epoch': '0.3708'}
+{'loss': '0.7603', 'grad_norm': '27.12', 'learning_rate': '6.292e-06', 'num_tokens': '1.403e+06', 'mean_token_accuracy': '0.8021', 'epoch': '0.371'}
+{'loss': '1.124', 'grad_norm': '31.38', 'learning_rate': '6.29e-06', 'num_tokens': '1.404e+06', 'mean_token_accuracy': '0.7355', 'epoch': '0.3711'}
+{'loss': '1.897', 'grad_norm': '20', 'learning_rate': '6.289e-06', 'num_tokens': '1.404e+06', 'mean_token_accuracy': '0.6316', 'epoch': '0.3713'}
+{'loss': '0.5804', 'grad_norm': '7.781', 'learning_rate': '6.287e-06', 'num_tokens': '1.405e+06', 'mean_token_accuracy': '0.8252', 'epoch': '0.3715'}
+{'loss': '1.082', 'grad_norm': '28.5', 'learning_rate': '6.285e-06', 'num_tokens': '1.406e+06', 'mean_token_accuracy': '0.7319', 'epoch': '0.3717'}
+{'loss': '1.118', 'grad_norm': '34.25', 'learning_rate': '6.283e-06', 'num_tokens': '1.406e+06', 'mean_token_accuracy': '0.7867', 'epoch': '0.3718'}
+{'loss': '1.795', 'grad_norm': '20.38', 'learning_rate': '6.282e-06', 'num_tokens': '1.406e+06', 'mean_token_accuracy': '0.6385', 'epoch': '0.372'}
+{'loss': '1.883', 'grad_norm': '13.31', 'learning_rate': '6.28e-06', 'num_tokens': '1.407e+06', 'mean_token_accuracy': '0.5916', 'epoch': '0.3722'}
+{'loss': '1.847', 'grad_norm': '31.5', 'learning_rate': '6.278e-06', 'num_tokens': '1.407e+06', 'mean_token_accuracy': '0.6218', 'epoch': '0.3724'}
+{'loss': '1.136', 'grad_norm': '11.56', 'learning_rate': '6.276e-06', 'num_tokens': '1.408e+06', 'mean_token_accuracy': '0.6744', 'epoch': '0.3725'}
+{'loss': '2.084', 'grad_norm': '14.56', 'learning_rate': '6.275e-06', 'num_tokens': '1.409e+06', 'mean_token_accuracy': '0.5787', 'epoch': '0.3727'}
+{'loss': '1.347', 'grad_norm': '13.12', 'learning_rate': '6.273e-06', 'num_tokens': '1.41e+06', 'mean_token_accuracy': '0.68', 'epoch': '0.3729'}
+{'loss': '0.9236', 'grad_norm': '11.5', 'learning_rate': '6.271e-06', 'num_tokens': '1.41e+06', 'mean_token_accuracy': '0.7819', 'epoch': '0.373'}
+{'loss': '1.605', 'grad_norm': '10.69', 'learning_rate': '6.27e-06', 'num_tokens': '1.411e+06', 'mean_token_accuracy': '0.6478', 'epoch': '0.3732'}
+{'loss': '0.9292', 'grad_norm': '12.56', 'learning_rate': '6.268e-06', 'num_tokens': '1.412e+06', 'mean_token_accuracy': '0.7489', 'epoch': '0.3734'}
+{'loss': '1.436', 'grad_norm': '13.56', 'learning_rate': '6.266e-06', 'num_tokens': '1.412e+06', 'mean_token_accuracy': '0.6807', 'epoch': '0.3736'}
+{'loss': '1.268', 'grad_norm': '16.25', 'learning_rate': '6.264e-06', 'num_tokens': '1.413e+06', 'mean_token_accuracy': '0.7147', 'epoch': '0.3737'}
+{'loss': '1.37', 'grad_norm': '33.25', 'learning_rate': '6.263e-06', 'num_tokens': '1.413e+06', 'mean_token_accuracy': '0.7807', 'epoch': '0.3739'}
+{'loss': '2.047', 'grad_norm': '13.31', 'learning_rate': '6.261e-06', 'num_tokens': '1.413e+06', 'mean_token_accuracy': '0.5765', 'epoch': '0.3741'}
+{'loss': '1.33', 'grad_norm': '12.25', 'learning_rate': '6.259e-06', 'num_tokens': '1.414e+06', 'mean_token_accuracy': '0.6327', 'epoch': '0.3743'}
+{'loss': '1.145', 'grad_norm': '10.75', 'learning_rate': '6.257e-06', 'num_tokens': '1.415e+06', 'mean_token_accuracy': '0.6871', 'epoch': '0.3744'}
+{'loss': '2.022', 'grad_norm': '12.5', 'learning_rate': '6.256e-06', 'num_tokens': '1.416e+06', 'mean_token_accuracy': '0.5588', 'epoch': '0.3746'}
+{'loss': '2.094', 'grad_norm': '14.06', 'learning_rate': '6.254e-06', 'num_tokens': '1.416e+06', 'mean_token_accuracy': '0.5641', 'epoch': '0.3748'}
+{'loss': '1.513', 'grad_norm': '15.31', 'learning_rate': '6.252e-06', 'num_tokens': '1.417e+06', 'mean_token_accuracy': '0.6771', 'epoch': '0.375'}
+{'loss': '1.284', 'grad_norm': '7.469', 'learning_rate': '6.25e-06', 'num_tokens': '1.419e+06', 'mean_token_accuracy': '0.6993', 'epoch': '0.3751'}
+{'loss': '2.186', 'grad_norm': '19', 'learning_rate': '6.249e-06', 'num_tokens': '1.419e+06', 'mean_token_accuracy': '0.5441', 'epoch': '0.3753'}
+{'loss': '2.029', 'grad_norm': '19.25', 'learning_rate': '6.247e-06', 'num_tokens': '1.42e+06', 'mean_token_accuracy': '0.6078', 'epoch': '0.3755'}
+{'loss': '1.778', 'grad_norm': '15.81', 'learning_rate': '6.245e-06', 'num_tokens': '1.42e+06', 'mean_token_accuracy': '0.6123', 'epoch': '0.3757'}
+{'loss': '1.033', 'grad_norm': '18.25', 'learning_rate': '6.243e-06', 'num_tokens': '1.421e+06', 'mean_token_accuracy': '0.7587', 'epoch': '0.3758'}
+{'loss': '1.327', 'grad_norm': '21', 'learning_rate': '6.242e-06', 'num_tokens': '1.421e+06', 'mean_token_accuracy': '0.7427', 'epoch': '0.376'}
+
 38%|███▊      | 2166/5758 [12:20<20:30,  2.92it/s]
 38%|███▊      | 2167/5758 [12:21<20:42,  2.89it/s]
                                                   

 38%|███▊      | 2167/5758 [12:21<20:42,  2.89it/s]
 38%|███▊      | 2168/5758 [12:21<19:38,  3.05it/s]
                                                   

 38%|███▊      | 2168/5758 [12:21<19:38,  3.05it/s]
 38%|███▊      | 2169/5758 [12:21<18:57,  3.15it/s]
                                                   

 38%|███▊      | 2169/5758 [12:21<18:57,  3.15it/s]
 38%|███▊      | 2170/5758 [12:21<18:38,  3.21it/s]
                                                   

 38%|███▊      | 2170/5758 [12:21<18:38,  3.21it/s]
 38%|███▊      | 2171/5758 [12:22<18:59,  3.15it/s]
                                                   

 38%|███▊      | 2171/5758 [12:22<18:59,  3.15it/s]
 38%|███▊      | 2172/5758 [12:22<18:52,  3.17it/s]
                                                   

 38%|███▊      | 2172/5758 [12:22<18:52,  3.17it/s]
 38%|███▊      | 2173/5758 [12:22<19:03,  3.13it/s]
                                                   

 38%|███▊      | 2173/5758 [12:22<19:03,  3.13it/s]
 38%|███▊      | 2174/5758 [12:23<19:40,  3.04it/s]
                                                   

 38%|███▊      | 2174/5758 [12:23<19:40,  3.04it/s]
 38%|███▊      | 2175/5758 [12:23<19:29,  3.06it/s]
                                                   

 38%|███▊      | 2175/5758 [12:23<19:29,  3.06it/s]
 38%|███▊      | 2176/5758 [12:23<19:00,  3.14it/s]
                                                   

 38%|███▊      | 2176/5758 [12:23<19:00,  3.14it/s]
 38%|███▊      | 2177/5758 [12:24<21:19,  2.80it/s]
                                                   

 38%|███▊      | 2177/5758 [12:24<21:19,  2.80it/s]
 38%|███▊      | 2178/5758 [12:24<21:47,  2.74it/s]
                                                   

 38%|███▊      | 2178/5758 [12:24<21:47,  2.74it/s]
 38%|███▊      | 2179/5758 [12:25<21:55,  2.72it/s]
                                                   

 38%|███▊      | 2179/5758 [12:25<21:55,  2.72it/s]
 38%|███▊      | 2180/5758 [12:25<21:59,  2.71it/s]
                                                   

 38%|███▊      | 2180/5758 [12:25<21:59,  2.71it/s]
 38%|███▊      | 2181/5758 [12:25<20:37,  2.89it/s]
                                                   

 38%|███▊      | 2181/5758 [12:25<20:37,  2.89it/s]
 38%|███▊      | 2182/5758 [12:26<21:11,  2.81it/s]
                                                   

 38%|███▊      | 2182/5758 [12:26<21:11,  2.81it/s]
 38%|███▊      | 2183/5758 [12:26<20:22,  2.92it/s]
                                                   

 38%|███▊      | 2183/5758 [12:26<20:22,  2.92it/s]
 38%|███▊      | 2184/5758 [12:26<19:37,  3.03it/s]
                                                   

 38%|███▊      | 2184/5758 [12:26<19:37,  3.03it/s]
 38%|███▊      | 2185/5758 [12:27<19:17,  3.09it/s]
                                                   

 38%|███▊      | 2185/5758 [12:27<19:17,  3.09it/s]
 38%|███▊      | 2186/5758 [12:27<19:09,  3.11it/s]
                                                   

 38%|███▊      | 2186/5758 [12:27<19:09,  3.11it/s]
 38%|███▊      | 2187/5758 [12:27<20:13,  2.94it/s]
                                                   

 38%|███▊      | 2187/5758 [12:27<20:13,  2.94it/s]
 38%|███▊      | 2188/5758 [12:28<20:20,  2.93it/s]
                                                   

 38%|███▊      | 2188/5758 [12:28<20:20,  2.93it/s]
 38%|███▊      | 2189/5758 [12:28<19:40,  3.02it/s]
                                                   

 38%|███▊      | 2189/5758 [12:28<19:40,  3.02it/s]
 38%|███▊      | 2190/5758 [12:28<19:01,  3.13it/s]
                                                   

 38%|███▊      | 2190/5758 [12:28<19:01,  3.13it/s]
 38%|███▊      | 2191/5758 [12:28<18:47,  3.16it/s]
                                                   

 38%|███▊      | 2191/5758 [12:28<18:47,  3.16it/s]
 38%|███▊      | 2192/5758 [12:29<20:04,  2.96it/s]
                                                   

 38%|███▊      | 2192/5758 [12:29<20:04,  2.96it/s]
 38%|███▊      | 2193/5758 [12:29<20:56,  2.84it/s]
                                                   

 38%|███▊      | 2193/5758 [12:29<20:56,  2.84it/s]
 38%|███▊      | 2194/5758 [12:30<19:53,  2.99it/s]
                                                   

 38%|███▊      | 2194/5758 [12:30<19:53,  2.99it/s]
 38%|███▊      | 2195/5758 [12:30<18:44,  3.17it/s]
                                                   

 38%|███▊      | 2195/5758 [12:30<18:44,  3.17it/s]
 38%|███▊      | 2196/5758 [12:30<19:04,  3.11it/s]
                                                   

 38%|███▊      | 2196/5758 [12:30<19:04,  3.11it/s]
 38%|███▊      | 2197/5758 [12:30<18:47,  3.16it/s]
                                                   

 38%|███▊      | 2197/5758 [12:30<18:47,  3.16it/s]
 38%|███▊      | 2198/5758 [12:31<18:28,  3.21it/s]
                                                   

 38%|███▊      | 2198/5758 [12:31<18:28,  3.21it/s]
 38%|███▊      | 2199/5758 [12:31<18:44,  3.16it/s]
                                                   

 38%|███▊      | 2199/5758 [12:31<18:44,  3.16it/s]
 38%|███▊      | 2200/5758 [12:31<19:46,  3.00it/s]
                                                   

 38%|███▊      | 2200/5758 [12:31<19:46,  3.00it/s]
 38%|███▊      | 2201/5758 [12:32<19:21,  3.06it/s]
                                                   

 38%|███▊      | 2201/5758 [12:32<19:21,  3.06it/s]
 38%|███▊      | 2202/5758 [12:32<20:25,  2.90it/s]
                                                   

 38%|███▊      | 2202/5758 [12:32<20:25,  2.90it/s]
 38%|███▊      | 2203/5758 [12:33<20:48,  2.85it/s]
                                                   

 38%|███▊      | 2203/5758 [12:33<20:48,  2.85it/s]
 38%|███▊      | 2204/5758 [12:33<23:09,  2.56it/s]
                                                   

 38%|███▊      | 2204/5758 [12:33<23:09,  2.56it/s]
 38%|███▊      | 2205/5758 [12:33<23:11,  2.55it/s]
                                                   

 38%|███▊      | 2205/5758 [12:33<23:11,  2.55it/s]
 38%|███▊      | 2206/5758 [12:34<21:36,  2.74it/s]
                                                   

 38%|███▊      | 2206/5758 [12:34<21:36,  2.74it/s]
 38%|███▊      | 2207/5758 [12:34<20:16,  2.92it/s]
                                                   

 38%|███▊      | 2207/5758 [12:34<20:16,  2.92it/s]
 38%|███▊      | 2208/5758 [12:34<20:02,  2.95it/s]
                                                   

 38%|███▊      | 2208/5758 [12:34<20:02,  2.95it/s]
 38%|███▊      | 2209/5758 [12:35<20:10,  2.93it/s]
                                                   

 38%|███▊      | 2209/5758 [12:35<20:10,  2.93it/s]
 38%|███▊      | 2210/5758 [12:35<19:13,  3.08it/s]
                                                   

 38%|███▊      | 2210/5758 [12:35<19:13,  3.08it/s]
 38%|███▊      | 2211/5758 [12:35<20:18,  2.91it/s]
                                                   

 38%|███▊      | 2211/5758 [12:35<20:18,  2.91it/s]
 38%|███▊      | 2212/5758 [12:36<19:41,  3.00it/s]
                                                   

 38%|███▊      | 2212/5758 [12:36<19:41,  3.00it/s]
 38%|███▊      | 2213/5758 [12:36<19:12,  3.07it/s]
                                                   

 38%|███▊      | 2213/5758 [12:36<19:12,  3.07it/s]
 38%|███▊      | 2214/5758 [12:36<19:24,  3.04it/s]
                                                   

 38%|███▊      | 2214/5758 [12:36<19:24,  3.04it/s]
 38%|███▊      | 2215/5758 [12:37<19:26,  3.04it/s]
                                                   

 38%|███▊      | 2215/5758 [12:37<19:26,  3.04it/s]
 38%|███▊      | 2216/5758 [12:37<18:57,  3.11it/s]
                                                   

 38%|███▊      | 2216/5758 [12:37<18:57,  3.11it/s]
 39%|███▊      | 2217/5758 [12:37<18:43,  3.15it/s]
                                                   

 39%|███▊      | 2217/5758 [12:37<18:43,  3.15it/s]
 39%|███▊      | 2218/5758 [12:38<18:35,  3.17it/s]
                                                   

 39%|███▊      | 2218/5758 [12:38<18:35,  3.17it/s]
 39%|███▊      | 2219/5758 [12:38<18:14,  3.23it/s]
                                                   

 39%|███▊      | 2219/5758 [12:38<18:14,  3.23it/s]
 39%|███▊      | 2220/5758 [12:38<18:09,  3.25it/s]
                                                   

 39%|███▊      | 2220/5758 [12:38<18:09,  3.25it/s]
 39%|███▊      | 2221/5758 [12:39<19:07,  3.08it/s]
                                                   
{'loss': '1.196', 'grad_norm': '26.75', 'learning_rate': '6.24e-06', 'num_tokens': '1.421e+06', 'mean_token_accuracy': '0.7287', 'epoch': '0.3762'}
+{'loss': '0.5901', 'grad_norm': '6.188', 'learning_rate': '6.238e-06', 'num_tokens': '1.423e+06', 'mean_token_accuracy': '0.8308', 'epoch': '0.3763'}
+{'loss': '1.251', 'grad_norm': '9.5', 'learning_rate': '6.237e-06', 'num_tokens': '1.424e+06', 'mean_token_accuracy': '0.6645', 'epoch': '0.3765'}
+{'loss': '1.809', 'grad_norm': '15.75', 'learning_rate': '6.235e-06', 'num_tokens': '1.425e+06', 'mean_token_accuracy': '0.6122', 'epoch': '0.3767'}
+{'loss': '0.9864', 'grad_norm': '12.31', 'learning_rate': '6.233e-06', 'num_tokens': '1.426e+06', 'mean_token_accuracy': '0.7072', 'epoch': '0.3769'}
+{'loss': '1.023', 'grad_norm': '17.88', 'learning_rate': '6.231e-06', 'num_tokens': '1.426e+06', 'mean_token_accuracy': '0.7482', 'epoch': '0.377'}
+{'loss': '2.244', 'grad_norm': '9.25', 'learning_rate': '6.23e-06', 'num_tokens': '1.427e+06', 'mean_token_accuracy': '0.5219', 'epoch': '0.3772'}
+{'loss': '0.6557', 'grad_norm': '8.375', 'learning_rate': '6.228e-06', 'num_tokens': '1.429e+06', 'mean_token_accuracy': '0.7951', 'epoch': '0.3774'}
+{'loss': '1.625', 'grad_norm': '40.5', 'learning_rate': '6.226e-06', 'num_tokens': '1.429e+06', 'mean_token_accuracy': '0.6574', 'epoch': '0.3776'}
+{'loss': '1.066', 'grad_norm': '12.25', 'learning_rate': '6.224e-06', 'num_tokens': '1.43e+06', 'mean_token_accuracy': '0.6913', 'epoch': '0.3777'}
+{'loss': '2.167', 'grad_norm': '7.656', 'learning_rate': '6.223e-06', 'num_tokens': '1.432e+06', 'mean_token_accuracy': '0.5403', 'epoch': '0.3779'}
+{'loss': '1.188', 'grad_norm': '11.56', 'learning_rate': '6.221e-06', 'num_tokens': '1.433e+06', 'mean_token_accuracy': '0.6851', 'epoch': '0.3781'}
+{'loss': '1.349', 'grad_norm': '25.62', 'learning_rate': '6.219e-06', 'num_tokens': '1.433e+06', 'mean_token_accuracy': '0.6945', 'epoch': '0.3783'}
+{'loss': '1.107', 'grad_norm': '14.19', 'learning_rate': '6.217e-06', 'num_tokens': '1.434e+06', 'mean_token_accuracy': '0.7119', 'epoch': '0.3784'}
+{'loss': '1.847', 'grad_norm': '25.75', 'learning_rate': '6.216e-06', 'num_tokens': '1.434e+06', 'mean_token_accuracy': '0.6113', 'epoch': '0.3786'}
+{'loss': '1.103', 'grad_norm': '13.12', 'learning_rate': '6.214e-06', 'num_tokens': '1.435e+06', 'mean_token_accuracy': '0.7072', 'epoch': '0.3788'}
+{'loss': '1.359', 'grad_norm': '18.12', 'learning_rate': '6.212e-06', 'num_tokens': '1.435e+06', 'mean_token_accuracy': '0.7036', 'epoch': '0.379'}
+{'loss': '2.205', 'grad_norm': '14.88', 'learning_rate': '6.21e-06', 'num_tokens': '1.436e+06', 'mean_token_accuracy': '0.5718', 'epoch': '0.3791'}
+{'loss': '1.579', 'grad_norm': '38.25', 'learning_rate': '6.209e-06', 'num_tokens': '1.436e+06', 'mean_token_accuracy': '0.6438', 'epoch': '0.3793'}
+{'loss': '2.218', 'grad_norm': '11.19', 'learning_rate': '6.207e-06', 'num_tokens': '1.437e+06', 'mean_token_accuracy': '0.5577', 'epoch': '0.3795'}
+{'loss': '2.274', 'grad_norm': '12.44', 'learning_rate': '6.205e-06', 'num_tokens': '1.437e+06', 'mean_token_accuracy': '0.543', 'epoch': '0.3796'}
+{'loss': '1.702', 'grad_norm': '12.56', 'learning_rate': '6.204e-06', 'num_tokens': '1.438e+06', 'mean_token_accuracy': '0.6369', 'epoch': '0.3798'}
+{'loss': '1.43', 'grad_norm': '11.06', 'learning_rate': '6.202e-06', 'num_tokens': '1.439e+06', 'mean_token_accuracy': '0.6768', 'epoch': '0.38'}
+{'loss': '0.5985', 'grad_norm': '21.88', 'learning_rate': '6.2e-06', 'num_tokens': '1.439e+06', 'mean_token_accuracy': '0.8639', 'epoch': '0.3802'}
+{'loss': '2.021', 'grad_norm': '16.38', 'learning_rate': '6.198e-06', 'num_tokens': '1.44e+06', 'mean_token_accuracy': '0.5935', 'epoch': '0.3803'}
+{'loss': '0.6774', 'grad_norm': '8.25', 'learning_rate': '6.197e-06', 'num_tokens': '1.441e+06', 'mean_token_accuracy': '0.7829', 'epoch': '0.3805'}
+{'loss': '1.5', 'grad_norm': '18', 'learning_rate': '6.195e-06', 'num_tokens': '1.441e+06', 'mean_token_accuracy': '0.6842', 'epoch': '0.3807'}
+{'loss': '0.8631', 'grad_norm': '11.75', 'learning_rate': '6.193e-06', 'num_tokens': '1.442e+06', 'mean_token_accuracy': '0.7766', 'epoch': '0.3809'}
+{'loss': '0.7879', 'grad_norm': '19.75', 'learning_rate': '6.191e-06', 'num_tokens': '1.442e+06', 'mean_token_accuracy': '0.7987', 'epoch': '0.381'}
+{'loss': '0.9425', 'grad_norm': '37', 'learning_rate': '6.19e-06', 'num_tokens': '1.442e+06', 'mean_token_accuracy': '0.8113', 'epoch': '0.3812'}
+{'loss': '1.144', 'grad_norm': '12.88', 'learning_rate': '6.188e-06', 'num_tokens': '1.443e+06', 'mean_token_accuracy': '0.7506', 'epoch': '0.3814'}
+{'loss': '1.151', 'grad_norm': '24', 'learning_rate': '6.186e-06', 'num_tokens': '1.443e+06', 'mean_token_accuracy': '0.747', 'epoch': '0.3816'}
+{'loss': '1.402', 'grad_norm': '12.94', 'learning_rate': '6.184e-06', 'num_tokens': '1.444e+06', 'mean_token_accuracy': '0.6222', 'epoch': '0.3817'}
+{'loss': '2.547', 'grad_norm': '30', 'learning_rate': '6.183e-06', 'num_tokens': '1.444e+06', 'mean_token_accuracy': '0.4975', 'epoch': '0.3819'}
+{'loss': '1.731', 'grad_norm': '19.75', 'learning_rate': '6.181e-06', 'num_tokens': '1.444e+06', 'mean_token_accuracy': '0.6581', 'epoch': '0.3821'}
+{'loss': '1.356', 'grad_norm': '12.62', 'learning_rate': '6.179e-06', 'num_tokens': '1.445e+06', 'mean_token_accuracy': '0.6332', 'epoch': '0.3823'}
+{'loss': '2.048', 'grad_norm': '18.75', 'learning_rate': '6.177e-06', 'num_tokens': '1.446e+06', 'mean_token_accuracy': '0.5927', 'epoch': '0.3824'}
+{'loss': '1.136', 'grad_norm': '14.44', 'learning_rate': '6.176e-06', 'num_tokens': '1.446e+06', 'mean_token_accuracy': '0.6979', 'epoch': '0.3826'}
+{'loss': '1.496', 'grad_norm': '18.88', 'learning_rate': '6.174e-06', 'num_tokens': '1.446e+06', 'mean_token_accuracy': '0.6732', 'epoch': '0.3828'}
+{'loss': '1.193', 'grad_norm': '28.38', 'learning_rate': '6.172e-06', 'num_tokens': '1.447e+06', 'mean_token_accuracy': '0.7761', 'epoch': '0.3829'}
+{'loss': '1.226', 'grad_norm': '12.75', 'learning_rate': '6.171e-06', 'num_tokens': '1.447e+06', 'mean_token_accuracy': '0.6732', 'epoch': '0.3831'}
+{'loss': '1.312', 'grad_norm': '14.06', 'learning_rate': '6.169e-06', 'num_tokens': '1.448e+06', 'mean_token_accuracy': '0.6511', 'epoch': '0.3833'}
+{'loss': '1.393', 'grad_norm': '12.69', 'learning_rate': '6.167e-06', 'num_tokens': '1.449e+06', 'mean_token_accuracy': '0.639', 'epoch': '0.3835'}
+{'loss': '2.097', 'grad_norm': '16.38', 'learning_rate': '6.165e-06', 'num_tokens': '1.449e+06', 'mean_token_accuracy': '0.5826', 'epoch': '0.3836'}
+{'loss': '1.719', 'grad_norm': '12.12', 'learning_rate': '6.164e-06', 'num_tokens': '1.45e+06', 'mean_token_accuracy': '0.6356', 'epoch': '0.3838'}
+{'loss': '1.652', 'grad_norm': '15.62', 'learning_rate': '6.162e-06', 'num_tokens': '1.45e+06', 'mean_token_accuracy': '0.6126', 'epoch': '0.384'}
+{'loss': '1.185', 'grad_norm': '8.062', 'learning_rate': '6.16e-06', 'num_tokens': '1.452e+06', 'mean_token_accuracy': '0.7409', 'epoch': '0.3842'}
+{'loss': '1.009', 'grad_norm': '38', 'learning_rate': '6.158e-06', 'num_tokens': '1.452e+06', 'mean_token_accuracy': '0.7317', 'epoch': '0.3843'}
+{'loss': '1.082', 'grad_norm': '29.62', 'learning_rate': '6.157e-06', 'num_tokens': '1.452e+06', 'mean_token_accuracy': '0.82', 'epoch': '0.3845'}
+{'loss': '0.5981', 'grad_norm': '6.906', 'learning_rate': '6.155e-06', 'num_tokens': '1.454e+06', 'mean_token_accuracy': '0.8183', 'epoch': '0.3847'}
+{'loss': '1.238', 'grad_norm': '13.25', 'learning_rate': '6.153e-06', 'num_tokens': '1.454e+06', 'mean_token_accuracy': '0.6485', 'epoch': '0.3849'}
+{'loss': '1.181', 'grad_norm': '13.94', 'learning_rate': '6.151e-06', 'num_tokens': '1.455e+06', 'mean_token_accuracy': '0.669', 'epoch': '0.385'}
+{'loss': '2.054', 'grad_norm': '7.656', 'learning_rate': '6.15e-06', 'num_tokens': '1.456e+06', 'mean_token_accuracy': '0.561', 'epoch': '0.3852'}
+{'loss': '0.6885', 'grad_norm': '9', 'learning_rate': '6.148e-06', 'num_tokens': '1.458e+06', 'mean_token_accuracy': '0.7962', 'epoch': '0.3854'}
+{'loss': '1.088', 'grad_norm': '13.31', 'learning_rate': '6.146e-06', 'num_tokens': '1.458e+06', 'mean_token_accuracy': '0.709', 'epoch': '0.3856'}
+
 39%|███▊      | 2221/5758 [12:39<19:07,  3.08it/s]
 39%|███▊      | 2222/5758 [12:39<20:02,  2.94it/s]
                                                   

 39%|███▊      | 2222/5758 [12:39<20:02,  2.94it/s]
 39%|███▊      | 2223/5758 [12:39<19:23,  3.04it/s]
                                                   

 39%|███▊      | 2223/5758 [12:39<19:23,  3.04it/s]
 39%|███▊      | 2224/5758 [12:40<20:08,  2.92it/s]
                                                   

 39%|███▊      | 2224/5758 [12:40<20:08,  2.92it/s]
 39%|███▊      | 2225/5758 [12:40<19:52,  2.96it/s]
                                                   

 39%|███▊      | 2225/5758 [12:40<19:52,  2.96it/s]
 39%|███▊      | 2226/5758 [12:40<19:56,  2.95it/s]
                                                   

 39%|███▊      | 2226/5758 [12:40<19:56,  2.95it/s]
 39%|███▊      | 2227/5758 [12:41<19:24,  3.03it/s]
                                                   

 39%|███▊      | 2227/5758 [12:41<19:24,  3.03it/s]
 39%|███▊      | 2228/5758 [12:41<20:09,  2.92it/s]
                                                   

 39%|███▊      | 2228/5758 [12:41<20:09,  2.92it/s]
 39%|███▊      | 2229/5758 [12:41<19:26,  3.03it/s]
                                                   

 39%|███▊      | 2229/5758 [12:41<19:26,  3.03it/s]
 39%|███▊      | 2230/5758 [12:42<18:54,  3.11it/s]
                                                   

 39%|███▊      | 2230/5758 [12:42<18:54,  3.11it/s]
 39%|███▊      | 2231/5758 [12:42<20:31,  2.87it/s]
                                                   

 39%|███▊      | 2231/5758 [12:42<20:31,  2.87it/s]
 39%|███▉      | 2232/5758 [12:42<19:47,  2.97it/s]
                                                   

 39%|███▉      | 2232/5758 [12:42<19:47,  2.97it/s]
 39%|███▉      | 2233/5758 [12:43<19:25,  3.02it/s]
                                                   

 39%|███▉      | 2233/5758 [12:43<19:25,  3.02it/s]
 39%|███▉      | 2234/5758 [12:43<19:19,  3.04it/s]
                                                   

 39%|███▉      | 2234/5758 [12:43<19:19,  3.04it/s]
 39%|███▉      | 2235/5758 [12:43<18:55,  3.10it/s]
                                                   

 39%|███▉      | 2235/5758 [12:43<18:55,  3.10it/s]
 39%|███▉      | 2236/5758 [12:43<18:44,  3.13it/s]
                                                   

 39%|███▉      | 2236/5758 [12:43<18:44,  3.13it/s]
 39%|███▉      | 2237/5758 [12:44<18:23,  3.19it/s]
                                                   

 39%|███▉      | 2237/5758 [12:44<18:23,  3.19it/s]
 39%|███▉      | 2238/5758 [12:44<18:15,  3.21it/s]
                                                   

 39%|███▉      | 2238/5758 [12:44<18:15,  3.21it/s]
 39%|███▉      | 2239/5758 [12:44<18:24,  3.19it/s]
                                                   

 39%|███▉      | 2239/5758 [12:44<18:24,  3.19it/s]
 39%|███▉      | 2240/5758 [12:45<18:47,  3.12it/s]
                                                   

 39%|███▉      | 2240/5758 [12:45<18:47,  3.12it/s]
 39%|███▉      | 2241/5758 [12:45<19:31,  3.00it/s]
                                                   

 39%|███▉      | 2241/5758 [12:45<19:31,  3.00it/s]
 39%|███▉      | 2242/5758 [12:45<20:04,  2.92it/s]
                                                   

 39%|███▉      | 2242/5758 [12:45<20:04,  2.92it/s]
 39%|███▉      | 2243/5758 [12:46<19:34,  2.99it/s]
                                                   

 39%|███▉      | 2243/5758 [12:46<19:34,  2.99it/s]
 39%|███▉      | 2244/5758 [12:46<20:52,  2.81it/s]
                                                   

 39%|███▉      | 2244/5758 [12:46<20:52,  2.81it/s]
 39%|███▉      | 2245/5758 [12:47<21:21,  2.74it/s]
                                                   

 39%|███▉      | 2245/5758 [12:47<21:21,  2.74it/s]
 39%|███▉      | 2246/5758 [12:47<21:50,  2.68it/s]
                                                   

 39%|███▉      | 2246/5758 [12:47<21:50,  2.68it/s]
 39%|███▉      | 2247/5758 [12:47<21:41,  2.70it/s]
                                                   

 39%|███▉      | 2247/5758 [12:47<21:41,  2.70it/s]
 39%|███▉      | 2248/5758 [12:48<20:55,  2.80it/s]
                                                   

 39%|███▉      | 2248/5758 [12:48<20:55,  2.80it/s]
 39%|███▉      | 2249/5758 [12:48<20:16,  2.88it/s]
                                                   

 39%|███▉      | 2249/5758 [12:48<20:16,  2.88it/s]
 39%|███▉      | 2250/5758 [12:48<19:43,  2.96it/s]
                                                   

 39%|███▉      | 2250/5758 [12:48<19:43,  2.96it/s]
 39%|███▉      | 2251/5758 [12:49<19:34,  2.98it/s]
                                                   

 39%|███▉      | 2251/5758 [12:49<19:34,  2.98it/s]
 39%|███▉      | 2252/5758 [12:49<19:12,  3.04it/s]
                                                   

 39%|███▉      | 2252/5758 [12:49<19:12,  3.04it/s]
 39%|███▉      | 2253/5758 [12:49<18:55,  3.09it/s]
                                                   

 39%|███▉      | 2253/5758 [12:49<18:55,  3.09it/s]
 39%|███▉      | 2254/5758 [12:50<18:26,  3.17it/s]
                                                   

 39%|███▉      | 2254/5758 [12:50<18:26,  3.17it/s]
 39%|███▉      | 2255/5758 [12:50<18:09,  3.21it/s]
                                                   

 39%|███▉      | 2255/5758 [12:50<18:09,  3.21it/s]
 39%|███▉      | 2256/5758 [12:50<18:29,  3.16it/s]
                                                   

 39%|███▉      | 2256/5758 [12:50<18:29,  3.16it/s]
 39%|███▉      | 2257/5758 [12:51<19:01,  3.07it/s]
                                                   

 39%|███▉      | 2257/5758 [12:51<19:01,  3.07it/s]
 39%|███▉      | 2258/5758 [12:51<19:39,  2.97it/s]
                                                   

 39%|███▉      | 2258/5758 [12:51<19:39,  2.97it/s]
 39%|███▉      | 2259/5758 [12:51<20:32,  2.84it/s]
                                                   

 39%|███▉      | 2259/5758 [12:51<20:32,  2.84it/s]
 39%|███▉      | 2260/5758 [12:52<21:50,  2.67it/s]
                                                   

 39%|███▉      | 2260/5758 [12:52<21:50,  2.67it/s]
 39%|███▉      | 2261/5758 [12:52<21:20,  2.73it/s]
                                                   

 39%|███▉      | 2261/5758 [12:52<21:20,  2.73it/s]
 39%|███▉      | 2262/5758 [12:52<20:09,  2.89it/s]
                                                   

 39%|███▉      | 2262/5758 [12:52<20:09,  2.89it/s]
 39%|███▉      | 2263/5758 [12:53<19:37,  2.97it/s]
                                                   

 39%|███▉      | 2263/5758 [12:53<19:37,  2.97it/s]
 39%|███▉      | 2264/5758 [12:53<18:52,  3.08it/s]
                                                   

 39%|███▉      | 2264/5758 [12:53<18:52,  3.08it/s]
 39%|███▉      | 2265/5758 [12:53<19:44,  2.95it/s]
                                                   

 39%|███▉      | 2265/5758 [12:53<19:44,  2.95it/s]
 39%|███▉      | 2266/5758 [12:54<19:13,  3.03it/s]
                                                   

 39%|███▉      | 2266/5758 [12:54<19:13,  3.03it/s]
 39%|███▉      | 2267/5758 [12:54<19:08,  3.04it/s]
                                                   

 39%|███▉      | 2267/5758 [12:54<19:08,  3.04it/s]
 39%|███▉      | 2268/5758 [12:54<18:36,  3.12it/s]
                                                   

 39%|███▉      | 2268/5758 [12:54<18:36,  3.12it/s]
 39%|███▉      | 2269/5758 [12:55<19:41,  2.95it/s]
                                                   

 39%|███▉      | 2269/5758 [12:55<19:41,  2.95it/s]
 39%|███▉      | 2270/5758 [12:55<18:44,  3.10it/s]
                                                   

 39%|███▉      | 2270/5758 [12:55<18:44,  3.10it/s]
 39%|███▉      | 2271/5758 [12:55<19:24,  2.99it/s]
                                                   

 39%|███▉      | 2271/5758 [12:55<19:24,  2.99it/s]
 39%|███▉      | 2272/5758 [12:56<19:17,  3.01it/s]
                                                   

 39%|███▉      | 2272/5758 [12:56<19:17,  3.01it/s]
 39%|███▉      | 2273/5758 [12:56<19:35,  2.96it/s]
                                                   

 39%|███▉      | 2273/5758 [12:56<19:35,  2.96it/s]
 39%|███▉      | 2274/5758 [12:56<19:46,  2.94it/s]
                                                   

 39%|███▉      | 2274/5758 [12:56<19:46,  2.94it/s]
 40%|███▉      | 2275/5758 [12:57<19:06,  3.04it/s]
                                                   

 40%|███▉      | 2275/5758 [12:57<19:06,  3.04it/s]
 40%|███▉      | 2276/5758 [12:57<19:01,  3.05it/s]
                                                   
{'loss': '1.337', 'grad_norm': '5.656', 'learning_rate': '6.144e-06', 'num_tokens': '1.463e+06', 'mean_token_accuracy': '0.6847', 'epoch': '0.3857'}
+{'loss': '1.678', 'grad_norm': '22.62', 'learning_rate': '6.143e-06', 'num_tokens': '1.463e+06', 'mean_token_accuracy': '0.6429', 'epoch': '0.3859'}
+{'loss': '0.8144', 'grad_norm': '12.12', 'learning_rate': '6.141e-06', 'num_tokens': '1.464e+06', 'mean_token_accuracy': '0.759', 'epoch': '0.3861'}
+{'loss': '0.6591', 'grad_norm': '30.12', 'learning_rate': '6.139e-06', 'num_tokens': '1.464e+06', 'mean_token_accuracy': '0.8443', 'epoch': '0.3862'}
+{'loss': '1', 'grad_norm': '20.5', 'learning_rate': '6.138e-06', 'num_tokens': '1.464e+06', 'mean_token_accuracy': '0.7653', 'epoch': '0.3864'}
+{'loss': '2.114', 'grad_norm': '15.44', 'learning_rate': '6.136e-06', 'num_tokens': '1.464e+06', 'mean_token_accuracy': '0.5553', 'epoch': '0.3866'}
+{'loss': '0.8198', 'grad_norm': '10.69', 'learning_rate': '6.134e-06', 'num_tokens': '1.465e+06', 'mean_token_accuracy': '0.7874', 'epoch': '0.3868'}
+{'loss': '0.9804', 'grad_norm': '18.88', 'learning_rate': '6.132e-06', 'num_tokens': '1.465e+06', 'mean_token_accuracy': '0.7675', 'epoch': '0.3869'}
+{'loss': '0.6012', 'grad_norm': '7.5', 'learning_rate': '6.131e-06', 'num_tokens': '1.467e+06', 'mean_token_accuracy': '0.8293', 'epoch': '0.3871'}
+{'loss': '1.244', 'grad_norm': '12.88', 'learning_rate': '6.129e-06', 'num_tokens': '1.467e+06', 'mean_token_accuracy': '0.6561', 'epoch': '0.3873'}
+{'loss': '0.825', 'grad_norm': '18.12', 'learning_rate': '6.127e-06', 'num_tokens': '1.468e+06', 'mean_token_accuracy': '0.8404', 'epoch': '0.3875'}
+{'loss': '1.29', 'grad_norm': '13.62', 'learning_rate': '6.125e-06', 'num_tokens': '1.468e+06', 'mean_token_accuracy': '0.6492', 'epoch': '0.3876'}
+{'loss': '0.9195', 'grad_norm': '21.25', 'learning_rate': '6.124e-06', 'num_tokens': '1.468e+06', 'mean_token_accuracy': '0.8138', 'epoch': '0.3878'}
+{'loss': '1.427', 'grad_norm': '32.5', 'learning_rate': '6.122e-06', 'num_tokens': '1.469e+06', 'mean_token_accuracy': '0.6684', 'epoch': '0.388'}
+{'loss': '1.406', 'grad_norm': '14.81', 'learning_rate': '6.12e-06', 'num_tokens': '1.469e+06', 'mean_token_accuracy': '0.6455', 'epoch': '0.3882'}
+{'loss': '0.943', 'grad_norm': '11.69', 'learning_rate': '6.118e-06', 'num_tokens': '1.47e+06', 'mean_token_accuracy': '0.7201', 'epoch': '0.3883'}
+{'loss': '0.9319', 'grad_norm': '13.31', 'learning_rate': '6.117e-06', 'num_tokens': '1.471e+06', 'mean_token_accuracy': '0.7434', 'epoch': '0.3885'}
+{'loss': '1.065', 'grad_norm': '12.62', 'learning_rate': '6.115e-06', 'num_tokens': '1.471e+06', 'mean_token_accuracy': '0.6896', 'epoch': '0.3887'}
+{'loss': '1.199', 'grad_norm': '12.31', 'learning_rate': '6.113e-06', 'num_tokens': '1.472e+06', 'mean_token_accuracy': '0.6775', 'epoch': '0.3889'}
+{'loss': '0.5683', 'grad_norm': '18.25', 'learning_rate': '6.111e-06', 'num_tokens': '1.472e+06', 'mean_token_accuracy': '0.8618', 'epoch': '0.389'}
+{'loss': '1.237', 'grad_norm': '5.688', 'learning_rate': '6.11e-06', 'num_tokens': '1.476e+06', 'mean_token_accuracy': '0.695', 'epoch': '0.3892'}
+{'loss': '1.071', 'grad_norm': '13.5', 'learning_rate': '6.108e-06', 'num_tokens': '1.477e+06', 'mean_token_accuracy': '0.6932', 'epoch': '0.3894'}
+{'loss': '1.273', 'grad_norm': '12.56', 'learning_rate': '6.106e-06', 'num_tokens': '1.478e+06', 'mean_token_accuracy': '0.6404', 'epoch': '0.3895'}
+{'loss': '1.423', 'grad_norm': '26.75', 'learning_rate': '6.105e-06', 'num_tokens': '1.478e+06', 'mean_token_accuracy': '0.6939', 'epoch': '0.3897'}
+{'loss': '1.144', 'grad_norm': '13.75', 'learning_rate': '6.103e-06', 'num_tokens': '1.478e+06', 'mean_token_accuracy': '0.6945', 'epoch': '0.3899'}
+{'loss': '1.984', 'grad_norm': '13.12', 'learning_rate': '6.101e-06', 'num_tokens': '1.479e+06', 'mean_token_accuracy': '0.5901', 'epoch': '0.3901'}
+{'loss': '0.4313', 'grad_norm': '15.25', 'learning_rate': '6.099e-06', 'num_tokens': '1.479e+06', 'mean_token_accuracy': '0.9046', 'epoch': '0.3902'}
+{'loss': '1.304', 'grad_norm': '33.5', 'learning_rate': '6.098e-06', 'num_tokens': '1.479e+06', 'mean_token_accuracy': '0.766', 'epoch': '0.3904'}
+{'loss': '0.7121', 'grad_norm': '20.12', 'learning_rate': '6.096e-06', 'num_tokens': '1.48e+06', 'mean_token_accuracy': '0.8298', 'epoch': '0.3906'}
+{'loss': '1.087', 'grad_norm': '9.312', 'learning_rate': '6.094e-06', 'num_tokens': '1.481e+06', 'mean_token_accuracy': '0.7442', 'epoch': '0.3908'}
+{'loss': '0.9665', 'grad_norm': '11.06', 'learning_rate': '6.092e-06', 'num_tokens': '1.482e+06', 'mean_token_accuracy': '0.7043', 'epoch': '0.3909'}
+{'loss': '0.8223', 'grad_norm': '26.38', 'learning_rate': '6.091e-06', 'num_tokens': '1.482e+06', 'mean_token_accuracy': '0.8081', 'epoch': '0.3911'}
+{'loss': '0.9558', 'grad_norm': '11.56', 'learning_rate': '6.089e-06', 'num_tokens': '1.483e+06', 'mean_token_accuracy': '0.7447', 'epoch': '0.3913'}
+{'loss': '2.21', 'grad_norm': '16.88', 'learning_rate': '6.087e-06', 'num_tokens': '1.483e+06', 'mean_token_accuracy': '0.5696', 'epoch': '0.3915'}
+{'loss': '0.4551', 'grad_norm': '10.44', 'learning_rate': '6.085e-06', 'num_tokens': '1.484e+06', 'mean_token_accuracy': '0.893', 'epoch': '0.3916'}
+{'loss': '1.247', 'grad_norm': '11.94', 'learning_rate': '6.084e-06', 'num_tokens': '1.485e+06', 'mean_token_accuracy': '0.6586', 'epoch': '0.3918'}
+{'loss': '0.9971', 'grad_norm': '14.75', 'learning_rate': '6.082e-06', 'num_tokens': '1.485e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.392'}
+{'loss': '1.428', 'grad_norm': '9.375', 'learning_rate': '6.08e-06', 'num_tokens': '1.486e+06', 'mean_token_accuracy': '0.6596', 'epoch': '0.3922'}
+{'loss': '1.11', 'grad_norm': '17.38', 'learning_rate': '6.078e-06', 'num_tokens': '1.487e+06', 'mean_token_accuracy': '0.7508', 'epoch': '0.3923'}
+{'loss': '1.135', 'grad_norm': '8.812', 'learning_rate': '6.077e-06', 'num_tokens': '1.488e+06', 'mean_token_accuracy': '0.7417', 'epoch': '0.3925'}
+{'loss': '0.9327', 'grad_norm': '19', 'learning_rate': '6.075e-06', 'num_tokens': '1.488e+06', 'mean_token_accuracy': '0.7725', 'epoch': '0.3927'}
+{'loss': '0.7147', 'grad_norm': '19.12', 'learning_rate': '6.073e-06', 'num_tokens': '1.488e+06', 'mean_token_accuracy': '0.8535', 'epoch': '0.3928'}
+{'loss': '1.668', 'grad_norm': '8.562', 'learning_rate': '6.072e-06', 'num_tokens': '1.49e+06', 'mean_token_accuracy': '0.6496', 'epoch': '0.393'}
+{'loss': '1.18', 'grad_norm': '14', 'learning_rate': '6.07e-06', 'num_tokens': '1.491e+06', 'mean_token_accuracy': '0.6618', 'epoch': '0.3932'}
+{'loss': '1.031', 'grad_norm': '16.88', 'learning_rate': '6.068e-06', 'num_tokens': '1.491e+06', 'mean_token_accuracy': '0.7757', 'epoch': '0.3934'}
+{'loss': '1.719', 'grad_norm': '30.62', 'learning_rate': '6.066e-06', 'num_tokens': '1.491e+06', 'mean_token_accuracy': '0.7062', 'epoch': '0.3935'}
+{'loss': '1.003', 'grad_norm': '11.38', 'learning_rate': '6.065e-06', 'num_tokens': '1.492e+06', 'mean_token_accuracy': '0.7088', 'epoch': '0.3937'}
+{'loss': '1.222', 'grad_norm': '12.56', 'learning_rate': '6.063e-06', 'num_tokens': '1.492e+06', 'mean_token_accuracy': '0.6609', 'epoch': '0.3939'}
+{'loss': '0.5728', 'grad_norm': '8.375', 'learning_rate': '6.061e-06', 'num_tokens': '1.494e+06', 'mean_token_accuracy': '0.8203', 'epoch': '0.3941'}
+{'loss': '0.6952', 'grad_norm': '41.25', 'learning_rate': '6.059e-06', 'num_tokens': '1.494e+06', 'mean_token_accuracy': '0.8125', 'epoch': '0.3942'}
+{'loss': '0.827', 'grad_norm': '13.5', 'learning_rate': '6.058e-06', 'num_tokens': '1.494e+06', 'mean_token_accuracy': '0.7398', 'epoch': '0.3944'}
+{'loss': '1.266', 'grad_norm': '7.312', 'learning_rate': '6.056e-06', 'num_tokens': '1.496e+06', 'mean_token_accuracy': '0.7214', 'epoch': '0.3946'}
+{'loss': '2.325', 'grad_norm': '15.81', 'learning_rate': '6.054e-06', 'num_tokens': '1.496e+06', 'mean_token_accuracy': '0.5682', 'epoch': '0.3948'}
+{'loss': '1.908', 'grad_norm': '10.56', 'learning_rate': '6.052e-06', 'num_tokens': '1.497e+06', 'mean_token_accuracy': '0.5964', 'epoch': '0.3949'}
+{'loss': '1.82', 'grad_norm': '14.19', 'learning_rate': '6.051e-06', 'num_tokens': '1.498e+06', 'mean_token_accuracy': '0.6077', 'epoch': '0.3951'}
+
 40%|███▉      | 2276/5758 [12:57<19:01,  3.05it/s]
 40%|███▉      | 2277/5758 [12:57<18:29,  3.14it/s]
                                                   

 40%|███▉      | 2277/5758 [12:57<18:29,  3.14it/s]
 40%|███▉      | 2278/5758 [12:58<18:21,  3.16it/s]
                                                   

 40%|███▉      | 2278/5758 [12:58<18:21,  3.16it/s]
 40%|███▉      | 2279/5758 [12:58<17:59,  3.22it/s]
                                                   

 40%|███▉      | 2279/5758 [12:58<17:59,  3.22it/s]
 40%|███▉      | 2280/5758 [12:58<18:12,  3.18it/s]
                                                   

 40%|███▉      | 2280/5758 [12:58<18:12,  3.18it/s]
 40%|███▉      | 2281/5758 [12:59<18:07,  3.20it/s]
                                                   

 40%|███▉      | 2281/5758 [12:59<18:07,  3.20it/s]
 40%|███▉      | 2282/5758 [12:59<17:50,  3.25it/s]
                                                   

 40%|███▉      | 2282/5758 [12:59<17:50,  3.25it/s]
 40%|███▉      | 2283/5758 [12:59<18:42,  3.10it/s]
                                                   

 40%|███▉      | 2283/5758 [12:59<18:42,  3.10it/s]
 40%|███▉      | 2284/5758 [12:59<18:40,  3.10it/s]
                                                   

 40%|███▉      | 2284/5758 [12:59<18:40,  3.10it/s]
 40%|███▉      | 2285/5758 [13:00<21:13,  2.73it/s]
                                                   

 40%|███▉      | 2285/5758 [13:00<21:13,  2.73it/s]
 40%|███▉      | 2286/5758 [13:00<20:41,  2.80it/s]
                                                   

 40%|███▉      | 2286/5758 [13:00<20:41,  2.80it/s]
 40%|███▉      | 2287/5758 [13:01<21:11,  2.73it/s]
                                                   

 40%|███▉      | 2287/5758 [13:01<21:11,  2.73it/s]
 40%|███▉      | 2288/5758 [13:01<20:18,  2.85it/s]
                                                   

 40%|███▉      | 2288/5758 [13:01<20:18,  2.85it/s]
 40%|███▉      | 2289/5758 [13:01<19:16,  3.00it/s]
                                                   

 40%|███▉      | 2289/5758 [13:01<19:16,  3.00it/s]
 40%|███▉      | 2290/5758 [13:02<18:57,  3.05it/s]
                                                   

 40%|███▉      | 2290/5758 [13:02<18:57,  3.05it/s]
 40%|███▉      | 2291/5758 [13:02<18:24,  3.14it/s]
                                                   

 40%|███▉      | 2291/5758 [13:02<18:24,  3.14it/s]
 40%|███▉      | 2292/5758 [13:02<18:04,  3.19it/s]
                                                   

 40%|███▉      | 2292/5758 [13:02<18:04,  3.19it/s]
 40%|███▉      | 2293/5758 [13:03<19:42,  2.93it/s]
                                                   

 40%|███▉      | 2293/5758 [13:03<19:42,  2.93it/s]
 40%|███▉      | 2294/5758 [13:03<20:06,  2.87it/s]
                                                   

 40%|███▉      | 2294/5758 [13:03<20:06,  2.87it/s]
 40%|███▉      | 2295/5758 [13:03<19:05,  3.02it/s]
                                                   

 40%|███▉      | 2295/5758 [13:03<19:05,  3.02it/s]
 40%|███▉      | 2296/5758 [13:04<18:45,  3.08it/s]
                                                   

 40%|███▉      | 2296/5758 [13:04<18:45,  3.08it/s]
 40%|███▉      | 2297/5758 [13:04<19:18,  2.99it/s]
                                                   

 40%|███▉      | 2297/5758 [13:04<19:18,  2.99it/s]
 40%|███▉      | 2298/5758 [13:04<19:59,  2.89it/s]
                                                   

 40%|███▉      | 2298/5758 [13:04<19:59,  2.89it/s]
 40%|███▉      | 2299/5758 [13:05<19:24,  2.97it/s]
                                                   

 40%|███▉      | 2299/5758 [13:05<19:24,  2.97it/s]
 40%|███▉      | 2300/5758 [13:05<19:06,  3.02it/s]
                                                   

 40%|███▉      | 2300/5758 [13:05<19:06,  3.02it/s]
 40%|███▉      | 2301/5758 [13:05<18:53,  3.05it/s]
                                                   

 40%|███▉      | 2301/5758 [13:05<18:53,  3.05it/s]
 40%|███▉      | 2302/5758 [13:06<18:23,  3.13it/s]
                                                   

 40%|███▉      | 2302/5758 [13:06<18:23,  3.13it/s]
 40%|███▉      | 2303/5758 [13:06<18:06,  3.18it/s]
                                                   

 40%|███▉      | 2303/5758 [13:06<18:06,  3.18it/s]
 40%|████      | 2304/5758 [13:06<19:14,  2.99it/s]
                                                   

 40%|████      | 2304/5758 [13:06<19:14,  2.99it/s]
 40%|████      | 2305/5758 [13:07<18:25,  3.12it/s]
                                                   

 40%|████      | 2305/5758 [13:07<18:25,  3.12it/s]
 40%|████      | 2306/5758 [13:07<18:35,  3.10it/s]
                                                   

 40%|████      | 2306/5758 [13:07<18:35,  3.10it/s]
 40%|████      | 2307/5758 [13:07<18:32,  3.10it/s]
                                                   

 40%|████      | 2307/5758 [13:07<18:32,  3.10it/s]
 40%|████      | 2308/5758 [13:08<19:27,  2.95it/s]
                                                   

 40%|████      | 2308/5758 [13:08<19:27,  2.95it/s]
 40%|████      | 2309/5758 [13:08<18:39,  3.08it/s]
                                                   

 40%|████      | 2309/5758 [13:08<18:39,  3.08it/s]
 40%|████      | 2310/5758 [13:08<18:40,  3.08it/s]
                                                   

 40%|████      | 2310/5758 [13:08<18:40,  3.08it/s]
 40%|████      | 2311/5758 [13:09<20:32,  2.80it/s]
                                                   

 40%|████      | 2311/5758 [13:09<20:32,  2.80it/s]
 40%|████      | 2312/5758 [13:09<20:20,  2.82it/s]
                                                   

 40%|████      | 2312/5758 [13:09<20:20,  2.82it/s]
 40%|████      | 2313/5758 [13:09<19:39,  2.92it/s]
                                                   

 40%|████      | 2313/5758 [13:09<19:39,  2.92it/s]
 40%|████      | 2314/5758 [13:10<18:37,  3.08it/s]
                                                   

 40%|████      | 2314/5758 [13:10<18:37,  3.08it/s]
 40%|████      | 2315/5758 [13:10<18:05,  3.17it/s]
                                                   

 40%|████      | 2315/5758 [13:10<18:05,  3.17it/s]
 40%|████      | 2316/5758 [13:10<18:06,  3.17it/s]
                                                   

 40%|████      | 2316/5758 [13:10<18:06,  3.17it/s]
 40%|████      | 2317/5758 [13:10<17:40,  3.24it/s]
                                                   

 40%|████      | 2317/5758 [13:10<17:40,  3.24it/s]
 40%|████      | 2318/5758 [13:11<18:15,  3.14it/s]
                                                   

 40%|████      | 2318/5758 [13:11<18:15,  3.14it/s]
 40%|████      | 2319/5758 [13:11<19:08,  2.99it/s]
                                                   

 40%|████      | 2319/5758 [13:11<19:08,  2.99it/s]
 40%|████      | 2320/5758 [13:11<18:18,  3.13it/s]
                                                   

 40%|████      | 2320/5758 [13:11<18:18,  3.13it/s]
 40%|████      | 2321/5758 [13:12<18:08,  3.16it/s]
                                                   

 40%|████      | 2321/5758 [13:12<18:08,  3.16it/s]
 40%|████      | 2322/5758 [13:12<17:32,  3.26it/s]
                                                   

 40%|████      | 2322/5758 [13:12<17:32,  3.26it/s]
 40%|████      | 2323/5758 [13:12<18:23,  3.11it/s]
                                                   

 40%|████      | 2323/5758 [13:12<18:23,  3.11it/s]
 40%|████      | 2324/5758 [13:13<18:05,  3.16it/s]
                                                   

 40%|████      | 2324/5758 [13:13<18:05,  3.16it/s]
 40%|████      | 2325/5758 [13:13<17:40,  3.24it/s]
                                                   

 40%|████      | 2325/5758 [13:13<17:40,  3.24it/s]
 40%|████      | 2326/5758 [13:13<18:58,  3.01it/s]
                                                   

 40%|████      | 2326/5758 [13:13<18:58,  3.01it/s]
 40%|████      | 2327/5758 [13:14<18:23,  3.11it/s]
                                                   

 40%|████      | 2327/5758 [13:14<18:23,  3.11it/s]
 40%|████      | 2328/5758 [13:14<18:12,  3.14it/s]
                                                   

 40%|████      | 2328/5758 [13:14<18:12,  3.14it/s]
 40%|████      | 2329/5758 [13:14<18:39,  3.06it/s]
                                                   

 40%|████      | 2329/5758 [13:14<18:39,  3.06it/s]
 40%|████      | 2330/5758 [13:15<19:21,  2.95it/s]
                                                   

 40%|████      | 2330/5758 [13:15<19:21,  2.95it/s]
 40%|████      | 2331/5758 [13:15<18:46,  3.04it/s]
                                                   
{'loss': '1.071', 'grad_norm': '11.62', 'learning_rate': '6.049e-06', 'num_tokens': '1.499e+06', 'mean_token_accuracy': '0.7145', 'epoch': '0.3953'}
+{'loss': '1.29', 'grad_norm': '13.88', 'learning_rate': '6.047e-06', 'num_tokens': '1.499e+06', 'mean_token_accuracy': '0.6396', 'epoch': '0.3954'}
+{'loss': '1.461', 'grad_norm': '12.44', 'learning_rate': '6.046e-06', 'num_tokens': '1.5e+06', 'mean_token_accuracy': '0.6321', 'epoch': '0.3956'}
+{'loss': '1.24', 'grad_norm': '16', 'learning_rate': '6.044e-06', 'num_tokens': '1.5e+06', 'mean_token_accuracy': '0.7083', 'epoch': '0.3958'}
+{'loss': '1.855', 'grad_norm': '10.81', 'learning_rate': '6.042e-06', 'num_tokens': '1.501e+06', 'mean_token_accuracy': '0.6197', 'epoch': '0.396'}
+{'loss': '1.286', 'grad_norm': '11.88', 'learning_rate': '6.04e-06', 'num_tokens': '1.502e+06', 'mean_token_accuracy': '0.6566', 'epoch': '0.3961'}
+{'loss': '1.266', 'grad_norm': '12.44', 'learning_rate': '6.039e-06', 'num_tokens': '1.503e+06', 'mean_token_accuracy': '0.6672', 'epoch': '0.3963'}
+{'loss': '1.548', 'grad_norm': '17.25', 'learning_rate': '6.037e-06', 'num_tokens': '1.503e+06', 'mean_token_accuracy': '0.6757', 'epoch': '0.3965'}
+{'loss': '1.181', 'grad_norm': '33.25', 'learning_rate': '6.035e-06', 'num_tokens': '1.503e+06', 'mean_token_accuracy': '0.7283', 'epoch': '0.3967'}
+{'loss': '1.227', 'grad_norm': '17.88', 'learning_rate': '6.033e-06', 'num_tokens': '1.504e+06', 'mean_token_accuracy': '0.725', 'epoch': '0.3968'}
+{'loss': '1.002', 'grad_norm': '11.38', 'learning_rate': '6.032e-06', 'num_tokens': '1.504e+06', 'mean_token_accuracy': '0.7274', 'epoch': '0.397'}
+{'loss': '1.004', 'grad_norm': '12.06', 'learning_rate': '6.03e-06', 'num_tokens': '1.505e+06', 'mean_token_accuracy': '0.7209', 'epoch': '0.3972'}
+{'loss': '1.269', 'grad_norm': '12.12', 'learning_rate': '6.028e-06', 'num_tokens': '1.506e+06', 'mean_token_accuracy': '0.7143', 'epoch': '0.3974'}
+{'loss': '2.033', 'grad_norm': '8.812', 'learning_rate': '6.026e-06', 'num_tokens': '1.507e+06', 'mean_token_accuracy': '0.586', 'epoch': '0.3975'}
+{'loss': '1.113', 'grad_norm': '29.25', 'learning_rate': '6.025e-06', 'num_tokens': '1.507e+06', 'mean_token_accuracy': '0.7386', 'epoch': '0.3977'}
+{'loss': '1.03', 'grad_norm': '41.5', 'learning_rate': '6.023e-06', 'num_tokens': '1.507e+06', 'mean_token_accuracy': '0.7662', 'epoch': '0.3979'}
+{'loss': '1.129', 'grad_norm': '7.625', 'learning_rate': '6.021e-06', 'num_tokens': '1.509e+06', 'mean_token_accuracy': '0.7482', 'epoch': '0.3981'}
+{'loss': '2.264', 'grad_norm': '31.5', 'learning_rate': '6.019e-06', 'num_tokens': '1.509e+06', 'mean_token_accuracy': '0.5548', 'epoch': '0.3982'}
+{'loss': '2.407', 'grad_norm': '53.25', 'learning_rate': '6.018e-06', 'num_tokens': '1.509e+06', 'mean_token_accuracy': '0.6083', 'epoch': '0.3984'}
+{'loss': '0.6846', 'grad_norm': '9.438', 'learning_rate': '6.016e-06', 'num_tokens': '1.51e+06', 'mean_token_accuracy': '0.7886', 'epoch': '0.3986'}
+{'loss': '1.139', 'grad_norm': '35', 'learning_rate': '6.014e-06', 'num_tokens': '1.51e+06', 'mean_token_accuracy': '0.7349', 'epoch': '0.3987'}
+{'loss': '0.8611', 'grad_norm': '22.12', 'learning_rate': '6.013e-06', 'num_tokens': '1.51e+06', 'mean_token_accuracy': '0.7579', 'epoch': '0.3989'}
+{'loss': '1.13', 'grad_norm': '27.88', 'learning_rate': '6.011e-06', 'num_tokens': '1.511e+06', 'mean_token_accuracy': '0.8095', 'epoch': '0.3991'}
+{'loss': '1.528', 'grad_norm': '15.94', 'learning_rate': '6.009e-06', 'num_tokens': '1.511e+06', 'mean_token_accuracy': '0.6542', 'epoch': '0.3993'}
+{'loss': '0.9798', 'grad_norm': '12.56', 'learning_rate': '6.007e-06', 'num_tokens': '1.512e+06', 'mean_token_accuracy': '0.7077', 'epoch': '0.3994'}
+{'loss': '1.105', 'grad_norm': '12.44', 'learning_rate': '6.006e-06', 'num_tokens': '1.512e+06', 'mean_token_accuracy': '0.6943', 'epoch': '0.3996'}
+{'loss': '1.394', 'grad_norm': '14.19', 'learning_rate': '6.004e-06', 'num_tokens': '1.513e+06', 'mean_token_accuracy': '0.6269', 'epoch': '0.3998'}
+{'loss': '1.271', 'grad_norm': '18', 'learning_rate': '6.002e-06', 'num_tokens': '1.513e+06', 'mean_token_accuracy': '0.7293', 'epoch': '0.4'}
+{'loss': '2.054', 'grad_norm': '14.06', 'learning_rate': '6e-06', 'num_tokens': '1.514e+06', 'mean_token_accuracy': '0.5769', 'epoch': '0.4001'}
+{'loss': '1.547', 'grad_norm': '16', 'learning_rate': '5.999e-06', 'num_tokens': '1.515e+06', 'mean_token_accuracy': '0.6461', 'epoch': '0.4003'}
+{'loss': '1.641', 'grad_norm': '10.81', 'learning_rate': '5.997e-06', 'num_tokens': '1.515e+06', 'mean_token_accuracy': '0.6347', 'epoch': '0.4005'}
+{'loss': '1.425', 'grad_norm': '39.5', 'learning_rate': '5.995e-06', 'num_tokens': '1.516e+06', 'mean_token_accuracy': '0.6957', 'epoch': '0.4007'}
+{'loss': '1.048', 'grad_norm': '5.875', 'learning_rate': '5.993e-06', 'num_tokens': '1.518e+06', 'mean_token_accuracy': '0.7306', 'epoch': '0.4008'}
+{'loss': '1.194', 'grad_norm': '13.75', 'learning_rate': '5.992e-06', 'num_tokens': '1.518e+06', 'mean_token_accuracy': '0.6813', 'epoch': '0.401'}
+{'loss': '2.144', 'grad_norm': '42', 'learning_rate': '5.99e-06', 'num_tokens': '1.519e+06', 'mean_token_accuracy': '0.5876', 'epoch': '0.4012'}
+{'loss': '2.068', 'grad_norm': '11.88', 'learning_rate': '5.988e-06', 'num_tokens': '1.519e+06', 'mean_token_accuracy': '0.5616', 'epoch': '0.4014'}
+{'loss': '1.159', 'grad_norm': '6.562', 'learning_rate': '5.986e-06', 'num_tokens': '1.521e+06', 'mean_token_accuracy': '0.7206', 'epoch': '0.4015'}
+{'loss': '1.586', 'grad_norm': '32.75', 'learning_rate': '5.985e-06', 'num_tokens': '1.521e+06', 'mean_token_accuracy': '0.7528', 'epoch': '0.4017'}
+{'loss': '1.257', 'grad_norm': '13.94', 'learning_rate': '5.983e-06', 'num_tokens': '1.522e+06', 'mean_token_accuracy': '0.634', 'epoch': '0.4019'}
+{'loss': '0.694', 'grad_norm': '16', 'learning_rate': '5.981e-06', 'num_tokens': '1.522e+06', 'mean_token_accuracy': '0.8196', 'epoch': '0.402'}
+{'loss': '1.097', 'grad_norm': '7.312', 'learning_rate': '5.98e-06', 'num_tokens': '1.524e+06', 'mean_token_accuracy': '0.7505', 'epoch': '0.4022'}
+{'loss': '1.156', 'grad_norm': '12.88', 'learning_rate': '5.978e-06', 'num_tokens': '1.524e+06', 'mean_token_accuracy': '0.6898', 'epoch': '0.4024'}
+{'loss': '2.361', 'grad_norm': '11.25', 'learning_rate': '5.976e-06', 'num_tokens': '1.525e+06', 'mean_token_accuracy': '0.4895', 'epoch': '0.4026'}
+{'loss': '1.47', 'grad_norm': '13.12', 'learning_rate': '5.974e-06', 'num_tokens': '1.526e+06', 'mean_token_accuracy': '0.653', 'epoch': '0.4027'}
+{'loss': '0.8724', 'grad_norm': '20.5', 'learning_rate': '5.973e-06', 'num_tokens': '1.526e+06', 'mean_token_accuracy': '0.802', 'epoch': '0.4029'}
+{'loss': '2.128', 'grad_norm': '17.5', 'learning_rate': '5.971e-06', 'num_tokens': '1.527e+06', 'mean_token_accuracy': '0.5889', 'epoch': '0.4031'}
+{'loss': '0.8466', 'grad_norm': '20.12', 'learning_rate': '5.969e-06', 'num_tokens': '1.527e+06', 'mean_token_accuracy': '0.8609', 'epoch': '0.4033'}
+{'loss': '1.644', 'grad_norm': '12.19', 'learning_rate': '5.967e-06', 'num_tokens': '1.527e+06', 'mean_token_accuracy': '0.6481', 'epoch': '0.4034'}
+{'loss': '1.299', 'grad_norm': '13.75', 'learning_rate': '5.966e-06', 'num_tokens': '1.528e+06', 'mean_token_accuracy': '0.6291', 'epoch': '0.4036'}
+{'loss': '1.72', 'grad_norm': '12.88', 'learning_rate': '5.964e-06', 'num_tokens': '1.529e+06', 'mean_token_accuracy': '0.6585', 'epoch': '0.4038'}
+{'loss': '1.973', 'grad_norm': '15.56', 'learning_rate': '5.962e-06', 'num_tokens': '1.529e+06', 'mean_token_accuracy': '0.5787', 'epoch': '0.404'}
+{'loss': '1.159', 'grad_norm': '11.69', 'learning_rate': '5.96e-06', 'num_tokens': '1.53e+06', 'mean_token_accuracy': '0.6756', 'epoch': '0.4041'}
+{'loss': '1.182', 'grad_norm': '12.81', 'learning_rate': '5.959e-06', 'num_tokens': '1.531e+06', 'mean_token_accuracy': '0.6743', 'epoch': '0.4043'}
+{'loss': '1.642', 'grad_norm': '20.38', 'learning_rate': '5.957e-06', 'num_tokens': '1.531e+06', 'mean_token_accuracy': '0.6569', 'epoch': '0.4045'}
+{'loss': '1.299', 'grad_norm': '14.75', 'learning_rate': '5.955e-06', 'num_tokens': '1.531e+06', 'mean_token_accuracy': '0.7283', 'epoch': '0.4047'}
+
 40%|████      | 2331/5758 [13:15<18:46,  3.04it/s]
 41%|████      | 2332/5758 [13:15<18:22,  3.11it/s]
                                                   

 41%|████      | 2332/5758 [13:15<18:22,  3.11it/s]
 41%|████      | 2333/5758 [13:16<18:03,  3.16it/s]
                                                   

 41%|████      | 2333/5758 [13:16<18:03,  3.16it/s]
 41%|████      | 2334/5758 [13:16<17:43,  3.22it/s]
                                                   

 41%|████      | 2334/5758 [13:16<17:43,  3.22it/s]
 41%|████      | 2335/5758 [13:16<17:43,  3.22it/s]
                                                   

 41%|████      | 2335/5758 [13:16<17:43,  3.22it/s]
 41%|███���      | 2336/5758 [13:17<18:12,  3.13it/s]
                                                   

 41%|████      | 2336/5758 [13:17<18:12,  3.13it/s]
 41%|████      | 2337/5758 [13:17<20:09,  2.83it/s]
                                                   

 41%|████      | 2337/5758 [13:17<20:09,  2.83it/s]
 41%|████      | 2338/5758 [13:17<20:09,  2.83it/s]
                                                   

 41%|████      | 2338/5758 [13:17<20:09,  2.83it/s]
 41%|████      | 2339/5758 [13:18<19:49,  2.87it/s]
                                                   

 41%|████      | 2339/5758 [13:18<19:49,  2.87it/s]
 41%|████      | 2340/5758 [13:18<18:55,  3.01it/s]
                                                   

 41%|████      | 2340/5758 [13:18<18:55,  3.01it/s]
 41%|████      | 2341/5758 [13:18<18:29,  3.08it/s]
                                                   

 41%|████      | 2341/5758 [13:18<18:29,  3.08it/s]
 41%|████      | 2342/5758 [13:19<18:49,  3.02it/s]
                                                   

 41%|████      | 2342/5758 [13:19<18:49,  3.02it/s]
 41%|████      | 2343/5758 [13:19<18:43,  3.04it/s]
                                                   

 41%|████      | 2343/5758 [13:19<18:43,  3.04it/s]
 41%|████      | 2344/5758 [13:19<18:14,  3.12it/s]
                                                   

 41%|████      | 2344/5758 [13:19<18:14,  3.12it/s]
 41%|████      | 2345/5758 [13:20<17:46,  3.20it/s]
                                                   

 41%|████      | 2345/5758 [13:20<17:46,  3.20it/s]
 41%|████      | 2346/5758 [13:20<17:49,  3.19it/s]
                                                   

 41%|████      | 2346/5758 [13:20<17:49,  3.19it/s]
 41%|████      | 2347/5758 [13:20<17:52,  3.18it/s]
                                                   

 41%|████      | 2347/5758 [13:20<17:52,  3.18it/s]
 41%|████      | 2348/5758 [13:20<17:29,  3.25it/s]
                                                   

 41%|████      | 2348/5758 [13:20<17:29,  3.25it/s]
 41%|████      | 2349/5758 [13:21<18:21,  3.10it/s]
                                                   

 41%|████      | 2349/5758 [13:21<18:21,  3.10it/s]
 41%|████      | 2350/5758 [13:21<18:54,  3.00it/s]
                                                   

 41%|████      | 2350/5758 [13:21<18:54,  3.00it/s]
 41%|████      | 2351/5758 [13:22<18:54,  3.00it/s]
                                                   

 41%|████      | 2351/5758 [13:22<18:54,  3.00it/s]
 41%|████      | 2352/5758 [13:22<18:12,  3.12it/s]
                                                   

 41%|████      | 2352/5758 [13:22<18:12,  3.12it/s]
 41%|████      | 2353/5758 [13:22<18:25,  3.08it/s]
                                                   

 41%|████      | 2353/5758 [13:22<18:25,  3.08it/s]
 41%|████      | 2354/5758 [13:22<18:29,  3.07it/s]
                                                   

 41%|████      | 2354/5758 [13:22<18:29,  3.07it/s]
 41%|████      | 2355/5758 [13:23<18:18,  3.10it/s]
                                                   

 41%|████      | 2355/5758 [13:23<18:18,  3.10it/s]
 41%|████      | 2356/5758 [13:23<17:44,  3.20it/s]
                                                   

 41%|████      | 2356/5758 [13:23<17:44,  3.20it/s]
 41%|████      | 2357/5758 [13:23<17:59,  3.15it/s]
                                                   

 41%|████      | 2357/5758 [13:23<17:59,  3.15it/s]
 41%|████      | 2358/5758 [13:24<18:02,  3.14it/s]
                                                   

 41%|████      | 2358/5758 [13:24<18:02,  3.14it/s]
 41%|████      | 2359/5758 [13:24<17:37,  3.22it/s]
                                                   

 41%|████      | 2359/5758 [13:24<17:37,  3.22it/s]
 41%|████      | 2360/5758 [13:24<17:09,  3.30it/s]
                                                   

 41%|████      | 2360/5758 [13:24<17:09,  3.30it/s]
 41%|████      | 2361/5758 [13:25<18:29,  3.06it/s]
                                                   

 41%|████      | 2361/5758 [13:25<18:29,  3.06it/s]
 41%|████      | 2362/5758 [13:25<18:06,  3.13it/s]
                                                   

 41%|████      | 2362/5758 [13:25<18:06,  3.13it/s]
 41%|████      | 2363/5758 [13:25<19:41,  2.87it/s]
                                                   

 41%|████      | 2363/5758 [13:25<19:41,  2.87it/s]
 41%|████      | 2364/5758 [13:26<19:20,  2.92it/s]
                                                   

 41%|████      | 2364/5758 [13:26<19:20,  2.92it/s]
 41%|████      | 2365/5758 [13:26<18:43,  3.02it/s]
                                                   

 41%|████      | 2365/5758 [13:26<18:43,  3.02it/s]
 41%|████      | 2366/5758 [13:26<18:23,  3.07it/s]
                                                   

 41%|████      | 2366/5758 [13:26<18:23,  3.07it/s]
 41%|████      | 2367/5758 [13:27<18:16,  3.09it/s]
                                                   

 41%|████      | 2367/5758 [13:27<18:16,  3.09it/s]
 41%|████      | 2368/5758 [13:27<18:33,  3.04it/s]
                                                   

 41%|████      | 2368/5758 [13:27<18:33,  3.04it/s]
 41%|████      | 2369/5758 [13:27<18:35,  3.04it/s]
                                                   

 41%|████      | 2369/5758 [13:27<18:35,  3.04it/s]
 41%|████      | 2370/5758 [13:28<17:53,  3.16it/s]
                                                   

 41%|████      | 2370/5758 [13:28<17:53,  3.16it/s]
 41%|████      | 2371/5758 [13:28<18:26,  3.06it/s]
                                                   

 41%|████      | 2371/5758 [13:28<18:26,  3.06it/s]
 41%|████      | 2372/5758 [13:28<19:19,  2.92it/s]
                                                   

 41%|████      | 2372/5758 [13:28<19:19,  2.92it/s]
 41%|████      | 2373/5758 [13:29<19:21,  2.92it/s]
                                                   

 41%|████      | 2373/5758 [13:29<19:21,  2.92it/s]
 41%|████      | 2374/5758 [13:29<19:36,  2.88it/s]
                                                   

 41%|████      | 2374/5758 [13:29<19:36,  2.88it/s]
 41%|████      | 2375/5758 [13:29<18:39,  3.02it/s]
                                                   

 41%|████      | 2375/5758 [13:29<18:39,  3.02it/s]
 41%|████▏     | 2376/5758 [13:30<18:03,  3.12it/s]
                                                   

 41%|████▏     | 2376/5758 [13:30<18:03,  3.12it/s]
 41%|████▏     | 2377/5758 [13:30<17:37,  3.20it/s]
                                                   

 41%|████▏     | 2377/5758 [13:30<17:37,  3.20it/s]
 41%|████▏     | 2378/5758 [13:30<18:50,  2.99it/s]
                                                   

 41%|████▏     | 2378/5758 [13:30<18:50,  2.99it/s]
 41%|████▏     | 2379/5758 [13:31<19:04,  2.95it/s]
                                                   

 41%|████▏     | 2379/5758 [13:31<19:04,  2.95it/s]
 41%|████▏     | 2380/5758 [13:31<18:36,  3.03it/s]
                                                   

 41%|████▏     | 2380/5758 [13:31<18:36,  3.03it/s]
 41%|████▏     | 2381/5758 [13:31<18:04,  3.11it/s]
                                                   

 41%|████▏     | 2381/5758 [13:31<18:04,  3.11it/s]
 41%|████▏     | 2382/5758 [13:32<17:38,  3.19it/s]
                                                   

 41%|████▏     | 2382/5758 [13:32<17:38,  3.19it/s]
 41%|████▏     | 2383/5758 [13:32<17:16,  3.25it/s]
                                                   

 41%|████▏     | 2383/5758 [13:32<17:16,  3.25it/s]
 41%|████▏     | 2384/5758 [13:32<18:24,  3.05it/s]
                                                   

 41%|████▏     | 2384/5758 [13:32<18:24,  3.05it/s]
 41%|████▏     | 2385/5758 [13:33<19:00,  2.96it/s]
                                                   

 41%|████▏     | 2385/5758 [13:33<19:00,  2.96it/s]
 41%|████▏     | 2386/5758 [13:33<18:20,  3.06it/s]
                                                   
{'loss': '1.117', 'grad_norm': '11.81', 'learning_rate': '5.953e-06', 'num_tokens': '1.532e+06', 'mean_token_accuracy': '0.6954', 'epoch': '0.4048'}
+{'loss': '1.355', 'grad_norm': '13.62', 'learning_rate': '5.952e-06', 'num_tokens': '1.533e+06', 'mean_token_accuracy': '0.6442', 'epoch': '0.405'}
+{'loss': '1.188', 'grad_norm': '9.438', 'learning_rate': '5.95e-06', 'num_tokens': '1.534e+06', 'mean_token_accuracy': '0.6756', 'epoch': '0.4052'}
+{'loss': '1.68', 'grad_norm': '12.56', 'learning_rate': '5.948e-06', 'num_tokens': '1.535e+06', 'mean_token_accuracy': '0.6434', 'epoch': '0.4053'}
+{'loss': '1.386', 'grad_norm': '17.38', 'learning_rate': '5.947e-06', 'num_tokens': '1.535e+06', 'mean_token_accuracy': '0.7364', 'epoch': '0.4055'}
+{'loss': '1.615', 'grad_norm': '12.62', 'learning_rate': '5.945e-06', 'num_tokens': '1.536e+06', 'mean_token_accuracy': '0.5918', 'epoch': '0.4057'}
+{'loss': '0.6079', 'grad_norm': '16.38', 'learning_rate': '5.943e-06', 'num_tokens': '1.536e+06', 'mean_token_accuracy': '0.8698', 'epoch': '0.4059'}
+{'loss': '1.125', 'grad_norm': '16.25', 'learning_rate': '5.941e-06', 'num_tokens': '1.536e+06', 'mean_token_accuracy': '0.7552', 'epoch': '0.406'}
+{'loss': '1.561', 'grad_norm': '12.62', 'learning_rate': '5.94e-06', 'num_tokens': '1.537e+06', 'mean_token_accuracy': '0.5973', 'epoch': '0.4062'}
+{'loss': '0.6205', 'grad_norm': '20.25', 'learning_rate': '5.938e-06', 'num_tokens': '1.538e+06', 'mean_token_accuracy': '0.8663', 'epoch': '0.4064'}
+{'loss': '1.217', 'grad_norm': '9.125', 'learning_rate': '5.936e-06', 'num_tokens': '1.539e+06', 'mean_token_accuracy': '0.7091', 'epoch': '0.4066'}
+{'loss': '1.81', 'grad_norm': '7.531', 'learning_rate': '5.934e-06', 'num_tokens': '1.541e+06', 'mean_token_accuracy': '0.6524', 'epoch': '0.4067'}
+{'loss': '1.17', 'grad_norm': '25.38', 'learning_rate': '5.933e-06', 'num_tokens': '1.541e+06', 'mean_token_accuracy': '0.7512', 'epoch': '0.4069'}
+{'loss': '0.7864', 'grad_norm': '8.375', 'learning_rate': '5.931e-06', 'num_tokens': '1.542e+06', 'mean_token_accuracy': '0.7771', 'epoch': '0.4071'}
+{'loss': '0.9841', 'grad_norm': '12.19', 'learning_rate': '5.929e-06', 'num_tokens': '1.543e+06', 'mean_token_accuracy': '0.7107', 'epoch': '0.4073'}
+{'loss': '2.058', 'grad_norm': '19.5', 'learning_rate': '5.927e-06', 'num_tokens': '1.543e+06', 'mean_token_accuracy': '0.6132', 'epoch': '0.4074'}
+{'loss': '1.161', 'grad_norm': '11.62', 'learning_rate': '5.926e-06', 'num_tokens': '1.544e+06', 'mean_token_accuracy': '0.664', 'epoch': '0.4076'}
+{'loss': '0.592', 'grad_norm': '8.188', 'learning_rate': '5.924e-06', 'num_tokens': '1.545e+06', 'mean_token_accuracy': '0.8224', 'epoch': '0.4078'}
+{'loss': '1.091', 'grad_norm': '21.25', 'learning_rate': '5.922e-06', 'num_tokens': '1.546e+06', 'mean_token_accuracy': '0.7259', 'epoch': '0.408'}
+{'loss': '1.851', 'grad_norm': '32.5', 'learning_rate': '5.92e-06', 'num_tokens': '1.546e+06', 'mean_token_accuracy': '0.6579', 'epoch': '0.4081'}
+{'loss': '2.043', 'grad_norm': '22.12', 'learning_rate': '5.919e-06', 'num_tokens': '1.546e+06', 'mean_token_accuracy': '0.5946', 'epoch': '0.4083'}
+{'loss': '1.032', 'grad_norm': '12.75', 'learning_rate': '5.917e-06', 'num_tokens': '1.547e+06', 'mean_token_accuracy': '0.7136', 'epoch': '0.4085'}
+{'loss': '1.847', 'grad_norm': '16.75', 'learning_rate': '5.915e-06', 'num_tokens': '1.547e+06', 'mean_token_accuracy': '0.6406', 'epoch': '0.4086'}
+{'loss': '0.7936', 'grad_norm': '16.88', 'learning_rate': '5.914e-06', 'num_tokens': '1.547e+06', 'mean_token_accuracy': '0.8066', 'epoch': '0.4088'}
+{'loss': '1.222', 'grad_norm': '11.94', 'learning_rate': '5.912e-06', 'num_tokens': '1.548e+06', 'mean_token_accuracy': '0.6618', 'epoch': '0.409'}
+{'loss': '1.028', 'grad_norm': '12.44', 'learning_rate': '5.91e-06', 'num_tokens': '1.549e+06', 'mean_token_accuracy': '0.7154', 'epoch': '0.4092'}
+{'loss': '1.991', 'grad_norm': '17.75', 'learning_rate': '5.908e-06', 'num_tokens': '1.549e+06', 'mean_token_accuracy': '0.6048', 'epoch': '0.4093'}
+{'loss': '0.6148', 'grad_norm': '7.594', 'learning_rate': '5.907e-06', 'num_tokens': '1.551e+06', 'mean_token_accuracy': '0.814', 'epoch': '0.4095'}
+{'loss': '1.451', 'grad_norm': '13.38', 'learning_rate': '5.905e-06', 'num_tokens': '1.551e+06', 'mean_token_accuracy': '0.6385', 'epoch': '0.4097'}
+{'loss': '1.19', 'grad_norm': '14.44', 'learning_rate': '5.903e-06', 'num_tokens': '1.552e+06', 'mean_token_accuracy': '0.6433', 'epoch': '0.4099'}
+{'loss': '2.04', 'grad_norm': '17.75', 'learning_rate': '5.901e-06', 'num_tokens': '1.552e+06', 'mean_token_accuracy': '0.5629', 'epoch': '0.41'}
+{'loss': '1.373', 'grad_norm': '12.62', 'learning_rate': '5.9e-06', 'num_tokens': '1.553e+06', 'mean_token_accuracy': '0.6264', 'epoch': '0.4102'}
+{'loss': '1.345', 'grad_norm': '13.38', 'learning_rate': '5.898e-06', 'num_tokens': '1.554e+06', 'mean_token_accuracy': '0.6561', 'epoch': '0.4104'}
+{'loss': '0.7362', 'grad_norm': '15.5', 'learning_rate': '5.896e-06', 'num_tokens': '1.554e+06', 'mean_token_accuracy': '0.828', 'epoch': '0.4106'}
+{'loss': '0.5903', 'grad_norm': '15.06', 'learning_rate': '5.894e-06', 'num_tokens': '1.555e+06', 'mean_token_accuracy': '0.8242', 'epoch': '0.4107'}
+{'loss': '1.51', 'grad_norm': '12.94', 'learning_rate': '5.893e-06', 'num_tokens': '1.555e+06', 'mean_token_accuracy': '0.616', 'epoch': '0.4109'}
+{'loss': '2.417', 'grad_norm': '50', 'learning_rate': '5.891e-06', 'num_tokens': '1.556e+06', 'mean_token_accuracy': '0.5761', 'epoch': '0.4111'}
+{'loss': '1.601', 'grad_norm': '17.62', 'learning_rate': '5.889e-06', 'num_tokens': '1.556e+06', 'mean_token_accuracy': '0.6581', 'epoch': '0.4113'}
+{'loss': '1.811', 'grad_norm': '18', 'learning_rate': '5.887e-06', 'num_tokens': '1.556e+06', 'mean_token_accuracy': '0.6185', 'epoch': '0.4114'}
+{'loss': '1.195', 'grad_norm': '12.75', 'learning_rate': '5.886e-06', 'num_tokens': '1.557e+06', 'mean_token_accuracy': '0.6728', 'epoch': '0.4116'}
+{'loss': '0.7298', 'grad_norm': '29.75', 'learning_rate': '5.884e-06', 'num_tokens': '1.557e+06', 'mean_token_accuracy': '0.8421', 'epoch': '0.4118'}
+{'loss': '0.6477', 'grad_norm': '9.438', 'learning_rate': '5.882e-06', 'num_tokens': '1.558e+06', 'mean_token_accuracy': '0.8102', 'epoch': '0.4119'}
+{'loss': '1.321', 'grad_norm': '14.31', 'learning_rate': '5.881e-06', 'num_tokens': '1.559e+06', 'mean_token_accuracy': '0.6909', 'epoch': '0.4121'}
+{'loss': '0.5127', 'grad_norm': '6.094', 'learning_rate': '5.879e-06', 'num_tokens': '1.561e+06', 'mean_token_accuracy': '0.8445', 'epoch': '0.4123'}
+{'loss': '0.5577', 'grad_norm': '7.312', 'learning_rate': '5.877e-06', 'num_tokens': '1.563e+06', 'mean_token_accuracy': '0.8287', 'epoch': '0.4125'}
+{'loss': '1.273', 'grad_norm': '13.69', 'learning_rate': '5.875e-06', 'num_tokens': '1.563e+06', 'mean_token_accuracy': '0.6572', 'epoch': '0.4126'}
+{'loss': '1.946', 'grad_norm': '16.75', 'learning_rate': '5.874e-06', 'num_tokens': '1.564e+06', 'mean_token_accuracy': '0.6078', 'epoch': '0.4128'}
+{'loss': '0.7035', 'grad_norm': '8.375', 'learning_rate': '5.872e-06', 'num_tokens': '1.565e+06', 'mean_token_accuracy': '0.7973', 'epoch': '0.413'}
+{'loss': '1.367', 'grad_norm': '14.81', 'learning_rate': '5.87e-06', 'num_tokens': '1.566e+06', 'mean_token_accuracy': '0.6401', 'epoch': '0.4132'}
+{'loss': '1.077', 'grad_norm': '7.125', 'learning_rate': '5.868e-06', 'num_tokens': '1.567e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.4133'}
+{'loss': '0.571', 'grad_norm': '6.781', 'learning_rate': '5.867e-06', 'num_tokens': '1.569e+06', 'mean_token_accuracy': '0.8242', 'epoch': '0.4135'}
+{'loss': '0.5251', 'grad_norm': '43.75', 'learning_rate': '5.865e-06', 'num_tokens': '1.569e+06', 'mean_token_accuracy': '0.8442', 'epoch': '0.4137'}
+{'loss': '1.526', 'grad_norm': '8', 'learning_rate': '5.863e-06', 'num_tokens': '1.57e+06', 'mean_token_accuracy': '0.6753', 'epoch': '0.4139'}
+{'loss': '2.401', 'grad_norm': '10.75', 'learning_rate': '5.861e-06', 'num_tokens': '1.571e+06', 'mean_token_accuracy': '0.5286', 'epoch': '0.414'}
+{'loss': '1.425', 'grad_norm': '7.438', 'learning_rate': '5.86e-06', 'num_tokens': '1.573e+06', 'mean_token_accuracy': '0.689', 'epoch': '0.4142'}
+
 41%|████▏     | 2386/5758 [13:33<18:20,  3.06it/s]
 41%|████▏     | 2387/5758 [13:33<18:54,  2.97it/s]
                                                   

 41%|████▏     | 2387/5758 [13:33<18:54,  2.97it/s]
 41%|████▏     | 2388/5758 [13:34<18:13,  3.08it/s]
                                                   

 41%|████▏     | 2388/5758 [13:34<18:13,  3.08it/s]
 41%|████▏     | 2389/5758 [13:34<19:41,  2.85it/s]
                                                   

 41%|████▏     | 2389/5758 [13:34<19:41,  2.85it/s]
 42%|████▏     | 2390/5758 [13:34<19:28,  2.88it/s]
                                                   

 42%|████▏     | 2390/5758 [13:34<19:28,  2.88it/s]
 42%|████▏     | 2391/5758 [13:35<19:24,  2.89it/s]
                                                   

 42%|████▏     | 2391/5758 [13:35<19:24,  2.89it/s]
 42%|████▏     | 2392/5758 [13:35<19:56,  2.81it/s]
                                                   

 42%|████▏     | 2392/5758 [13:35<19:56,  2.81it/s]
 42%|████▏     | 2393/5758 [13:35<19:14,  2.91it/s]
                                                   

 42%|████▏     | 2393/5758 [13:35<19:14,  2.91it/s]
 42%|████▏     | 2394/5758 [13:36<19:53,  2.82it/s]
                                                   

 42%|████▏     | 2394/5758 [13:36<19:53,  2.82it/s]
 42%|████▏     | 2395/5758 [13:36<20:07,  2.78it/s]
                                                   

 42%|████▏     | 2395/5758 [13:36<20:07,  2.78it/s]
 42%|████▏     | 2396/5758 [13:36<20:04,  2.79it/s]
                                                   

 42%|████▏     | 2396/5758 [13:36<20:04,  2.79it/s]
 42%|████▏     | 2397/5758 [13:37<21:30,  2.61it/s]
                                                   

 42%|████▏     | 2397/5758 [13:37<21:30,  2.61it/s]
 42%|████▏     | 2398/5758 [13:37<19:36,  2.86it/s]
                                                   

 42%|████▏     | 2398/5758 [13:37<19:36,  2.86it/s]
 42%|████▏     | 2399/5758 [13:37<18:57,  2.95it/s]
                                                   

 42%|████▏     | 2399/5758 [13:37<18:57,  2.95it/s]
 42%|████▏     | 2400/5758 [13:38<18:17,  3.06it/s]
                                                   

 42%|████▏     | 2400/5758 [13:38<18:17,  3.06it/s]
 42%|████▏     | 2401/5758 [13:38<19:19,  2.90it/s]
                                                   

 42%|████▏     | 2401/5758 [13:38<19:19,  2.90it/s]
 42%|████▏     | 2402/5758 [13:39<19:59,  2.80it/s]
                                                   

 42%|████▏     | 2402/5758 [13:39<19:59,  2.80it/s]
 42%|████▏     | 2403/5758 [13:39<19:00,  2.94it/s]
                                                   

 42%|████▏     | 2403/5758 [13:39<19:00,  2.94it/s]
 42%|████▏     | 2404/5758 [13:39<18:20,  3.05it/s]
                                                   

 42%|████▏     | 2404/5758 [13:39<18:20,  3.05it/s]
 42%|████▏     | 2405/5758 [13:39<18:08,  3.08it/s]
                                                   

 42%|████▏     | 2405/5758 [13:39<18:08,  3.08it/s]
 42%|████▏     | 2406/5758 [13:40<17:34,  3.18it/s]
                                                   

 42%|████▏     | 2406/5758 [13:40<17:34,  3.18it/s]
 42%|████▏     | 2407/5758 [13:40<18:35,  3.01it/s]
                                                   

 42%|████▏     | 2407/5758 [13:40<18:35,  3.01it/s]
 42%|████▏     | 2408/5758 [13:40<18:30,  3.02it/s]
                                                   

 42%|████▏     | 2408/5758 [13:40<18:30,  3.02it/s]
 42%|████▏     | 2409/5758 [13:41<17:56,  3.11it/s]
                                                   

 42%|████▏     | 2409/5758 [13:41<17:56,  3.11it/s]
 42%|████▏     | 2410/5758 [13:41<17:28,  3.19it/s]
                                                   

 42%|████▏     | 2410/5758 [13:41<17:28,  3.19it/s]
 42%|████▏     | 2411/5758 [13:41<18:17,  3.05it/s]
                                                   

 42%|████▏     | 2411/5758 [13:41<18:17,  3.05it/s]
 42%|████▏     | 2412/5758 [13:42<17:50,  3.13it/s]
                                                   

 42%|████▏     | 2412/5758 [13:42<17:50,  3.13it/s]
 42%|████▏     | 2413/5758 [13:42<17:18,  3.22it/s]
                                                   

 42%|████▏     | 2413/5758 [13:42<17:18,  3.22it/s]
 42%|████▏     | 2414/5758 [13:42<17:07,  3.25it/s]
                                                   

 42%|████▏     | 2414/5758 [13:42<17:07,  3.25it/s]
 42%|████▏     | 2415/5758 [13:43<16:46,  3.32it/s]
                                                   

 42%|████▏     | 2415/5758 [13:43<16:46,  3.32it/s]
 42%|████▏     | 2416/5758 [13:43<17:42,  3.15it/s]
                                                   

 42%|████▏     | 2416/5758 [13:43<17:42,  3.15it/s]
 42%|████▏     | 2417/5758 [13:43<17:10,  3.24it/s]
                                                   

 42%|████▏     | 2417/5758 [13:43<17:10,  3.24it/s]
 42%|████▏     | 2418/5758 [13:44<19:47,  2.81it/s]
                                                   

 42%|████▏     | 2418/5758 [13:44<19:47,  2.81it/s]
 42%|████▏     | 2419/5758 [13:44<18:56,  2.94it/s]
                                                   

 42%|████▏     | 2419/5758 [13:44<18:56,  2.94it/s]
 42%|████▏     | 2420/5758 [13:44<18:55,  2.94it/s]
                                                   

 42%|████▏     | 2420/5758 [13:44<18:55,  2.94it/s]
 42%|████▏     | 2421/5758 [13:45<18:52,  2.95it/s]
                                                   

 42%|████▏     | 2421/5758 [13:45<18:52,  2.95it/s]
 42%|████▏     | 2422/5758 [13:45<18:06,  3.07it/s]
                                                   

 42%|████▏     | 2422/5758 [13:45<18:06,  3.07it/s]
 42%|████▏     | 2423/5758 [13:45<17:36,  3.16it/s]
                                                   

 42%|████▏     | 2423/5758 [13:45<17:36,  3.16it/s]
 42%|████▏     | 2424/5758 [13:46<18:12,  3.05it/s]
                                                   

 42%|████▏     | 2424/5758 [13:46<18:12,  3.05it/s]
 42%|████▏     | 2425/5758 [13:46<17:56,  3.10it/s]
                                                   

 42%|████▏     | 2425/5758 [13:46<17:56,  3.10it/s]
 42%|████▏     | 2426/5758 [13:46<18:01,  3.08it/s]
                                                   

 42%|████▏     | 2426/5758 [13:46<18:01,  3.08it/s]
 42%|████▏     | 2427/5758 [13:47<17:52,  3.11it/s]
                                                   

 42%|████▏     | 2427/5758 [13:47<17:52,  3.11it/s]
 42%|████▏     | 2428/5758 [13:47<18:31,  3.00it/s]
                                                   

 42%|████▏     | 2428/5758 [13:47<18:31,  3.00it/s]
 42%|████▏     | 2429/5758 [13:47<17:56,  3.09it/s]
                                                   

 42%|████▏     | 2429/5758 [13:47<17:56,  3.09it/s]
 42%|████▏     | 2430/5758 [13:48<18:13,  3.04it/s]
                                                   

 42%|████▏     | 2430/5758 [13:48<18:13,  3.04it/s]
 42%|████▏     | 2431/5758 [13:48<18:21,  3.02it/s]
                                                   

 42%|████▏     | 2431/5758 [13:48<18:21,  3.02it/s]
 42%|████▏     | 2432/5758 [13:48<19:36,  2.83it/s]
                                                   

 42%|████▏     | 2432/5758 [13:48<19:36,  2.83it/s]
 42%|████▏     | 2433/5758 [13:49<20:04,  2.76it/s]
                                                   

 42%|████▏     | 2433/5758 [13:49<20:04,  2.76it/s]
 42%|████▏     | 2434/5758 [13:49<19:27,  2.85it/s]
                                                   

 42%|████▏     | 2434/5758 [13:49<19:27,  2.85it/s]
 42%|████▏     | 2435/5758 [13:49<18:24,  3.01it/s]
                                                   

 42%|████▏     | 2435/5758 [13:49<18:24,  3.01it/s]
 42%|████▏     | 2436/5758 [13:50<19:02,  2.91it/s]
                                                   

 42%|████▏     | 2436/5758 [13:50<19:02,  2.91it/s]
 42%|████▏     | 2437/5758 [13:50<18:17,  3.03it/s]
                                                   

 42%|████▏     | 2437/5758 [13:50<18:17,  3.03it/s]
 42%|████▏     | 2438/5758 [13:50<17:45,  3.12it/s]
                                                   

 42%|████▏     | 2438/5758 [13:50<17:45,  3.12it/s]
 42%|████▏     | 2439/5758 [13:51<17:17,  3.20it/s]
                                                   

 42%|████▏     | 2439/5758 [13:51<17:17,  3.20it/s]
 42%|████▏     | 2440/5758 [13:51<17:01,  3.25it/s]
                                                   

 42%|████▏     | 2440/5758 [13:51<17:01,  3.25it/s]
 42%|████▏     | 2441/5758 [13:51<16:52,  3.28it/s]
                                                   
{'loss': '1.464', 'grad_norm': '13.38', 'learning_rate': '5.858e-06', 'num_tokens': '1.574e+06', 'mean_token_accuracy': '0.6647', 'epoch': '0.4144'}
+{'loss': '1.152', 'grad_norm': '14.38', 'learning_rate': '5.856e-06', 'num_tokens': '1.574e+06', 'mean_token_accuracy': '0.6882', 'epoch': '0.4146'}
+{'loss': '1.327', 'grad_norm': '14.06', 'learning_rate': '5.854e-06', 'num_tokens': '1.575e+06', 'mean_token_accuracy': '0.6462', 'epoch': '0.4147'}
+{'loss': '1.706', 'grad_norm': '20', 'learning_rate': '5.853e-06', 'num_tokens': '1.575e+06', 'mean_token_accuracy': '0.6906', 'epoch': '0.4149'}
+{'loss': '1.608', 'grad_norm': '19.75', 'learning_rate': '5.851e-06', 'num_tokens': '1.576e+06', 'mean_token_accuracy': '0.6872', 'epoch': '0.4151'}
+{'loss': '0.895', 'grad_norm': '9.5', 'learning_rate': '5.849e-06', 'num_tokens': '1.577e+06', 'mean_token_accuracy': '0.7791', 'epoch': '0.4152'}
+{'loss': '1.979', 'grad_norm': '18.12', 'learning_rate': '5.848e-06', 'num_tokens': '1.577e+06', 'mean_token_accuracy': '0.6107', 'epoch': '0.4154'}
+{'loss': '1.937', 'grad_norm': '18.5', 'learning_rate': '5.846e-06', 'num_tokens': '1.578e+06', 'mean_token_accuracy': '0.5876', 'epoch': '0.4156'}
+{'loss': '1.964', 'grad_norm': '17.38', 'learning_rate': '5.844e-06', 'num_tokens': '1.578e+06', 'mean_token_accuracy': '0.6', 'epoch': '0.4158'}
+{'loss': '1.373', 'grad_norm': '10.56', 'learning_rate': '5.842e-06', 'num_tokens': '1.579e+06', 'mean_token_accuracy': '0.6877', 'epoch': '0.4159'}
+{'loss': '1.651', 'grad_norm': '22.12', 'learning_rate': '5.841e-06', 'num_tokens': '1.58e+06', 'mean_token_accuracy': '0.698', 'epoch': '0.4161'}
+{'loss': '1.75', 'grad_norm': '5.531', 'learning_rate': '5.839e-06', 'num_tokens': '1.584e+06', 'mean_token_accuracy': '0.6536', 'epoch': '0.4163'}
+{'loss': '0.8479', 'grad_norm': '37.75', 'learning_rate': '5.837e-06', 'num_tokens': '1.584e+06', 'mean_token_accuracy': '0.8448', 'epoch': '0.4165'}
+{'loss': '1.04', 'grad_norm': '14.31', 'learning_rate': '5.835e-06', 'num_tokens': '1.584e+06', 'mean_token_accuracy': '0.7486', 'epoch': '0.4166'}
+{'loss': '1.002', 'grad_norm': '12.75', 'learning_rate': '5.834e-06', 'num_tokens': '1.585e+06', 'mean_token_accuracy': '0.6994', 'epoch': '0.4168'}
+{'loss': '1.744', 'grad_norm': '15.62', 'learning_rate': '5.832e-06', 'num_tokens': '1.586e+06', 'mean_token_accuracy': '0.6254', 'epoch': '0.417'}
+{'loss': '1.045', 'grad_norm': '17.75', 'learning_rate': '5.83e-06', 'num_tokens': '1.586e+06', 'mean_token_accuracy': '0.7416', 'epoch': '0.4172'}
+{'loss': '1.91', 'grad_norm': '12.69', 'learning_rate': '5.828e-06', 'num_tokens': '1.587e+06', 'mean_token_accuracy': '0.5948', 'epoch': '0.4173'}
+{'loss': '0.6075', 'grad_norm': '7.906', 'learning_rate': '5.827e-06', 'num_tokens': '1.588e+06', 'mean_token_accuracy': '0.8174', 'epoch': '0.4175'}
+{'loss': '1.68', 'grad_norm': '30.62', 'learning_rate': '5.825e-06', 'num_tokens': '1.589e+06', 'mean_token_accuracy': '0.651', 'epoch': '0.4177'}
+{'loss': '1.915', 'grad_norm': '13.75', 'learning_rate': '5.823e-06', 'num_tokens': '1.589e+06', 'mean_token_accuracy': '0.5921', 'epoch': '0.4179'}
+{'loss': '1.366', 'grad_norm': '17.88', 'learning_rate': '5.821e-06', 'num_tokens': '1.59e+06', 'mean_token_accuracy': '0.6395', 'epoch': '0.418'}
+{'loss': '1.403', 'grad_norm': '12.88', 'learning_rate': '5.82e-06', 'num_tokens': '1.591e+06', 'mean_token_accuracy': '0.6712', 'epoch': '0.4182'}
+{'loss': '0.8438', 'grad_norm': '39.5', 'learning_rate': '5.818e-06', 'num_tokens': '1.591e+06', 'mean_token_accuracy': '0.7945', 'epoch': '0.4184'}
+{'loss': '1.605', 'grad_norm': '9.25', 'learning_rate': '5.816e-06', 'num_tokens': '1.592e+06', 'mean_token_accuracy': '0.6532', 'epoch': '0.4185'}
+{'loss': '1.17', 'grad_norm': '14.31', 'learning_rate': '5.815e-06', 'num_tokens': '1.592e+06', 'mean_token_accuracy': '0.6952', 'epoch': '0.4187'}
+{'loss': '1.382', 'grad_norm': '27.62', 'learning_rate': '5.813e-06', 'num_tokens': '1.593e+06', 'mean_token_accuracy': '0.7191', 'epoch': '0.4189'}
+{'loss': '2.134', 'grad_norm': '21.62', 'learning_rate': '5.811e-06', 'num_tokens': '1.593e+06', 'mean_token_accuracy': '0.5425', 'epoch': '0.4191'}
+{'loss': '1.863', 'grad_norm': '11.31', 'learning_rate': '5.809e-06', 'num_tokens': '1.594e+06', 'mean_token_accuracy': '0.6049', 'epoch': '0.4192'}
+{'loss': '0.9172', 'grad_norm': '37.75', 'learning_rate': '5.808e-06', 'num_tokens': '1.594e+06', 'mean_token_accuracy': '0.8571', 'epoch': '0.4194'}
+{'loss': '1.315', 'grad_norm': '29.88', 'learning_rate': '5.806e-06', 'num_tokens': '1.594e+06', 'mean_token_accuracy': '0.7699', 'epoch': '0.4196'}
+{'loss': '1.719', 'grad_norm': '16.5', 'learning_rate': '5.804e-06', 'num_tokens': '1.594e+06', 'mean_token_accuracy': '0.615', 'epoch': '0.4198'}
+{'loss': '1.306', 'grad_norm': '23.88', 'learning_rate': '5.802e-06', 'num_tokens': '1.595e+06', 'mean_token_accuracy': '0.7322', 'epoch': '0.4199'}
+{'loss': '0.8018', 'grad_norm': '10.75', 'learning_rate': '5.801e-06', 'num_tokens': '1.595e+06', 'mean_token_accuracy': '0.7666', 'epoch': '0.4201'}
+{'loss': '1.972', 'grad_norm': '10.25', 'learning_rate': '5.799e-06', 'num_tokens': '1.596e+06', 'mean_token_accuracy': '0.6071', 'epoch': '0.4203'}
+{'loss': '0.5902', 'grad_norm': '23.38', 'learning_rate': '5.797e-06', 'num_tokens': '1.596e+06', 'mean_token_accuracy': '0.8694', 'epoch': '0.4205'}
+{'loss': '0.9454', 'grad_norm': '24.5', 'learning_rate': '5.795e-06', 'num_tokens': '1.597e+06', 'mean_token_accuracy': '0.7888', 'epoch': '0.4206'}
+{'loss': '1.06', 'grad_norm': '12.31', 'learning_rate': '5.794e-06', 'num_tokens': '1.597e+06', 'mean_token_accuracy': '0.713', 'epoch': '0.4208'}
+{'loss': '1.034', 'grad_norm': '20.12', 'learning_rate': '5.792e-06', 'num_tokens': '1.598e+06', 'mean_token_accuracy': '0.7967', 'epoch': '0.421'}
+{'loss': '0.5859', 'grad_norm': '7.5', 'learning_rate': '5.79e-06', 'num_tokens': '1.599e+06', 'mean_token_accuracy': '0.8112', 'epoch': '0.4212'}
+{'loss': '1.253', 'grad_norm': '24.5', 'learning_rate': '5.788e-06', 'num_tokens': '1.599e+06', 'mean_token_accuracy': '0.7083', 'epoch': '0.4213'}
+{'loss': '1.382', 'grad_norm': '13.75', 'learning_rate': '5.787e-06', 'num_tokens': '1.6e+06', 'mean_token_accuracy': '0.6206', 'epoch': '0.4215'}
+{'loss': '1.629', 'grad_norm': '15.31', 'learning_rate': '5.785e-06', 'num_tokens': '1.601e+06', 'mean_token_accuracy': '0.6374', 'epoch': '0.4217'}
+{'loss': '1.02', 'grad_norm': '12.75', 'learning_rate': '5.783e-06', 'num_tokens': '1.601e+06', 'mean_token_accuracy': '0.7068', 'epoch': '0.4218'}
+{'loss': '2.227', 'grad_norm': '24', 'learning_rate': '5.782e-06', 'num_tokens': '1.601e+06', 'mean_token_accuracy': '0.5526', 'epoch': '0.422'}
+{'loss': '1.046', 'grad_norm': '13.88', 'learning_rate': '5.78e-06', 'num_tokens': '1.602e+06', 'mean_token_accuracy': '0.7586', 'epoch': '0.4222'}
+{'loss': '0.9082', 'grad_norm': '23.5', 'learning_rate': '5.778e-06', 'num_tokens': '1.602e+06', 'mean_token_accuracy': '0.7973', 'epoch': '0.4224'}
+{'loss': '1.862', 'grad_norm': '16.38', 'learning_rate': '5.776e-06', 'num_tokens': '1.603e+06', 'mean_token_accuracy': '0.609', 'epoch': '0.4225'}
+{'loss': '1.146', 'grad_norm': '11.88', 'learning_rate': '5.775e-06', 'num_tokens': '1.604e+06', 'mean_token_accuracy': '0.6697', 'epoch': '0.4227'}
+{'loss': '1.718', 'grad_norm': '19.62', 'learning_rate': '5.773e-06', 'num_tokens': '1.604e+06', 'mean_token_accuracy': '0.648', 'epoch': '0.4229'}
+{'loss': '1.947', 'grad_norm': '9.312', 'learning_rate': '5.771e-06', 'num_tokens': '1.605e+06', 'mean_token_accuracy': '0.5815', 'epoch': '0.4231'}
+{'loss': '1.403', 'grad_norm': '18.5', 'learning_rate': '5.769e-06', 'num_tokens': '1.605e+06', 'mean_token_accuracy': '0.6905', 'epoch': '0.4232'}
+{'loss': '1.094', 'grad_norm': '13.38', 'learning_rate': '5.768e-06', 'num_tokens': '1.606e+06', 'mean_token_accuracy': '0.6845', 'epoch': '0.4234'}
+{'loss': '0.7101', 'grad_norm': '9.5', 'learning_rate': '5.766e-06', 'num_tokens': '1.607e+06', 'mean_token_accuracy': '0.77', 'epoch': '0.4236'}
+{'loss': '1.11', 'grad_norm': '12.81', 'learning_rate': '5.764e-06', 'num_tokens': '1.608e+06', 'mean_token_accuracy': '0.7202', 'epoch': '0.4238'}
+
 42%|████▏     | 2441/5758 [13:51<16:52,  3.28it/s]
 42%|████▏     | 2442/5758 [13:52<17:49,  3.10it/s]
                                                   

 42%|████▏     | 2442/5758 [13:52<17:49,  3.10it/s]
 42%|████▏     | 2443/5758 [13:52<18:24,  3.00it/s]
                                                   

 42%|████▏     | 2443/5758 [13:52<18:24,  3.00it/s]
 42%|████▏     | 2444/5758 [13:52<20:22,  2.71it/s]
                                                   

 42%|████▏     | 2444/5758 [13:52<20:22,  2.71it/s]
 42%|████▏     | 2445/5758 [13:53<19:23,  2.85it/s]
                                                   

 42%|████▏     | 2445/5758 [13:53<19:23,  2.85it/s]
 42%|████▏     | 2446/5758 [13:53<18:22,  3.00it/s]
                                                   

 42%|████▏     | 2446/5758 [13:53<18:22,  3.00it/s]
 42%|████▏     | 2447/5758 [13:53<18:24,  3.00it/s]
                                                   

 42%|████▏     | 2447/5758 [13:53<18:24,  3.00it/s]
 43%|████▎     | 2448/5758 [13:54<18:53,  2.92it/s]
                                                   

 43%|████▎     | 2448/5758 [13:54<18:53,  2.92it/s]
 43%|████▎     | 2449/5758 [13:54<17:53,  3.08it/s]
                                                   

 43%|████▎     | 2449/5758 [13:54<17:53,  3.08it/s]
 43%|████▎     | 2450/5758 [13:54<18:30,  2.98it/s]
                                                   

 43%|████▎     | 2450/5758 [13:54<18:30,  2.98it/s]
 43%|████▎     | 2451/5758 [13:55<18:07,  3.04it/s]
                                                   

 43%|████▎     | 2451/5758 [13:55<18:07,  3.04it/s]
 43%|████▎     | 2452/5758 [13:55<17:20,  3.18it/s]
                                                   

 43%|████▎     | 2452/5758 [13:55<17:20,  3.18it/s]
 43%|████▎     | 2453/5758 [13:55<18:28,  2.98it/s]
                                                   

 43%|████▎     | 2453/5758 [13:55<18:28,  2.98it/s]
 43%|████▎     | 2454/5758 [13:56<19:28,  2.83it/s]
                                                   

 43%|████▎     | 2454/5758 [13:56<19:28,  2.83it/s]
 43%|████▎     | 2455/5758 [13:56<18:43,  2.94it/s]
                                                   

 43%|████▎     | 2455/5758 [13:56<18:43,  2.94it/s]
 43%|████▎     | 2456/5758 [13:56<19:08,  2.87it/s]
                                                   

 43%|████▎     | 2456/5758 [13:56<19:08,  2.87it/s]
 43%|████▎     | 2457/5758 [13:57<18:11,  3.02it/s]
                                                   

 43%|████▎     | 2457/5758 [13:57<18:11,  3.02it/s]
 43%|████▎     | 2458/5758 [13:57<17:53,  3.07it/s]
                                                   

 43%|████▎     | 2458/5758 [13:57<17:53,  3.07it/s]
 43%|████▎     | 2459/5758 [13:57<18:45,  2.93it/s]
                                                   

 43%|████▎     | 2459/5758 [13:57<18:45,  2.93it/s]
 43%|████▎     | 2460/5758 [13:58<18:17,  3.01it/s]
                                                   

 43%|████▎     | 2460/5758 [13:58<18:17,  3.01it/s]
 43%|████▎     | 2461/5758 [13:58<17:56,  3.06it/s]
                                                   

 43%|████▎     | 2461/5758 [13:58<17:56,  3.06it/s]
 43%|████▎     | 2462/5758 [13:58<17:36,  3.12it/s]
                                                   

 43%|████▎     | 2462/5758 [13:58<17:36,  3.12it/s]
 43%|████▎     | 2463/5758 [13:59<18:05,  3.04it/s]
                                                   

 43%|████▎     | 2463/5758 [13:59<18:05,  3.04it/s]
 43%|████▎     | 2464/5758 [13:59<18:10,  3.02it/s]
                                                   

 43%|████▎     | 2464/5758 [13:59<18:10,  3.02it/s]
 43%|████▎     | 2465/5758 [13:59<18:40,  2.94it/s]
                                                   

 43%|████▎     | 2465/5758 [13:59<18:40,  2.94it/s]
 43%|████▎     | 2466/5758 [14:00<17:58,  3.05it/s]
                                                   

 43%|████▎     | 2466/5758 [14:00<17:58,  3.05it/s]
 43%|████▎     | 2467/5758 [14:00<18:31,  2.96it/s]
                                                   

 43%|████▎     | 2467/5758 [14:00<18:31,  2.96it/s]
 43%|████▎     | 2468/5758 [14:00<17:46,  3.08it/s]
                                                   

 43%|████▎     | 2468/5758 [14:00<17:46,  3.08it/s]
 43%|████▎     | 2469/5758 [14:01<18:26,  2.97it/s]
                                                   

 43%|████▎     | 2469/5758 [14:01<18:26,  2.97it/s]
 43%|████▎     | 2470/5758 [14:01<19:55,  2.75it/s]
                                                   

 43%|████▎     | 2470/5758 [14:01<19:55,  2.75it/s]
 43%|████▎     | 2471/5758 [14:01<20:20,  2.69it/s]
                                                   

 43%|████▎     | 2471/5758 [14:01<20:20,  2.69it/s]
 43%|████▎     | 2472/5758 [14:02<19:40,  2.78it/s]
                                                   

 43%|████▎     | 2472/5758 [14:02<19:40,  2.78it/s]
 43%|████▎     | 2473/5758 [14:02<18:38,  2.94it/s]
                                                   

 43%|████▎     | 2473/5758 [14:02<18:38,  2.94it/s]
 43%|████▎     | 2474/5758 [14:02<18:21,  2.98it/s]
                                                   

 43%|████▎     | 2474/5758 [14:02<18:21,  2.98it/s]
 43%|████▎     | 2475/5758 [14:03<17:35,  3.11it/s]
                                                   

 43%|████▎     | 2475/5758 [14:03<17:35,  3.11it/s]
 43%|████▎     | 2476/5758 [14:03<17:05,  3.20it/s]
                                                   

 43%|████▎     | 2476/5758 [14:03<17:05,  3.20it/s]
 43%|████▎     | 2477/5758 [14:03<17:55,  3.05it/s]
                                                   

 43%|████▎     | 2477/5758 [14:03<17:55,  3.05it/s]
 43%|████▎     | 2478/5758 [14:04<17:26,  3.13it/s]
                                                   

 43%|████▎     | 2478/5758 [14:04<17:26,  3.13it/s]
 43%|████▎     | 2479/5758 [14:04<17:59,  3.04it/s]
                                                   

 43%|████▎     | 2479/5758 [14:04<17:59,  3.04it/s]
 43%|████▎     | 2480/5758 [14:04<18:01,  3.03it/s]
                                                   

 43%|████▎     | 2480/5758 [14:04<18:01,  3.03it/s]
 43%|████▎     | 2481/5758 [14:05<17:59,  3.04it/s]
                                                   

 43%|████▎     | 2481/5758 [14:05<17:59,  3.04it/s]
 43%|████▎     | 2482/5758 [14:05<17:47,  3.07it/s]
                                                   

 43%|████▎     | 2482/5758 [14:05<17:47,  3.07it/s]
 43%|████▎     | 2483/5758 [14:05<17:17,  3.16it/s]
                                                   

 43%|████▎     | 2483/5758 [14:05<17:17,  3.16it/s]
 43%|████▎     | 2484/5758 [14:06<16:54,  3.23it/s]
                                                   

 43%|████▎     | 2484/5758 [14:06<16:54,  3.23it/s]
 43%|████▎     | 2485/5758 [14:06<17:10,  3.18it/s]
                                                   

 43%|████▎     | 2485/5758 [14:06<17:10,  3.18it/s]
 43%|████▎     | 2486/5758 [14:06<17:03,  3.20it/s]
                                                   

 43%|████▎     | 2486/5758 [14:06<17:03,  3.20it/s]
 43%|████▎     | 2487/5758 [14:07<17:50,  3.05it/s]
                                                   

 43%|████▎     | 2487/5758 [14:07<17:50,  3.05it/s]
 43%|████▎     | 2488/5758 [14:07<17:59,  3.03it/s]
                                                   

 43%|████▎     | 2488/5758 [14:07<17:59,  3.03it/s]
 43%|████▎     | 2489/5758 [14:07<18:17,  2.98it/s]
                                                   

 43%|████▎     | 2489/5758 [14:07<18:17,  2.98it/s]
 43%|████▎     | 2490/5758 [14:08<18:03,  3.02it/s]
                                                   

 43%|████▎     | 2490/5758 [14:08<18:03,  3.02it/s]
 43%|████▎     | 2491/5758 [14:08<17:26,  3.12it/s]
                                                   

 43%|████▎     | 2491/5758 [14:08<17:26,  3.12it/s]
 43%|████▎     | 2492/5758 [14:08<18:09,  3.00it/s]
                                                   

 43%|████▎     | 2492/5758 [14:08<18:09,  3.00it/s]
 43%|████▎     | 2493/5758 [14:09<17:43,  3.07it/s]
                                                   

 43%|████▎     | 2493/5758 [14:09<17:43,  3.07it/s]
 43%|████▎     | 2494/5758 [14:09<18:37,  2.92it/s]
                                                   

 43%|████▎     | 2494/5758 [14:09<18:37,  2.92it/s]
 43%|████▎     | 2495/5758 [14:09<19:08,  2.84it/s]
                                                   

 43%|████▎     | 2495/5758 [14:09<19:08,  2.84it/s]
 43%|████▎     | 2496/5758 [14:10<20:36,  2.64it/s]
                                                   
{'loss': '1.483', 'grad_norm': '18.5', 'learning_rate': '5.762e-06', 'num_tokens': '1.608e+06', 'mean_token_accuracy': '0.6855', 'epoch': '0.4239'}
+{'loss': '0.8377', 'grad_norm': '13.62', 'learning_rate': '5.761e-06', 'num_tokens': '1.609e+06', 'mean_token_accuracy': '0.7373', 'epoch': '0.4241'}
+{'loss': '1.802', 'grad_norm': '9.125', 'learning_rate': '5.759e-06', 'num_tokens': '1.61e+06', 'mean_token_accuracy': '0.6088', 'epoch': '0.4243'}
+{'loss': '1.202', 'grad_norm': '23.75', 'learning_rate': '5.757e-06', 'num_tokens': '1.61e+06', 'mean_token_accuracy': '0.7544', 'epoch': '0.4245'}
+{'loss': '2.261', 'grad_norm': '16.75', 'learning_rate': '5.755e-06', 'num_tokens': '1.611e+06', 'mean_token_accuracy': '0.5501', 'epoch': '0.4246'}
+{'loss': '1.832', 'grad_norm': '18.5', 'learning_rate': '5.754e-06', 'num_tokens': '1.611e+06', 'mean_token_accuracy': '0.6221', 'epoch': '0.4248'}
+{'loss': '1.142', 'grad_norm': '9.812', 'learning_rate': '5.752e-06', 'num_tokens': '1.612e+06', 'mean_token_accuracy': '0.7497', 'epoch': '0.425'}
+{'loss': '1.147', 'grad_norm': '6.312', 'learning_rate': '5.75e-06', 'num_tokens': '1.614e+06', 'mean_token_accuracy': '0.7162', 'epoch': '0.4251'}
+{'loss': '1.589', 'grad_norm': '12.5', 'learning_rate': '5.749e-06', 'num_tokens': '1.615e+06', 'mean_token_accuracy': '0.6633', 'epoch': '0.4253'}
+{'loss': '1.218', 'grad_norm': '6.656', 'learning_rate': '5.747e-06', 'num_tokens': '1.617e+06', 'mean_token_accuracy': '0.709', 'epoch': '0.4255'}
+{'loss': '0.7841', 'grad_norm': '13.75', 'learning_rate': '5.745e-06', 'num_tokens': '1.617e+06', 'mean_token_accuracy': '0.8154', 'epoch': '0.4257'}
+{'loss': '0.8638', 'grad_norm': '43.75', 'learning_rate': '5.743e-06', 'num_tokens': '1.617e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.4258'}
+{'loss': '0.5542', 'grad_norm': '19', 'learning_rate': '5.742e-06', 'num_tokens': '1.617e+06', 'mean_token_accuracy': '0.8797', 'epoch': '0.426'}
+{'loss': '0.94', 'grad_norm': '31.75', 'learning_rate': '5.74e-06', 'num_tokens': '1.618e+06', 'mean_token_accuracy': '0.781', 'epoch': '0.4262'}
+{'loss': '0.8528', 'grad_norm': '8.5', 'learning_rate': '5.738e-06', 'num_tokens': '1.619e+06', 'mean_token_accuracy': '0.776', 'epoch': '0.4264'}
+{'loss': '1.696', 'grad_norm': '13.81', 'learning_rate': '5.736e-06', 'num_tokens': '1.62e+06', 'mean_token_accuracy': '0.6331', 'epoch': '0.4265'}
+{'loss': '1.807', 'grad_norm': '19.75', 'learning_rate': '5.735e-06', 'num_tokens': '1.62e+06', 'mean_token_accuracy': '0.65', 'epoch': '0.4267'}
+{'loss': '1.339', 'grad_norm': '13.19', 'learning_rate': '5.733e-06', 'num_tokens': '1.621e+06', 'mean_token_accuracy': '0.6448', 'epoch': '0.4269'}
+{'loss': '2.439', 'grad_norm': '8.688', 'learning_rate': '5.731e-06', 'num_tokens': '1.623e+06', 'mean_token_accuracy': '0.5315', 'epoch': '0.4271'}
+{'loss': '1.074', 'grad_norm': '11.25', 'learning_rate': '5.729e-06', 'num_tokens': '1.623e+06', 'mean_token_accuracy': '0.7038', 'epoch': '0.4272'}
+{'loss': '0.9931', 'grad_norm': '12.19', 'learning_rate': '5.728e-06', 'num_tokens': '1.624e+06', 'mean_token_accuracy': '0.7565', 'epoch': '0.4274'}
+{'loss': '1.199', 'grad_norm': '14.88', 'learning_rate': '5.726e-06', 'num_tokens': '1.625e+06', 'mean_token_accuracy': '0.6876', 'epoch': '0.4276'}
+{'loss': '0.9922', 'grad_norm': '14.88', 'learning_rate': '5.724e-06', 'num_tokens': '1.625e+06', 'mean_token_accuracy': '0.7339', 'epoch': '0.4278'}
+{'loss': '1.907', 'grad_norm': '23.75', 'learning_rate': '5.722e-06', 'num_tokens': '1.626e+06', 'mean_token_accuracy': '0.6402', 'epoch': '0.4279'}
+{'loss': '0.9922', 'grad_norm': '7.969', 'learning_rate': '5.721e-06', 'num_tokens': '1.628e+06', 'mean_token_accuracy': '0.7353', 'epoch': '0.4281'}
+{'loss': '1.785', 'grad_norm': '28.12', 'learning_rate': '5.719e-06', 'num_tokens': '1.628e+06', 'mean_token_accuracy': '0.6784', 'epoch': '0.4283'}
+{'loss': '0.6628', 'grad_norm': '17', 'learning_rate': '5.717e-06', 'num_tokens': '1.628e+06', 'mean_token_accuracy': '0.8411', 'epoch': '0.4284'}
+{'loss': '1.395', 'grad_norm': '13.56', 'learning_rate': '5.716e-06', 'num_tokens': '1.629e+06', 'mean_token_accuracy': '0.6421', 'epoch': '0.4286'}
+{'loss': '0.8154', 'grad_norm': '29', 'learning_rate': '5.714e-06', 'num_tokens': '1.629e+06', 'mean_token_accuracy': '0.819', 'epoch': '0.4288'}
+{'loss': '1.349', 'grad_norm': '13.88', 'learning_rate': '5.712e-06', 'num_tokens': '1.629e+06', 'mean_token_accuracy': '0.6343', 'epoch': '0.429'}
+{'loss': '2.292', 'grad_norm': '37.5', 'learning_rate': '5.71e-06', 'num_tokens': '1.63e+06', 'mean_token_accuracy': '0.6015', 'epoch': '0.4291'}
+{'loss': '1.376', 'grad_norm': '12.12', 'learning_rate': '5.709e-06', 'num_tokens': '1.63e+06', 'mean_token_accuracy': '0.6367', 'epoch': '0.4293'}
+{'loss': '1.503', 'grad_norm': '17.12', 'learning_rate': '5.707e-06', 'num_tokens': '1.631e+06', 'mean_token_accuracy': '0.6585', 'epoch': '0.4295'}
+{'loss': '1.097', 'grad_norm': '8.188', 'learning_rate': '5.705e-06', 'num_tokens': '1.632e+06', 'mean_token_accuracy': '0.7169', 'epoch': '0.4297'}
+{'loss': '2.308', 'grad_norm': '17.38', 'learning_rate': '5.703e-06', 'num_tokens': '1.633e+06', 'mean_token_accuracy': '0.5124', 'epoch': '0.4298'}
+{'loss': '0.7305', 'grad_norm': '18.38', 'learning_rate': '5.702e-06', 'num_tokens': '1.633e+06', 'mean_token_accuracy': '0.8072', 'epoch': '0.43'}
+{'loss': '0.8413', 'grad_norm': '18.25', 'learning_rate': '5.7e-06', 'num_tokens': '1.633e+06', 'mean_token_accuracy': '0.8016', 'epoch': '0.4302'}
+{'loss': '1.83', 'grad_norm': '16.38', 'learning_rate': '5.698e-06', 'num_tokens': '1.634e+06', 'mean_token_accuracy': '0.6135', 'epoch': '0.4304'}
+{'loss': '1.225', 'grad_norm': '23.75', 'learning_rate': '5.696e-06', 'num_tokens': '1.634e+06', 'mean_token_accuracy': '0.6929', 'epoch': '0.4305'}
+{'loss': '0.7973', 'grad_norm': '20.75', 'learning_rate': '5.695e-06', 'num_tokens': '1.635e+06', 'mean_token_accuracy': '0.8128', 'epoch': '0.4307'}
+{'loss': '1.339', 'grad_norm': '25.38', 'learning_rate': '5.693e-06', 'num_tokens': '1.635e+06', 'mean_token_accuracy': '0.7677', 'epoch': '0.4309'}
+{'loss': '1.177', 'grad_norm': '31.12', 'learning_rate': '5.691e-06', 'num_tokens': '1.635e+06', 'mean_token_accuracy': '0.7083', 'epoch': '0.4311'}
+{'loss': '1.798', 'grad_norm': '17', 'learning_rate': '5.689e-06', 'num_tokens': '1.635e+06', 'mean_token_accuracy': '0.6556', 'epoch': '0.4312'}
+{'loss': '1.361', 'grad_norm': '12.94', 'learning_rate': '5.688e-06', 'num_tokens': '1.636e+06', 'mean_token_accuracy': '0.6499', 'epoch': '0.4314'}
+{'loss': '1.018', 'grad_norm': '11.38', 'learning_rate': '5.686e-06', 'num_tokens': '1.637e+06', 'mean_token_accuracy': '0.7202', 'epoch': '0.4316'}
+{'loss': '0.9726', 'grad_norm': '33.5', 'learning_rate': '5.684e-06', 'num_tokens': '1.637e+06', 'mean_token_accuracy': '0.7625', 'epoch': '0.4317'}
+{'loss': '0.9931', 'grad_norm': '19.25', 'learning_rate': '5.683e-06', 'num_tokens': '1.637e+06', 'mean_token_accuracy': '0.788', 'epoch': '0.4319'}
+{'loss': '1.338', 'grad_norm': '10.81', 'learning_rate': '5.681e-06', 'num_tokens': '1.638e+06', 'mean_token_accuracy': '0.6907', 'epoch': '0.4321'}
+{'loss': '1.368', 'grad_norm': '17.62', 'learning_rate': '5.679e-06', 'num_tokens': '1.638e+06', 'mean_token_accuracy': '0.7241', 'epoch': '0.4323'}
+{'loss': '1.817', 'grad_norm': '11.19', 'learning_rate': '5.677e-06', 'num_tokens': '1.639e+06', 'mean_token_accuracy': '0.6305', 'epoch': '0.4324'}
+{'loss': '0.9709', 'grad_norm': '12.75', 'learning_rate': '5.676e-06', 'num_tokens': '1.64e+06', 'mean_token_accuracy': '0.7392', 'epoch': '0.4326'}
+{'loss': '1.985', 'grad_norm': '19.12', 'learning_rate': '5.674e-06', 'num_tokens': '1.64e+06', 'mean_token_accuracy': '0.5781', 'epoch': '0.4328'}
+{'loss': '0.9909', 'grad_norm': '12.31', 'learning_rate': '5.672e-06', 'num_tokens': '1.641e+06', 'mean_token_accuracy': '0.7139', 'epoch': '0.433'}
+{'loss': '1.618', 'grad_norm': '34.75', 'learning_rate': '5.67e-06', 'num_tokens': '1.641e+06', 'mean_token_accuracy': '0.6746', 'epoch': '0.4331'}
+{'loss': '1.848', 'grad_norm': '23.25', 'learning_rate': '5.669e-06', 'num_tokens': '1.641e+06', 'mean_token_accuracy': '0.6381', 'epoch': '0.4333'}
+
 43%|████▎     | 2496/5758 [14:10<20:36,  2.64it/s]
 43%|████▎     | 2497/5758 [14:10<19:17,  2.82it/s]
                                                   

 43%|████▎     | 2497/5758 [14:10<19:17,  2.82it/s]
 43%|████▎     | 2498/5758 [14:10<19:50,  2.74it/s]
                                                   

 43%|████▎     | 2498/5758 [14:10<19:50,  2.74it/s]
 43%|████▎     | 2499/5758 [14:11<19:27,  2.79it/s]
                                                   

 43%|████▎     | 2499/5758 [14:11<19:27,  2.79it/s]
 43%|████▎     | 2500/5758 [14:11<18:30,  2.93it/s]
                                                   

 43%|████▎     | 2500/5758 [14:11<18:30,  2.93it/s]{'loss': '1.905', 'grad_norm': '11.69', 'learning_rate': '5.667e-06', 'num_tokens': '1.642e+06', 'mean_token_accuracy': '0.5948', 'epoch': '0.4335'}
+{'loss': '1.492', 'grad_norm': '15.12', 'learning_rate': '5.665e-06', 'num_tokens': '1.643e+06', 'mean_token_accuracy': '0.6182', 'epoch': '0.4337'}
+{'loss': '1.288', 'grad_norm': '14.56', 'learning_rate': '5.663e-06', 'num_tokens': '1.643e+06', 'mean_token_accuracy': '0.6615', 'epoch': '0.4338'}
+{'loss': '0.9342', 'grad_norm': '10.31', 'learning_rate': '5.662e-06', 'num_tokens': '1.644e+06', 'mean_token_accuracy': '0.7258', 'epoch': '0.434'}
+{'loss': '1.065', 'grad_norm': '8.312', 'learning_rate': '5.66e-06', 'num_tokens': '1.646e+06', 'mean_token_accuracy': '0.7611', 'epoch': '0.4342'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.59it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.59it/s]
+
 43%|████▎     | 2501/5758 [14:13<46:19,  1.17it/s]
                                                   

 43%|████▎     | 2501/5758 [14:13<46:19,  1.17it/s]
 43%|████▎     | 2502/5758 [14:13<37:17,  1.46it/s]
                                                   

 43%|████▎     | 2502/5758 [14:13<37:17,  1.46it/s]
 43%|████▎     | 2503/5758 [14:14<31:15,  1.74it/s]
                                                   

 43%|████▎     | 2503/5758 [14:14<31:15,  1.74it/s]
 43%|████▎     | 2504/5758 [14:14<28:03,  1.93it/s]
                                                   

 43%|████▎     | 2504/5758 [14:14<28:03,  1.93it/s]
 44%|████▎     | 2505/5758 [14:14<25:36,  2.12it/s]
                                                   

 44%|████▎     | 2505/5758 [14:14<25:36,  2.12it/s]
 44%|████▎     | 2506/5758 [14:15<22:52,  2.37it/s]
                                                   

 44%|████▎     | 2506/5758 [14:15<22:52,  2.37it/s]
 44%|████▎     | 2507/5758 [14:15<22:01,  2.46it/s]
                                                   

 44%|████▎     | 2507/5758 [14:15<22:01,  2.46it/s]
 44%|████▎     | 2508/5758 [14:15<20:17,  2.67it/s]
                                                   

 44%|████▎     | 2508/5758 [14:15<20:17,  2.67it/s]
 44%|████▎     | 2509/5758 [14:16<19:48,  2.73it/s]
                                                   

 44%|████▎     | 2509/5758 [14:16<19:48,  2.73it/s]
 44%|████▎     | 2510/5758 [14:16<19:52,  2.72it/s]
                                                   

 44%|████▎     | 2510/5758 [14:16<19:52,  2.72it/s]
 44%|████▎     | 2511/5758 [14:17<19:51,  2.73it/s]
                                                   

 44%|████▎     | 2511/5758 [14:17<19:51,  2.73it/s]
 44%|████▎     | 2512/5758 [14:17<18:46,  2.88it/s]
                                                   

 44%|████▎     | 2512/5758 [14:17<18:46,  2.88it/s]
 44%|████▎     | 2513/5758 [14:17<18:13,  2.97it/s]
                                                   

 44%|████▎     | 2513/5758 [14:17<18:13,  2.97it/s]
 44%|████▎     | 2514/5758 [14:17<18:28,  2.93it/s]
                                                   

 44%|████▎     | 2514/5758 [14:18<18:28,  2.93it/s]
 44%|████▎     | 2515/5758 [14:18<17:58,  3.01it/s]
                                                   

 44%|████▎     | 2515/5758 [14:18<17:58,  3.01it/s]
 44%|████▎     | 2516/5758 [14:18<19:10,  2.82it/s]
                                                   

 44%|████▎     | 2516/5758 [14:18<19:10,  2.82it/s]
 44%|████▎     | 2517/5758 [14:19<19:22,  2.79it/s]
                                                   

 44%|████▎     | 2517/5758 [14:19<19:22,  2.79it/s]
 44%|████▎     | 2518/5758 [14:19<19:28,  2.77it/s]
                                                   

 44%|████▎     | 2518/5758 [14:19<19:28,  2.77it/s]
 44%|████▎     | 2519/5758 [14:19<18:35,  2.90it/s]
                                                   

 44%|████▎     | 2519/5758 [14:19<18:35,  2.90it/s]
 44%|████▍     | 2520/5758 [14:20<17:48,  3.03it/s]
                                                   

 44%|████▍     | 2520/5758 [14:20<17:48,  3.03it/s]
 44%|████▍     | 2521/5758 [14:20<19:29,  2.77it/s]
                                                   

 44%|████▍     | 2521/5758 [14:20<19:29,  2.77it/s]
 44%|████▍     | 2522/5758 [14:20<20:03,  2.69it/s]
                                                   

 44%|████▍     | 2522/5758 [14:20<20:03,  2.69it/s]
 44%|████▍     | 2523/5758 [14:21<19:21,  2.79it/s]
                                                   

 44%|████▍     | 2523/5758 [14:21<19:21,  2.79it/s]
 44%|████▍     | 2524/5758 [14:21<19:10,  2.81it/s]
                                                   

 44%|████▍     | 2524/5758 [14:21<19:10,  2.81it/s]
 44%|████▍     | 2525/5758 [14:21<19:41,  2.74it/s]
                                                   

 44%|████▍     | 2525/5758 [14:21<19:41,  2.74it/s]
 44%|████▍     | 2526/5758 [14:22<18:31,  2.91it/s]
                                                   

 44%|████▍     | 2526/5758 [14:22<18:31,  2.91it/s]
 44%|████▍     | 2527/5758 [14:22<17:35,  3.06it/s]
                                                   

 44%|████▍     | 2527/5758 [14:22<17:35,  3.06it/s]
 44%|████▍     | 2528/5758 [14:22<18:01,  2.99it/s]
                                                   

 44%|████▍     | 2528/5758 [14:22<18:01,  2.99it/s]
 44%|████▍     | 2529/5758 [14:23<17:14,  3.12it/s]
                                                   

 44%|████▍     | 2529/5758 [14:23<17:14,  3.12it/s]
 44%|████▍     | 2530/5758 [14:23<16:22,  3.29it/s]
                                                   

 44%|████▍     | 2530/5758 [14:23<16:22,  3.29it/s]
 44%|████▍     | 2531/5758 [14:23<16:54,  3.18it/s]
                                                   

 44%|████▍     | 2531/5758 [14:23<16:54,  3.18it/s]
 44%|████▍     | 2532/5758 [14:24<17:36,  3.05it/s]
                                                   

 44%|████▍     | 2532/5758 [14:24<17:36,  3.05it/s]
 44%|████▍     | 2533/5758 [14:24<17:11,  3.13it/s]
                                                   

 44%|████▍     | 2533/5758 [14:24<17:11,  3.13it/s]
 44%|████▍     | 2534/5758 [14:24<16:50,  3.19it/s]
                                                   

 44%|████▍     | 2534/5758 [14:24<16:50,  3.19it/s]
 44%|████▍     | 2535/5758 [14:25<17:13,  3.12it/s]
                                                   

 44%|████▍     | 2535/5758 [14:25<17:13,  3.12it/s]
 44%|████▍     | 2536/5758 [14:25<17:46,  3.02it/s]
                                                   

 44%|████▍     | 2536/5758 [14:25<17:46,  3.02it/s]
 44%|████▍     | 2537/5758 [14:25<18:31,  2.90it/s]
                                                   

 44%|████▍     | 2537/5758 [14:25<18:31,  2.90it/s]
 44%|████▍     | 2538/5758 [14:26<17:51,  3.00it/s]
                                                   

 44%|████▍     | 2538/5758 [14:26<17:51,  3.00it/s]
 44%|████▍     | 2539/5758 [14:26<17:30,  3.07it/s]
                                                   

 44%|████▍     | 2539/5758 [14:26<17:30,  3.07it/s]
 44%|████▍     | 2540/5758 [14:26<17:34,  3.05it/s]
                                                   

 44%|████▍     | 2540/5758 [14:26<17:34,  3.05it/s]
 44%|████▍     | 2541/5758 [14:27<17:49,  3.01it/s]
                                                   

 44%|████▍     | 2541/5758 [14:27<17:49,  3.01it/s]
 44%|████▍     | 2542/5758 [14:27<18:36,  2.88it/s]
                                                   

 44%|████▍     | 2542/5758 [14:27<18:36,  2.88it/s]
 44%|████▍     | 2543/5758 [14:27<18:48,  2.85it/s]
                                                   

 44%|████▍     | 2543/5758 [14:27<18:48,  2.85it/s]
 44%|████▍     | 2544/5758 [14:28<18:24,  2.91it/s]
                                                   

 44%|████▍     | 2544/5758 [14:28<18:24,  2.91it/s]
 44%|████▍     | 2545/5758 [14:28<18:18,  2.92it/s]
                                                   

 44%|████▍     | 2545/5758 [14:28<18:18,  2.92it/s]
 44%|████▍     | 2546/5758 [14:28<20:11,  2.65it/s]
                                                   

 44%|████▍     | 2546/5758 [14:28<20:11,  2.65it/s]
 44%|████▍     | 2547/5758 [14:29<18:52,  2.84it/s]
                                                   

 44%|████▍     | 2547/5758 [14:29<18:52,  2.84it/s]
 44%|████▍     | 2548/5758 [14:29<17:56,  2.98it/s]
                                                   

 44%|████▍     | 2548/5758 [14:29<17:56,  2.98it/s]
 44%|████▍     | 2549/5758 [14:29<17:38,  3.03it/s]
                                                   

 44%|████▍     | 2549/5758 [14:29<17:38,  3.03it/s]
 44%|████▍     | 2550/5758 [14:30<16:58,  3.15it/s]
                                                   

 44%|████▍     | 2550/5758 [14:30<16:58,  3.15it/s]
 44%|████▍     | 2551/5758 [14:30<16:47,  3.18it/s]
                                                   

 44%|████▍     | 2551/5758 [14:30<16:47,  3.18it/s]
 44%|████▍     | 2552/5758 [14:30<17:17,  3.09it/s]
                                                   

 44%|████▍     | 2552/5758 [14:30<17:17,  3.09it/s]
 44%|████▍     | 2553/5758 [14:31<17:45,  3.01it/s]
                                                   

 44%|████▍     | 2553/5758 [14:31<17:45,  3.01it/s]
 44%|████▍     | 2554/5758 [14:31<18:39,  2.86it/s]
                                                   

 44%|████▍     | 2554/5758 [14:31<18:39,  2.86it/s]
 44%|████▍     | 2555/5758 [14:31<19:18,  2.77it/s]
                                                   

 44%|████▍     | 2555/5758 [14:31<19:18,  2.77it/s]
 44%|████▍     | 2556/5758 [14:32<18:20,  2.91it/s]
                                                   
{'loss': '1.706', 'grad_norm': '15.81', 'learning_rate': '5.658e-06', 'num_tokens': '1.646e+06', 'mean_token_accuracy': '0.6636', 'epoch': '0.4344'}
+{'loss': '1.538', 'grad_norm': '18.75', 'learning_rate': '5.656e-06', 'num_tokens': '1.646e+06', 'mean_token_accuracy': '0.6908', 'epoch': '0.4345'}
+{'loss': '1.571', 'grad_norm': '19.5', 'learning_rate': '5.655e-06', 'num_tokens': '1.647e+06', 'mean_token_accuracy': '0.6894', 'epoch': '0.4347'}
+{'loss': '1.858', 'grad_norm': '22.5', 'learning_rate': '5.653e-06', 'num_tokens': '1.647e+06', 'mean_token_accuracy': '0.6439', 'epoch': '0.4349'}
+{'loss': '0.8465', 'grad_norm': '13.44', 'learning_rate': '5.651e-06', 'num_tokens': '1.648e+06', 'mean_token_accuracy': '0.8054', 'epoch': '0.435'}
+{'loss': '1.156', 'grad_norm': '11.75', 'learning_rate': '5.65e-06', 'num_tokens': '1.648e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.4352'}
+{'loss': '1.536', 'grad_norm': '30.88', 'learning_rate': '5.648e-06', 'num_tokens': '1.648e+06', 'mean_token_accuracy': '0.5952', 'epoch': '0.4354'}
+{'loss': '1.186', 'grad_norm': '13.19', 'learning_rate': '5.646e-06', 'num_tokens': '1.649e+06', 'mean_token_accuracy': '0.6492', 'epoch': '0.4356'}
+{'loss': '2.305', 'grad_norm': '7.344', 'learning_rate': '5.644e-06', 'num_tokens': '1.651e+06', 'mean_token_accuracy': '0.5583', 'epoch': '0.4357'}
+{'loss': '1.171', 'grad_norm': '14', 'learning_rate': '5.643e-06', 'num_tokens': '1.652e+06', 'mean_token_accuracy': '0.7444', 'epoch': '0.4359'}
+{'loss': '0.7032', 'grad_norm': '11.19', 'learning_rate': '5.641e-06', 'num_tokens': '1.652e+06', 'mean_token_accuracy': '0.8306', 'epoch': '0.4361'}
+{'loss': '0.9032', 'grad_norm': '17', 'learning_rate': '5.639e-06', 'num_tokens': '1.652e+06', 'mean_token_accuracy': '0.7884', 'epoch': '0.4363'}
+{'loss': '1.346', 'grad_norm': '12.5', 'learning_rate': '5.637e-06', 'num_tokens': '1.653e+06', 'mean_token_accuracy': '0.6511', 'epoch': '0.4364'}
+{'loss': '1.635', 'grad_norm': '16.25', 'learning_rate': '5.636e-06', 'num_tokens': '1.654e+06', 'mean_token_accuracy': '0.6759', 'epoch': '0.4366'}
+{'loss': '0.7957', 'grad_norm': '12.88', 'learning_rate': '5.634e-06', 'num_tokens': '1.654e+06', 'mean_token_accuracy': '0.7565', 'epoch': '0.4368'}
+{'loss': '1.651', 'grad_norm': '6.781', 'learning_rate': '5.632e-06', 'num_tokens': '1.657e+06', 'mean_token_accuracy': '0.6469', 'epoch': '0.437'}
+{'loss': '1.395', 'grad_norm': '21.25', 'learning_rate': '5.63e-06', 'num_tokens': '1.657e+06', 'mean_token_accuracy': '0.6784', 'epoch': '0.4371'}
+{'loss': '1.438', 'grad_norm': '31.75', 'learning_rate': '5.629e-06', 'num_tokens': '1.657e+06', 'mean_token_accuracy': '0.7477', 'epoch': '0.4373'}
+{'loss': '1.345', 'grad_norm': '23.38', 'learning_rate': '5.627e-06', 'num_tokens': '1.657e+06', 'mean_token_accuracy': '0.7455', 'epoch': '0.4375'}
+{'loss': '0.5752', 'grad_norm': '7.688', 'learning_rate': '5.625e-06', 'num_tokens': '1.659e+06', 'mean_token_accuracy': '0.8313', 'epoch': '0.4377'}
+{'loss': '1.44', 'grad_norm': '8.5', 'learning_rate': '5.623e-06', 'num_tokens': '1.66e+06', 'mean_token_accuracy': '0.6891', 'epoch': '0.4378'}
+{'loss': '1.209', 'grad_norm': '14.19', 'learning_rate': '5.622e-06', 'num_tokens': '1.661e+06', 'mean_token_accuracy': '0.7061', 'epoch': '0.438'}
+{'loss': '0.8173', 'grad_norm': '14.31', 'learning_rate': '5.62e-06', 'num_tokens': '1.661e+06', 'mean_token_accuracy': '0.7961', 'epoch': '0.4382'}
+{'loss': '0.9493', 'grad_norm': '13.88', 'learning_rate': '5.618e-06', 'num_tokens': '1.662e+06', 'mean_token_accuracy': '0.7576', 'epoch': '0.4383'}
+{'loss': '1.306', 'grad_norm': '14.5', 'learning_rate': '5.617e-06', 'num_tokens': '1.662e+06', 'mean_token_accuracy': '0.6588', 'epoch': '0.4385'}
+{'loss': '0.7034', 'grad_norm': '22.38', 'learning_rate': '5.615e-06', 'num_tokens': '1.662e+06', 'mean_token_accuracy': '0.8395', 'epoch': '0.4387'}
+{'loss': '0.7774', 'grad_norm': '23.25', 'learning_rate': '5.613e-06', 'num_tokens': '1.662e+06', 'mean_token_accuracy': '0.7792', 'epoch': '0.4389'}
+{'loss': '0.6426', 'grad_norm': '6.625', 'learning_rate': '5.611e-06', 'num_tokens': '1.664e+06', 'mean_token_accuracy': '0.8029', 'epoch': '0.439'}
+{'loss': '1.449', 'grad_norm': '17.62', 'learning_rate': '5.61e-06', 'num_tokens': '1.665e+06', 'mean_token_accuracy': '0.6758', 'epoch': '0.4392'}
+{'loss': '0.5238', 'grad_norm': '43', 'learning_rate': '5.608e-06', 'num_tokens': '1.665e+06', 'mean_token_accuracy': '0.8793', 'epoch': '0.4394'}
+{'loss': '1.281', 'grad_norm': '14.25', 'learning_rate': '5.606e-06', 'num_tokens': '1.665e+06', 'mean_token_accuracy': '0.714', 'epoch': '0.4396'}
+{'loss': '2.082', 'grad_norm': '9.312', 'learning_rate': '5.604e-06', 'num_tokens': '1.666e+06', 'mean_token_accuracy': '0.5441', 'epoch': '0.4397'}
+{'loss': '1.079', 'grad_norm': '5.594', 'learning_rate': '5.603e-06', 'num_tokens': '1.669e+06', 'mean_token_accuracy': '0.7256', 'epoch': '0.4399'}
+{'loss': '2.492', 'grad_norm': '21.38', 'learning_rate': '5.601e-06', 'num_tokens': '1.67e+06', 'mean_token_accuracy': '0.453', 'epoch': '0.4401'}
+{'loss': '2.102', 'grad_norm': '11', 'learning_rate': '5.599e-06', 'num_tokens': '1.67e+06', 'mean_token_accuracy': '0.5603', 'epoch': '0.4403'}
+{'loss': '2.037', 'grad_norm': '19.25', 'learning_rate': '5.597e-06', 'num_tokens': '1.671e+06', 'mean_token_accuracy': '0.5734', 'epoch': '0.4404'}
+{'loss': '1.39', 'grad_norm': '8', 'learning_rate': '5.596e-06', 'num_tokens': '1.673e+06', 'mean_token_accuracy': '0.6819', 'epoch': '0.4406'}
+{'loss': '1.982', 'grad_norm': '15.75', 'learning_rate': '5.594e-06', 'num_tokens': '1.674e+06', 'mean_token_accuracy': '0.5952', 'epoch': '0.4408'}
+{'loss': '1.564', 'grad_norm': '16.88', 'learning_rate': '5.592e-06', 'num_tokens': '1.674e+06', 'mean_token_accuracy': '0.6777', 'epoch': '0.441'}
+{'loss': '1.352', 'grad_norm': '8', 'learning_rate': '5.59e-06', 'num_tokens': '1.676e+06', 'mean_token_accuracy': '0.6796', 'epoch': '0.4411'}
+{'loss': '1.61', 'grad_norm': '11.06', 'learning_rate': '5.589e-06', 'num_tokens': '1.677e+06', 'mean_token_accuracy': '0.6645', 'epoch': '0.4413'}
+{'loss': '1.952', 'grad_norm': '16.88', 'learning_rate': '5.587e-06', 'num_tokens': '1.678e+06', 'mean_token_accuracy': '0.6109', 'epoch': '0.4415'}
+{'loss': '1.978', 'grad_norm': '10.31', 'learning_rate': '5.585e-06', 'num_tokens': '1.679e+06', 'mean_token_accuracy': '0.5902', 'epoch': '0.4416'}
+{'loss': '1.347', 'grad_norm': '12.5', 'learning_rate': '5.584e-06', 'num_tokens': '1.679e+06', 'mean_token_accuracy': '0.6219', 'epoch': '0.4418'}
+{'loss': '0.5203', 'grad_norm': '6', 'learning_rate': '5.582e-06', 'num_tokens': '1.681e+06', 'mean_token_accuracy': '0.842', 'epoch': '0.442'}
+{'loss': '0.6959', 'grad_norm': '6.594', 'learning_rate': '5.58e-06', 'num_tokens': '1.683e+06', 'mean_token_accuracy': '0.8079', 'epoch': '0.4422'}
+{'loss': '1.883', 'grad_norm': '20.12', 'learning_rate': '5.578e-06', 'num_tokens': '1.683e+06', 'mean_token_accuracy': '0.6233', 'epoch': '0.4423'}
+{'loss': '1.079', 'grad_norm': '12.19', 'learning_rate': '5.577e-06', 'num_tokens': '1.684e+06', 'mean_token_accuracy': '0.6966', 'epoch': '0.4425'}
+{'loss': '1.33', 'grad_norm': '44.75', 'learning_rate': '5.575e-06', 'num_tokens': '1.684e+06', 'mean_token_accuracy': '0.7558', 'epoch': '0.4427'}
+{'loss': '2.057', 'grad_norm': '12.56', 'learning_rate': '5.573e-06', 'num_tokens': '1.684e+06', 'mean_token_accuracy': '0.5308', 'epoch': '0.4429'}
+{'loss': '1.146', 'grad_norm': '12.44', 'learning_rate': '5.571e-06', 'num_tokens': '1.685e+06', 'mean_token_accuracy': '0.6681', 'epoch': '0.443'}
+{'loss': '1.333', 'grad_norm': '11.56', 'learning_rate': '5.57e-06', 'num_tokens': '1.686e+06', 'mean_token_accuracy': '0.6356', 'epoch': '0.4432'}
+{'loss': '1.439', 'grad_norm': '11.12', 'learning_rate': '5.568e-06', 'num_tokens': '1.687e+06', 'mean_token_accuracy': '0.6303', 'epoch': '0.4434'}
+{'loss': '1.013', 'grad_norm': '18.62', 'learning_rate': '5.566e-06', 'num_tokens': '1.687e+06', 'mean_token_accuracy': '0.7645', 'epoch': '0.4436'}
+{'loss': '0.85', 'grad_norm': '12.88', 'learning_rate': '5.564e-06', 'num_tokens': '1.688e+06', 'mean_token_accuracy': '0.82', 'epoch': '0.4437'}
+
 44%|████▍     | 2556/5758 [14:32<18:20,  2.91it/s]
 44%|████▍     | 2557/5758 [14:32<18:09,  2.94it/s]
                                                   

 44%|████▍     | 2557/5758 [14:32<18:09,  2.94it/s]
 44%|████▍     | 2558/5758 [14:32<18:50,  2.83it/s]
                                                   

 44%|████▍     | 2558/5758 [14:32<18:50,  2.83it/s]
 44%|████▍     | 2559/5758 [14:33<19:05,  2.79it/s]
                                                   

 44%|████▍     | 2559/5758 [14:33<19:05,  2.79it/s]
 44%|████▍     | 2560/5758 [14:33<18:56,  2.81it/s]
                                                   

 44%|████▍     | 2560/5758 [14:33<18:56,  2.81it/s]
 44%|████▍     | 2561/5758 [14:33<18:09,  2.93it/s]
                                                   

 44%|████▍     | 2561/5758 [14:33<18:09,  2.93it/s]
 44%|████▍     | 2562/5758 [14:34<17:27,  3.05it/s]
                                                   

 44%|████▍     | 2562/5758 [14:34<17:27,  3.05it/s]
 45%|████▍     | 2563/5758 [14:34<17:32,  3.03it/s]
                                                   

 45%|████▍     | 2563/5758 [14:34<17:32,  3.03it/s]
 45%|████▍     | 2564/5758 [14:34<18:26,  2.89it/s]
                                                   

 45%|████▍     | 2564/5758 [14:35<18:26,  2.89it/s]
 45%|████▍     | 2565/5758 [14:35<18:49,  2.83it/s]
                                                   

 45%|████▍     | 2565/5758 [14:35<18:49,  2.83it/s]
 45%|████▍     | 2566/5758 [14:35<18:09,  2.93it/s]
                                                   

 45%|████▍     | 2566/5758 [14:35<18:09,  2.93it/s]
 45%|████▍     | 2567/5758 [14:36<18:15,  2.91it/s]
                                                   

 45%|████▍     | 2567/5758 [14:36<18:15,  2.91it/s]
 45%|████▍     | 2568/5758 [14:36<17:25,  3.05it/s]
                                                   

 45%|████▍     | 2568/5758 [14:36<17:25,  3.05it/s]
 45%|████▍     | 2569/5758 [14:36<16:54,  3.14it/s]
                                                   

 45%|████▍     | 2569/5758 [14:36<16:54,  3.14it/s]
 45%|████▍     | 2570/5758 [14:36<16:54,  3.14it/s]
                                                   

 45%|████▍     | 2570/5758 [14:36<16:54,  3.14it/s]
 45%|████▍     | 2571/5758 [14:37<17:46,  2.99it/s]
                                                   

 45%|████▍     | 2571/5758 [14:37<17:46,  2.99it/s]
 45%|████▍     | 2572/5758 [14:37<17:14,  3.08it/s]
                                                   

 45%|████▍     | 2572/5758 [14:37<17:14,  3.08it/s]
 45%|████▍     | 2573/5758 [14:38<20:09,  2.63it/s]
                                                   

 45%|████▍     | 2573/5758 [14:38<20:09,  2.63it/s]
 45%|████▍     | 2574/5758 [14:38<19:04,  2.78it/s]
                                                   

 45%|████▍     | 2574/5758 [14:38<19:04,  2.78it/s]
 45%|████▍     | 2575/5758 [14:38<18:41,  2.84it/s]
                                                   

 45%|████▍     | 2575/5758 [14:38<18:41,  2.84it/s]
 45%|████▍     | 2576/5758 [14:39<19:08,  2.77it/s]
                                                   

 45%|████▍     | 2576/5758 [14:39<19:08,  2.77it/s]
 45%|████▍     | 2577/5758 [14:39<19:25,  2.73it/s]
                                                   

 45%|████▍     | 2577/5758 [14:39<19:25,  2.73it/s]
 45%|████▍     | 2578/5758 [14:39<19:11,  2.76it/s]
                                                   

 45%|████▍     | 2578/5758 [14:39<19:11,  2.76it/s]
 45%|████▍     | 2579/5758 [14:40<18:23,  2.88it/s]
                                                   

 45%|████▍     | 2579/5758 [14:40<18:23,  2.88it/s]
 45%|████▍     | 2580/5758 [14:40<18:28,  2.87it/s]
                                                   

 45%|████▍     | 2580/5758 [14:40<18:28,  2.87it/s]
 45%|████▍     | 2581/5758 [14:40<17:54,  2.96it/s]
                                                   

 45%|████▍     | 2581/5758 [14:40<17:54,  2.96it/s]
 45%|████▍     | 2582/5758 [14:41<18:34,  2.85it/s]
                                                   

 45%|████▍     | 2582/5758 [14:41<18:34,  2.85it/s]
 45%|████▍     | 2583/5758 [14:41<18:28,  2.86it/s]
                                                   

 45%|████▍     | 2583/5758 [14:41<18:28,  2.86it/s]
 45%|████▍     | 2584/5758 [14:41<18:38,  2.84it/s]
                                                   

 45%|████▍     | 2584/5758 [14:41<18:38,  2.84it/s]
 45%|████▍     | 2585/5758 [14:42<18:18,  2.89it/s]
                                                   

 45%|████▍     | 2585/5758 [14:42<18:18,  2.89it/s]
 45%|████▍     | 2586/5758 [14:42<17:51,  2.96it/s]
                                                   

 45%|████▍     | 2586/5758 [14:42<17:51,  2.96it/s]
 45%|████▍     | 2587/5758 [14:42<17:34,  3.01it/s]
                                                   

 45%|████▍     | 2587/5758 [14:42<17:34,  3.01it/s]
 45%|████▍     | 2588/5758 [14:43<17:45,  2.98it/s]
                                                   

 45%|████▍     | 2588/5758 [14:43<17:45,  2.98it/s]
 45%|████▍     | 2589/5758 [14:43<17:24,  3.03it/s]
                                                   

 45%|████▍     | 2589/5758 [14:43<17:24,  3.03it/s]
 45%|████▍     | 2590/5758 [14:43<16:47,  3.14it/s]
                                                   

 45%|████▍     | 2590/5758 [14:43<16:47,  3.14it/s]
 45%|████▍     | 2591/5758 [14:44<16:22,  3.22it/s]
                                                   

 45%|████▍     | 2591/5758 [14:44<16:22,  3.22it/s]
 45%|████▌     | 2592/5758 [14:44<17:16,  3.05it/s]
                                                   

 45%|████▌     | 2592/5758 [14:44<17:16,  3.05it/s]
 45%|████▌     | 2593/5758 [14:44<17:21,  3.04it/s]
                                                   

 45%|████▌     | 2593/5758 [14:44<17:21,  3.04it/s]
 45%|████▌     | 2594/5758 [14:45<17:21,  3.04it/s]
                                                   

 45%|████▌     | 2594/5758 [14:45<17:21,  3.04it/s]
 45%|████▌     | 2595/5758 [14:45<17:14,  3.06it/s]
                                                   

 45%|████▌     | 2595/5758 [14:45<17:14,  3.06it/s]
 45%|████▌     | 2596/5758 [14:45<17:26,  3.02it/s]
                                                   

 45%|████▌     | 2596/5758 [14:45<17:26,  3.02it/s]
 45%|████▌     | 2597/5758 [14:46<17:27,  3.02it/s]
                                                   

 45%|████▌     | 2597/5758 [14:46<17:27,  3.02it/s]
 45%|████▌     | 2598/5758 [14:46<18:20,  2.87it/s]
                                                   

 45%|████▌     | 2598/5758 [14:46<18:20,  2.87it/s]
 45%|████▌     | 2599/5758 [14:46<18:51,  2.79it/s]
                                                   

 45%|████▌     | 2599/5758 [14:46<18:51,  2.79it/s]
 45%|████▌     | 2600/5758 [14:47<18:16,  2.88it/s]
                                                   

 45%|████▌     | 2600/5758 [14:47<18:16,  2.88it/s]
 45%|████▌     | 2601/5758 [14:47<17:39,  2.98it/s]
                                                   

 45%|████▌     | 2601/5758 [14:47<17:39,  2.98it/s]
 45%|████▌     | 2602/5758 [14:48<20:13,  2.60it/s]
                                                   

 45%|████▌     | 2602/5758 [14:48<20:13,  2.60it/s]
 45%|████▌     | 2603/5758 [14:48<19:09,  2.74it/s]
                                                   

 45%|████▌     | 2603/5758 [14:48<19:09,  2.74it/s]
 45%|████▌     | 2604/5758 [14:48<18:47,  2.80it/s]
                                                   

 45%|████▌     | 2604/5758 [14:48<18:47,  2.80it/s]
 45%|████▌     | 2605/5758 [14:49<18:06,  2.90it/s]
                                                   

 45%|████▌     | 2605/5758 [14:49<18:06,  2.90it/s]
 45%|████▌     | 2606/5758 [14:49<17:41,  2.97it/s]
                                                   

 45%|████▌     | 2606/5758 [14:49<17:41,  2.97it/s]
 45%|████▌     | 2607/5758 [14:49<17:19,  3.03it/s]
                                                   

 45%|████▌     | 2607/5758 [14:49<17:19,  3.03it/s]
 45%|████▌     | 2608/5758 [14:49<16:58,  3.09it/s]
                                                   

 45%|████▌     | 2608/5758 [14:49<16:58,  3.09it/s]
 45%|████▌     | 2609/5758 [14:50<17:11,  3.05it/s]
                                                   

 45%|████▌     | 2609/5758 [14:50<17:11,  3.05it/s]
 45%|████▌     | 2610/5758 [14:50<16:34,  3.16it/s]
                                                   

 45%|████▌     | 2610/5758 [14:50<16:34,  3.16it/s]
 45%|████▌     | 2611/5758 [14:50<16:58,  3.09it/s]
                                                   
{'loss': '1.294', 'grad_norm': '18.62', 'learning_rate': '5.563e-06', 'num_tokens': '1.688e+06', 'mean_token_accuracy': '0.7138', 'epoch': '0.4439'}
+{'loss': '0.6683', 'grad_norm': '7.062', 'learning_rate': '5.561e-06', 'num_tokens': '1.69e+06', 'mean_token_accuracy': '0.8062', 'epoch': '0.4441'}
+{'loss': '1.701', 'grad_norm': '17', 'learning_rate': '5.559e-06', 'num_tokens': '1.69e+06', 'mean_token_accuracy': '0.6219', 'epoch': '0.4443'}
+{'loss': '1.313', 'grad_norm': '26.88', 'learning_rate': '5.557e-06', 'num_tokens': '1.691e+06', 'mean_token_accuracy': '0.686', 'epoch': '0.4444'}
+{'loss': '1.127', 'grad_norm': '26.75', 'learning_rate': '5.556e-06', 'num_tokens': '1.691e+06', 'mean_token_accuracy': '0.7729', 'epoch': '0.4446'}
+{'loss': '1.444', 'grad_norm': '18', 'learning_rate': '5.554e-06', 'num_tokens': '1.691e+06', 'mean_token_accuracy': '0.688', 'epoch': '0.4448'}
+{'loss': '1.695', 'grad_norm': '12.75', 'learning_rate': '5.552e-06', 'num_tokens': '1.692e+06', 'mean_token_accuracy': '0.6502', 'epoch': '0.4449'}
+{'loss': '1.155', 'grad_norm': '15', 'learning_rate': '5.551e-06', 'num_tokens': '1.692e+06', 'mean_token_accuracy': '0.7494', 'epoch': '0.4451'}
+{'loss': '1.79', 'grad_norm': '18.25', 'learning_rate': '5.549e-06', 'num_tokens': '1.692e+06', 'mean_token_accuracy': '0.6159', 'epoch': '0.4453'}
+{'loss': '1.281', 'grad_norm': '18.25', 'learning_rate': '5.547e-06', 'num_tokens': '1.693e+06', 'mean_token_accuracy': '0.7164', 'epoch': '0.4455'}
+{'loss': '1.201', 'grad_norm': '18.12', 'learning_rate': '5.545e-06', 'num_tokens': '1.693e+06', 'mean_token_accuracy': '0.6882', 'epoch': '0.4456'}
+{'loss': '1.378', 'grad_norm': '14.69', 'learning_rate': '5.544e-06', 'num_tokens': '1.694e+06', 'mean_token_accuracy': '0.7134', 'epoch': '0.4458'}
+{'loss': '1.875', 'grad_norm': '17.75', 'learning_rate': '5.542e-06', 'num_tokens': '1.694e+06', 'mean_token_accuracy': '0.6109', 'epoch': '0.446'}
+{'loss': '1.359', 'grad_norm': '13', 'learning_rate': '5.54e-06', 'num_tokens': '1.695e+06', 'mean_token_accuracy': '0.6375', 'epoch': '0.4462'}
+{'loss': '1.436', 'grad_norm': '13', 'learning_rate': '5.538e-06', 'num_tokens': '1.696e+06', 'mean_token_accuracy': '0.6438', 'epoch': '0.4463'}
+{'loss': '1.183', 'grad_norm': '9.438', 'learning_rate': '5.537e-06', 'num_tokens': '1.697e+06', 'mean_token_accuracy': '0.6861', 'epoch': '0.4465'}
+{'loss': '1.26', 'grad_norm': '12.81', 'learning_rate': '5.535e-06', 'num_tokens': '1.697e+06', 'mean_token_accuracy': '0.6377', 'epoch': '0.4467'}
+{'loss': '1.15', 'grad_norm': '14.12', 'learning_rate': '5.533e-06', 'num_tokens': '1.698e+06', 'mean_token_accuracy': '0.6877', 'epoch': '0.4469'}
+{'loss': '2.136', 'grad_norm': '19.25', 'learning_rate': '5.531e-06', 'num_tokens': '1.698e+06', 'mean_token_accuracy': '0.5504', 'epoch': '0.447'}
+{'loss': '2.014', 'grad_norm': '20.88', 'learning_rate': '5.53e-06', 'num_tokens': '1.699e+06', 'mean_token_accuracy': '0.5831', 'epoch': '0.4472'}
+{'loss': '1.969', 'grad_norm': '16.5', 'learning_rate': '5.528e-06', 'num_tokens': '1.699e+06', 'mean_token_accuracy': '0.6024', 'epoch': '0.4474'}
+{'loss': '1.394', 'grad_norm': '16.5', 'learning_rate': '5.526e-06', 'num_tokens': '1.7e+06', 'mean_token_accuracy': '0.6519', 'epoch': '0.4476'}
+{'loss': '1.283', 'grad_norm': '11.5', 'learning_rate': '5.524e-06', 'num_tokens': '1.701e+06', 'mean_token_accuracy': '0.6373', 'epoch': '0.4477'}
+{'loss': '1.646', 'grad_norm': '24', 'learning_rate': '5.523e-06', 'num_tokens': '1.701e+06', 'mean_token_accuracy': '0.6789', 'epoch': '0.4479'}
+{'loss': '1.583', 'grad_norm': '14.44', 'learning_rate': '5.521e-06', 'num_tokens': '1.702e+06', 'mean_token_accuracy': '0.6452', 'epoch': '0.4481'}
+{'loss': '2.12', 'grad_norm': '15.5', 'learning_rate': '5.519e-06', 'num_tokens': '1.702e+06', 'mean_token_accuracy': '0.5952', 'epoch': '0.4482'}
+{'loss': '1.086', 'grad_norm': '17.25', 'learning_rate': '5.518e-06', 'num_tokens': '1.703e+06', 'mean_token_accuracy': '0.7544', 'epoch': '0.4484'}
+{'loss': '1.931', 'grad_norm': '10.75', 'learning_rate': '5.516e-06', 'num_tokens': '1.704e+06', 'mean_token_accuracy': '0.6168', 'epoch': '0.4486'}
+{'loss': '1.202', 'grad_norm': '18.5', 'learning_rate': '5.514e-06', 'num_tokens': '1.704e+06', 'mean_token_accuracy': '0.7383', 'epoch': '0.4488'}
+{'loss': '1.393', 'grad_norm': '11.56', 'learning_rate': '5.512e-06', 'num_tokens': '1.705e+06', 'mean_token_accuracy': '0.6437', 'epoch': '0.4489'}
+{'loss': '1.262', 'grad_norm': '20.62', 'learning_rate': '5.511e-06', 'num_tokens': '1.705e+06', 'mean_token_accuracy': '0.7588', 'epoch': '0.4491'}
+{'loss': '0.5722', 'grad_norm': '6.781', 'learning_rate': '5.509e-06', 'num_tokens': '1.707e+06', 'mean_token_accuracy': '0.8329', 'epoch': '0.4493'}
+{'loss': '1.216', 'grad_norm': '10.75', 'learning_rate': '5.507e-06', 'num_tokens': '1.707e+06', 'mean_token_accuracy': '0.6684', 'epoch': '0.4495'}
+{'loss': '0.8288', 'grad_norm': '5.344', 'learning_rate': '5.505e-06', 'num_tokens': '1.71e+06', 'mean_token_accuracy': '0.7896', 'epoch': '0.4496'}
+{'loss': '1.054', 'grad_norm': '8.312', 'learning_rate': '5.504e-06', 'num_tokens': '1.712e+06', 'mean_token_accuracy': '0.7642', 'epoch': '0.4498'}
+{'loss': '1.501', 'grad_norm': '14.38', 'learning_rate': '5.502e-06', 'num_tokens': '1.712e+06', 'mean_token_accuracy': '0.6594', 'epoch': '0.45'}
+{'loss': '1.839', 'grad_norm': '7.656', 'learning_rate': '5.5e-06', 'num_tokens': '1.715e+06', 'mean_token_accuracy': '0.6172', 'epoch': '0.4502'}
+{'loss': '1.256', 'grad_norm': '16', 'learning_rate': '5.498e-06', 'num_tokens': '1.715e+06', 'mean_token_accuracy': '0.66', 'epoch': '0.4503'}
+{'loss': '1.125', 'grad_norm': '27.38', 'learning_rate': '5.497e-06', 'num_tokens': '1.715e+06', 'mean_token_accuracy': '0.7396', 'epoch': '0.4505'}
+{'loss': '1.107', 'grad_norm': '28.5', 'learning_rate': '5.495e-06', 'num_tokens': '1.715e+06', 'mean_token_accuracy': '0.7273', 'epoch': '0.4507'}
+{'loss': '0.947', 'grad_norm': '36.25', 'learning_rate': '5.493e-06', 'num_tokens': '1.715e+06', 'mean_token_accuracy': '0.7745', 'epoch': '0.4509'}
+{'loss': '1.463', 'grad_norm': '13', 'learning_rate': '5.491e-06', 'num_tokens': '1.716e+06', 'mean_token_accuracy': '0.6125', 'epoch': '0.451'}
+{'loss': '1.861', 'grad_norm': '27.38', 'learning_rate': '5.49e-06', 'num_tokens': '1.716e+06', 'mean_token_accuracy': '0.6308', 'epoch': '0.4512'}
+{'loss': '1.348', 'grad_norm': '10.69', 'learning_rate': '5.488e-06', 'num_tokens': '1.718e+06', 'mean_token_accuracy': '0.6232', 'epoch': '0.4514'}
+{'loss': '0.5338', 'grad_norm': '20.62', 'learning_rate': '5.486e-06', 'num_tokens': '1.718e+06', 'mean_token_accuracy': '0.8621', 'epoch': '0.4515'}
+{'loss': '2.132', 'grad_norm': '16.88', 'learning_rate': '5.485e-06', 'num_tokens': '1.718e+06', 'mean_token_accuracy': '0.5815', 'epoch': '0.4517'}
+{'loss': '1.636', 'grad_norm': '16.62', 'learning_rate': '5.483e-06', 'num_tokens': '1.719e+06', 'mean_token_accuracy': '0.645', 'epoch': '0.4519'}
+{'loss': '0.7207', 'grad_norm': '20.25', 'learning_rate': '5.481e-06', 'num_tokens': '1.719e+06', 'mean_token_accuracy': '0.8649', 'epoch': '0.4521'}
+{'loss': '1.037', 'grad_norm': '10.75', 'learning_rate': '5.479e-06', 'num_tokens': '1.72e+06', 'mean_token_accuracy': '0.7389', 'epoch': '0.4522'}
+{'loss': '0.9118', 'grad_norm': '34.75', 'learning_rate': '5.478e-06', 'num_tokens': '1.72e+06', 'mean_token_accuracy': '0.8313', 'epoch': '0.4524'}
+{'loss': '1.333', 'grad_norm': '12.25', 'learning_rate': '5.476e-06', 'num_tokens': '1.721e+06', 'mean_token_accuracy': '0.6355', 'epoch': '0.4526'}
+{'loss': '0.6184', 'grad_norm': '7.375', 'learning_rate': '5.474e-06', 'num_tokens': '1.722e+06', 'mean_token_accuracy': '0.8257', 'epoch': '0.4528'}
+{'loss': '0.4856', 'grad_norm': '6.562', 'learning_rate': '5.472e-06', 'num_tokens': '1.723e+06', 'mean_token_accuracy': '0.8539', 'epoch': '0.4529'}
+{'loss': '1.042', 'grad_norm': '18.25', 'learning_rate': '5.471e-06', 'num_tokens': '1.724e+06', 'mean_token_accuracy': '0.8054', 'epoch': '0.4531'}
+{'loss': '1.732', 'grad_norm': '15.62', 'learning_rate': '5.469e-06', 'num_tokens': '1.724e+06', 'mean_token_accuracy': '0.6373', 'epoch': '0.4533'}
+
 45%|████▌     | 2611/5758 [14:50<16:58,  3.09it/s]
 45%|████▌     | 2612/5758 [14:51<17:37,  2.97it/s]
                                                   

 45%|████▌     | 2612/5758 [14:51<17:37,  2.97it/s]
 45%|████▌     | 2613/5758 [14:51<18:07,  2.89it/s]
                                                   

 45%|████▌     | 2613/5758 [14:51<18:07,  2.89it/s]
 45%|████▌     | 2614/5758 [14:52<18:08,  2.89it/s]
                                                   

 45%|████▌     | 2614/5758 [14:52<18:08,  2.89it/s]
 45%|████▌     | 2615/5758 [14:52<17:44,  2.95it/s]
                                                   

 45%|████▌     | 2615/5758 [14:52<17:44,  2.95it/s]
 45%|████▌     | 2616/5758 [14:52<18:30,  2.83it/s]
                                                   

 45%|████▌     | 2616/5758 [14:52<18:30,  2.83it/s]
 45%|████▌     | 2617/5758 [14:53<17:54,  2.92it/s]
                                                   

 45%|████▌     | 2617/5758 [14:53<17:54,  2.92it/s]
 45%|████▌     | 2618/5758 [14:53<18:30,  2.83it/s]
                                                   

 45%|████▌     | 2618/5758 [14:53<18:30,  2.83it/s]
 45%|████▌     | 2619/5758 [14:53<18:21,  2.85it/s]
                                                   

 45%|████▌     | 2619/5758 [14:53<18:21,  2.85it/s]
 46%|███��▌     | 2620/5758 [14:54<17:24,  3.00it/s]
                                                   

 46%|████▌     | 2620/5758 [14:54<17:24,  3.00it/s]
 46%|████▌     | 2621/5758 [14:54<16:52,  3.10it/s]
                                                   

 46%|████▌     | 2621/5758 [14:54<16:52,  3.10it/s]
 46%|████▌     | 2622/5758 [14:54<16:53,  3.10it/s]
                                                   

 46%|████▌     | 2622/5758 [14:54<16:53,  3.10it/s]
 46%|████▌     | 2623/5758 [14:55<17:04,  3.06it/s]
                                                   

 46%|████▌     | 2623/5758 [14:55<17:04,  3.06it/s]
 46%|████▌     | 2624/5758 [14:55<17:24,  3.00it/s]
                                                   

 46%|████▌     | 2624/5758 [14:55<17:24,  3.00it/s]
 46%|████▌     | 2625/5758 [14:55<16:39,  3.13it/s]
                                                   

 46%|████▌     | 2625/5758 [14:55<16:39,  3.13it/s]
 46%|████▌     | 2626/5758 [14:55<16:36,  3.14it/s]
                                                   

 46%|████▌     | 2626/5758 [14:55<16:36,  3.14it/s]
 46%|████▌     | 2627/5758 [14:56<16:21,  3.19it/s]
                                                   

 46%|████▌     | 2627/5758 [14:56<16:21,  3.19it/s]
 46%|████▌     | 2628/5758 [14:56<17:58,  2.90it/s]
                                                   

 46%|████▌     | 2628/5758 [14:56<17:58,  2.90it/s]
 46%|████▌     | 2629/5758 [14:56<17:04,  3.06it/s]
                                                   

 46%|████▌     | 2629/5758 [14:56<17:04,  3.06it/s]
 46%|████▌     | 2630/5758 [14:57<16:35,  3.14it/s]
                                                   

 46%|████▌     | 2630/5758 [14:57<16:35,  3.14it/s]
 46%|████▌     | 2631/5758 [14:57<17:22,  3.00it/s]
                                                   

 46%|████▌     | 2631/5758 [14:57<17:22,  3.00it/s]
 46%|████▌     | 2632/5758 [14:58<18:05,  2.88it/s]
                                                   

 46%|████▌     | 2632/5758 [14:58<18:05,  2.88it/s]
 46%|████▌     | 2633/5758 [14:58<17:47,  2.93it/s]
                                                   

 46%|████▌     | 2633/5758 [14:58<17:47,  2.93it/s]
 46%|████▌     | 2634/5758 [14:58<17:37,  2.95it/s]
                                                   

 46%|████▌     | 2634/5758 [14:58<17:37,  2.95it/s]
 46%|████▌     | 2635/5758 [14:59<18:22,  2.83it/s]
                                                   

 46%|████▌     | 2635/5758 [14:59<18:22,  2.83it/s]
 46%|████▌     | 2636/5758 [14:59<18:39,  2.79it/s]
                                                   

 46%|████▌     | 2636/5758 [14:59<18:39,  2.79it/s]
 46%|████▌     | 2637/5758 [14:59<17:46,  2.93it/s]
                                                   

 46%|████▌     | 2637/5758 [14:59<17:46,  2.93it/s]
 46%|████▌     | 2638/5758 [15:00<17:24,  2.99it/s]
                                                   

 46%|████▌     | 2638/5758 [15:00<17:24,  2.99it/s]
 46%|████▌     | 2639/5758 [15:00<17:56,  2.90it/s]
                                                   

 46%|████▌     | 2639/5758 [15:00<17:56,  2.90it/s]
 46%|████▌     | 2640/5758 [15:00<17:07,  3.03it/s]
                                                   

 46%|████▌     | 2640/5758 [15:00<17:07,  3.03it/s]
 46%|████▌     | 2641/5758 [15:01<17:59,  2.89it/s]
                                                   

 46%|████▌     | 2641/5758 [15:01<17:59,  2.89it/s]
 46%|████▌     | 2642/5758 [15:01<18:55,  2.75it/s]
                                                   

 46%|████▌     | 2642/5758 [15:01<18:55,  2.75it/s]
 46%|████▌     | 2643/5758 [15:01<17:55,  2.90it/s]
                                                   

 46%|████▌     | 2643/5758 [15:01<17:55,  2.90it/s]
 46%|████▌     | 2644/5758 [15:02<17:20,  2.99it/s]
                                                   

 46%|████▌     | 2644/5758 [15:02<17:20,  2.99it/s]
 46%|████▌     | 2645/5758 [15:02<16:52,  3.07it/s]
                                                   

 46%|████▌     | 2645/5758 [15:02<16:52,  3.07it/s]
 46%|████▌     | 2646/5758 [15:02<17:04,  3.04it/s]
                                                   

 46%|████▌     | 2646/5758 [15:02<17:04,  3.04it/s]
 46%|████▌     | 2647/5758 [15:03<17:22,  2.98it/s]
                                                   

 46%|████▌     | 2647/5758 [15:03<17:22,  2.98it/s]
 46%|████▌     | 2648/5758 [15:03<17:30,  2.96it/s]
                                                   

 46%|████▌     | 2648/5758 [15:03<17:30,  2.96it/s]
 46%|████▌     | 2649/5758 [15:03<16:43,  3.10it/s]
                                                   

 46%|████▌     | 2649/5758 [15:03<16:43,  3.10it/s]
 46%|████▌     | 2650/5758 [15:04<16:22,  3.16it/s]
                                                   

 46%|████▌     | 2650/5758 [15:04<16:22,  3.16it/s]
 46%|████▌     | 2651/5758 [15:04<16:00,  3.23it/s]
                                                   

 46%|████▌     | 2651/5758 [15:04<16:00,  3.23it/s]
 46%|████▌     | 2652/5758 [15:04<16:11,  3.20it/s]
                                                   

 46%|████▌     | 2652/5758 [15:04<16:11,  3.20it/s]
 46%|████▌     | 2653/5758 [15:04<16:15,  3.18it/s]
                                                   

 46%|████▌     | 2653/5758 [15:05<16:15,  3.18it/s]
 46%|████▌     | 2654/5758 [15:05<19:12,  2.69it/s]
                                                   

 46%|████▌     | 2654/5758 [15:05<19:12,  2.69it/s]
 46%|████▌     | 2655/5758 [15:05<19:38,  2.63it/s]
                                                   

 46%|████▌     | 2655/5758 [15:05<19:38,  2.63it/s]
 46%|████▌     | 2656/5758 [15:06<18:25,  2.81it/s]
                                                   

 46%|████▌     | 2656/5758 [15:06<18:25,  2.81it/s]
 46%|████▌     | 2657/5758 [15:06<18:47,  2.75it/s]
                                                   

 46%|████▌     | 2657/5758 [15:06<18:47,  2.75it/s]
 46%|████▌     | 2658/5758 [15:06<17:52,  2.89it/s]
                                                   

 46%|████▌     | 2658/5758 [15:06<17:52,  2.89it/s]
 46%|████▌     | 2659/5758 [15:07<17:53,  2.89it/s]
                                                   

 46%|████▌     | 2659/5758 [15:07<17:53,  2.89it/s]
 46%|████▌     | 2660/5758 [15:07<18:35,  2.78it/s]
                                                   

 46%|████▌     | 2660/5758 [15:07<18:35,  2.78it/s]
 46%|████▌     | 2661/5758 [15:07<18:32,  2.78it/s]
                                                   

 46%|████▌     | 2661/5758 [15:07<18:32,  2.78it/s]
 46%|████▌     | 2662/5758 [15:08<17:44,  2.91it/s]
                                                   

 46%|████▌     | 2662/5758 [15:08<17:44,  2.91it/s]
 46%|████▌     | 2663/5758 [15:08<17:05,  3.02it/s]
                                                   

 46%|████▌     | 2663/5758 [15:08<17:05,  3.02it/s]
 46%|████▋     | 2664/5758 [15:08<16:58,  3.04it/s]
                                                   

 46%|████▋     | 2664/5758 [15:08<16:58,  3.04it/s]
 46%|████▋     | 2665/5758 [15:09<16:42,  3.09it/s]
                                                   

 46%|████▋     | 2665/5758 [15:09<16:42,  3.09it/s]
 46%|████▋     | 2666/5758 [15:09<16:32,  3.12it/s]
                                                   
{'loss': '1.163', 'grad_norm': '16.62', 'learning_rate': '5.467e-06', 'num_tokens': '1.725e+06', 'mean_token_accuracy': '0.7102', 'epoch': '0.4535'}
+{'loss': '1.063', 'grad_norm': '6.906', 'learning_rate': '5.465e-06', 'num_tokens': '1.727e+06', 'mean_token_accuracy': '0.7331', 'epoch': '0.4536'}
+{'loss': '1.85', 'grad_norm': '14.25', 'learning_rate': '5.464e-06', 'num_tokens': '1.727e+06', 'mean_token_accuracy': '0.6212', 'epoch': '0.4538'}
+{'loss': '1.152', 'grad_norm': '11.31', 'learning_rate': '5.462e-06', 'num_tokens': '1.728e+06', 'mean_token_accuracy': '0.6751', 'epoch': '0.454'}
+{'loss': '0.8762', 'grad_norm': '20.62', 'learning_rate': '5.46e-06', 'num_tokens': '1.728e+06', 'mean_token_accuracy': '0.7424', 'epoch': '0.4542'}
+{'loss': '0.9464', 'grad_norm': '13.75', 'learning_rate': '5.458e-06', 'num_tokens': '1.729e+06', 'mean_token_accuracy': '0.7464', 'epoch': '0.4543'}
+{'loss': '2.373', 'grad_norm': '29.62', 'learning_rate': '5.457e-06', 'num_tokens': '1.729e+06', 'mean_token_accuracy': '0.5882', 'epoch': '0.4545'}
+{'loss': '1.152', 'grad_norm': '14.06', 'learning_rate': '5.455e-06', 'num_tokens': '1.73e+06', 'mean_token_accuracy': '0.6906', 'epoch': '0.4547'}
+{'loss': '1.44', 'grad_norm': '11.56', 'learning_rate': '5.453e-06', 'num_tokens': '1.731e+06', 'mean_token_accuracy': '0.6103', 'epoch': '0.4548'}
+{'loss': '1.749', 'grad_norm': '11.62', 'learning_rate': '5.452e-06', 'num_tokens': '1.731e+06', 'mean_token_accuracy': '0.601', 'epoch': '0.455'}
+{'loss': '1.267', 'grad_norm': '12.88', 'learning_rate': '5.45e-06', 'num_tokens': '1.732e+06', 'mean_token_accuracy': '0.6499', 'epoch': '0.4552'}
+{'loss': '1.236', 'grad_norm': '9.375', 'learning_rate': '5.448e-06', 'num_tokens': '1.734e+06', 'mean_token_accuracy': '0.6928', 'epoch': '0.4554'}
+{'loss': '1.429', 'grad_norm': '7.062', 'learning_rate': '5.446e-06', 'num_tokens': '1.737e+06', 'mean_token_accuracy': '0.6742', 'epoch': '0.4555'}
+{'loss': '0.5825', 'grad_norm': '9.312', 'learning_rate': '5.445e-06', 'num_tokens': '1.738e+06', 'mean_token_accuracy': '0.8138', 'epoch': '0.4557'}
+{'loss': '1.912', 'grad_norm': '9.625', 'learning_rate': '5.443e-06', 'num_tokens': '1.739e+06', 'mean_token_accuracy': '0.5987', 'epoch': '0.4559'}
+{'loss': '1.226', 'grad_norm': '12.06', 'learning_rate': '5.441e-06', 'num_tokens': '1.74e+06', 'mean_token_accuracy': '0.6646', 'epoch': '0.4561'}
+{'loss': '1.211', 'grad_norm': '12.5', 'learning_rate': '5.439e-06', 'num_tokens': '1.741e+06', 'mean_token_accuracy': '0.6638', 'epoch': '0.4562'}
+{'loss': '0.7526', 'grad_norm': '15.56', 'learning_rate': '5.438e-06', 'num_tokens': '1.741e+06', 'mean_token_accuracy': '0.8284', 'epoch': '0.4564'}
+{'loss': '1.911', 'grad_norm': '13.25', 'learning_rate': '5.436e-06', 'num_tokens': '1.742e+06', 'mean_token_accuracy': '0.6098', 'epoch': '0.4566'}
+{'loss': '1.065', 'grad_norm': '12.25', 'learning_rate': '5.434e-06', 'num_tokens': '1.742e+06', 'mean_token_accuracy': '0.6946', 'epoch': '0.4568'}
+{'loss': '1.626', 'grad_norm': '10.62', 'learning_rate': '5.432e-06', 'num_tokens': '1.743e+06', 'mean_token_accuracy': '0.6423', 'epoch': '0.4569'}
+{'loss': '0.9452', 'grad_norm': '12.06', 'learning_rate': '5.431e-06', 'num_tokens': '1.744e+06', 'mean_token_accuracy': '0.7522', 'epoch': '0.4571'}
+{'loss': '1.997', 'grad_norm': '8.938', 'learning_rate': '5.429e-06', 'num_tokens': '1.746e+06', 'mean_token_accuracy': '0.5783', 'epoch': '0.4573'}
+{'loss': '1.812', 'grad_norm': '21.5', 'learning_rate': '5.427e-06', 'num_tokens': '1.746e+06', 'mean_token_accuracy': '0.6464', 'epoch': '0.4575'}
+{'loss': '1.763', 'grad_norm': '42', 'learning_rate': '5.425e-06', 'num_tokens': '1.746e+06', 'mean_token_accuracy': '0.6119', 'epoch': '0.4576'}
+{'loss': '1.287', 'grad_norm': '10.44', 'learning_rate': '5.424e-06', 'num_tokens': '1.747e+06', 'mean_token_accuracy': '0.6362', 'epoch': '0.4578'}
+{'loss': '0.6441', 'grad_norm': '25.38', 'learning_rate': '5.422e-06', 'num_tokens': '1.747e+06', 'mean_token_accuracy': '0.8631', 'epoch': '0.458'}
+{'loss': '1.21', 'grad_norm': '12.94', 'learning_rate': '5.42e-06', 'num_tokens': '1.748e+06', 'mean_token_accuracy': '0.6569', 'epoch': '0.4581'}
+{'loss': '1.379', 'grad_norm': '32.5', 'learning_rate': '5.419e-06', 'num_tokens': '1.748e+06', 'mean_token_accuracy': '0.7377', 'epoch': '0.4583'}
+{'loss': '0.6427', 'grad_norm': '8.812', 'learning_rate': '5.417e-06', 'num_tokens': '1.75e+06', 'mean_token_accuracy': '0.7982', 'epoch': '0.4585'}
+{'loss': '0.8443', 'grad_norm': '10', 'learning_rate': '5.415e-06', 'num_tokens': '1.751e+06', 'mean_token_accuracy': '0.7466', 'epoch': '0.4587'}
+{'loss': '0.4316', 'grad_norm': '18.5', 'learning_rate': '5.413e-06', 'num_tokens': '1.751e+06', 'mean_token_accuracy': '0.9048', 'epoch': '0.4588'}
+{'loss': '2.018', 'grad_norm': '17.25', 'learning_rate': '5.412e-06', 'num_tokens': '1.751e+06', 'mean_token_accuracy': '0.5702', 'epoch': '0.459'}
+{'loss': '1.009', 'grad_norm': '12.56', 'learning_rate': '5.41e-06', 'num_tokens': '1.752e+06', 'mean_token_accuracy': '0.6832', 'epoch': '0.4592'}
+{'loss': '0.9968', 'grad_norm': '11.75', 'learning_rate': '5.408e-06', 'num_tokens': '1.753e+06', 'mean_token_accuracy': '0.7165', 'epoch': '0.4594'}
+{'loss': '1.581', 'grad_norm': '21', 'learning_rate': '5.406e-06', 'num_tokens': '1.753e+06', 'mean_token_accuracy': '0.7022', 'epoch': '0.4595'}
+{'loss': '1.721', 'grad_norm': '21.88', 'learning_rate': '5.405e-06', 'num_tokens': '1.753e+06', 'mean_token_accuracy': '0.6452', 'epoch': '0.4597'}
+{'loss': '0.8085', 'grad_norm': '28.12', 'learning_rate': '5.403e-06', 'num_tokens': '1.753e+06', 'mean_token_accuracy': '0.785', 'epoch': '0.4599'}
+{'loss': '1.164', 'grad_norm': '13.25', 'learning_rate': '5.401e-06', 'num_tokens': '1.754e+06', 'mean_token_accuracy': '0.6602', 'epoch': '0.4601'}
+{'loss': '1.967', 'grad_norm': '15.81', 'learning_rate': '5.399e-06', 'num_tokens': '1.755e+06', 'mean_token_accuracy': '0.6198', 'epoch': '0.4602'}
+{'loss': '1.314', 'grad_norm': '13.62', 'learning_rate': '5.398e-06', 'num_tokens': '1.755e+06', 'mean_token_accuracy': '0.6375', 'epoch': '0.4604'}
+{'loss': '1.176', 'grad_norm': '12.06', 'learning_rate': '5.396e-06', 'num_tokens': '1.756e+06', 'mean_token_accuracy': '0.6683', 'epoch': '0.4606'}
+{'loss': '0.8978', 'grad_norm': '8', 'learning_rate': '5.394e-06', 'num_tokens': '1.758e+06', 'mean_token_accuracy': '0.7956', 'epoch': '0.4608'}
+{'loss': '1.699', 'grad_norm': '14.62', 'learning_rate': '5.392e-06', 'num_tokens': '1.758e+06', 'mean_token_accuracy': '0.6424', 'epoch': '0.4609'}
+{'loss': '1.819', 'grad_norm': '21', 'learning_rate': '5.391e-06', 'num_tokens': '1.758e+06', 'mean_token_accuracy': '0.6463', 'epoch': '0.4611'}
+{'loss': '1.825', 'grad_norm': '16', 'learning_rate': '5.389e-06', 'num_tokens': '1.759e+06', 'mean_token_accuracy': '0.6178', 'epoch': '0.4613'}
+{'loss': '1.381', 'grad_norm': '13.88', 'learning_rate': '5.387e-06', 'num_tokens': '1.76e+06', 'mean_token_accuracy': '0.6842', 'epoch': '0.4614'}
+{'loss': '1.149', 'grad_norm': '12.19', 'learning_rate': '5.386e-06', 'num_tokens': '1.76e+06', 'mean_token_accuracy': '0.6779', 'epoch': '0.4616'}
+{'loss': '2.288', 'grad_norm': '17.38', 'learning_rate': '5.384e-06', 'num_tokens': '1.761e+06', 'mean_token_accuracy': '0.5737', 'epoch': '0.4618'}
+{'loss': '1.388', 'grad_norm': '14.62', 'learning_rate': '5.382e-06', 'num_tokens': '1.761e+06', 'mean_token_accuracy': '0.6263', 'epoch': '0.462'}
+{'loss': '1.249', 'grad_norm': '11', 'learning_rate': '5.38e-06', 'num_tokens': '1.762e+06', 'mean_token_accuracy': '0.666', 'epoch': '0.4621'}
+{'loss': '0.5613', 'grad_norm': '29.88', 'learning_rate': '5.379e-06', 'num_tokens': '1.762e+06', 'mean_token_accuracy': '0.8537', 'epoch': '0.4623'}
+{'loss': '1.259', 'grad_norm': '17.75', 'learning_rate': '5.377e-06', 'num_tokens': '1.763e+06', 'mean_token_accuracy': '0.691', 'epoch': '0.4625'}
+{'loss': '1.578', 'grad_norm': '9.625', 'learning_rate': '5.375e-06', 'num_tokens': '1.764e+06', 'mean_token_accuracy': '0.6645', 'epoch': '0.4627'}
+{'loss': '1.072', 'grad_norm': '11.75', 'learning_rate': '5.373e-06', 'num_tokens': '1.765e+06', 'mean_token_accuracy': '0.7015', 'epoch': '0.4628'}
+
 46%|████▋     | 2666/5758 [15:09<16:32,  3.12it/s]
 46%|████▋     | 2667/5758 [15:09<17:33,  2.93it/s]
                                                   

 46%|████▋     | 2667/5758 [15:09<17:33,  2.93it/s]
 46%|████▋     | 2668/5758 [15:10<16:59,  3.03it/s]
                                                   

 46%|████▋     | 2668/5758 [15:10<16:59,  3.03it/s]
 46%|████▋     | 2669/5758 [15:10<16:29,  3.12it/s]
                                                   

 46%|████▋     | 2669/5758 [15:10<16:29,  3.12it/s]
 46%|████▋     | 2670/5758 [15:10<17:32,  2.93it/s]
                                                   

 46%|████▋     | 2670/5758 [15:10<17:32,  2.93it/s]
 46%|████▋     | 2671/5758 [15:11<17:23,  2.96it/s]
                                                   

 46%|████▋     | 2671/5758 [15:11<17:23,  2.96it/s]
 46%|████▋     | 2672/5758 [15:11<17:32,  2.93it/s]
                                                   

 46%|████▋     | 2672/5758 [15:11<17:32,  2.93it/s]
 46%|████▋     | 2673/5758 [15:11<17:11,  2.99it/s]
                                                   

 46%|████▋     | 2673/5758 [15:11<17:11,  2.99it/s]
 46%|████▋     | 2674/5758 [15:12<16:50,  3.05it/s]
                                                   

 46%|████▋     | 2674/5758 [15:12<16:50,  3.05it/s]
 46%|████▋     | 2675/5758 [15:12<17:06,  3.00it/s]
                                                   

 46%|████▋     | 2675/5758 [15:12<17:06,  3.00it/s]
 46%|████▋     | 2676/5758 [15:12<17:04,  3.01it/s]
                                                   

 46%|████▋     | 2676/5758 [15:12<17:04,  3.01it/s]
 46%|████▋     | 2677/5758 [15:13<16:33,  3.10it/s]
                                                   

 46%|████▋     | 2677/5758 [15:13<16:33,  3.10it/s]
 47%|████▋     | 2678/5758 [15:13<16:37,  3.09it/s]
                                                   

 47%|████▋     | 2678/5758 [15:13<16:37,  3.09it/s]
 47%|████▋     | 2679/5758 [15:13<16:54,  3.03it/s]
                                                   

 47%|████▋     | 2679/5758 [15:13<16:54,  3.03it/s]
 47%|████▋     | 2680/5758 [15:14<18:08,  2.83it/s]
                                                   

 47%|████▋     | 2680/5758 [15:14<18:08,  2.83it/s]
 47%|████▋     | 2681/5758 [15:14<18:30,  2.77it/s]
                                                   

 47%|████▋     | 2681/5758 [15:14<18:30,  2.77it/s]
 47%|████▋     | 2682/5758 [15:14<17:45,  2.89it/s]
                                                   

 47%|████▋     | 2682/5758 [15:14<17:45,  2.89it/s]
 47%|████▋     | 2683/5758 [15:15<17:46,  2.88it/s]
                                                   

 47%|████▋     | 2683/5758 [15:15<17:46,  2.88it/s]
 47%|████▋     | 2684/5758 [15:15<17:04,  3.00it/s]
                                                   

 47%|████▋     | 2684/5758 [15:15<17:04,  3.00it/s]
 47%|████▋     | 2685/5758 [15:15<16:44,  3.06it/s]
                                                   

 47%|████▋     | 2685/5758 [15:15<16:44,  3.06it/s]
 47%|████▋     | 2686/5758 [15:16<16:28,  3.11it/s]
                                                   

 47%|████▋     | 2686/5758 [15:16<16:28,  3.11it/s]
 47%|████▋     | 2687/5758 [15:16<16:23,  3.12it/s]
                                                   

 47%|████▋     | 2687/5758 [15:16<16:23,  3.12it/s]
 47%|████▋     | 2688/5758 [15:16<16:35,  3.08it/s]
                                                   

 47%|████▋     | 2688/5758 [15:16<16:35,  3.08it/s]
 47%|████▋     | 2689/5758 [15:17<16:47,  3.04it/s]
                                                   

 47%|████▋     | 2689/5758 [15:17<16:47,  3.04it/s]
 47%|████▋     | 2690/5758 [15:17<16:52,  3.03it/s]
                                                   

 47%|████▋     | 2690/5758 [15:17<16:52,  3.03it/s]
 47%|████▋     | 2691/5758 [15:17<16:45,  3.05it/s]
                                                   

 47%|████▋     | 2691/5758 [15:17<16:45,  3.05it/s]
 47%|████▋     | 2692/5758 [15:18<16:14,  3.15it/s]
                                                   

 47%|████▋     | 2692/5758 [15:18<16:14,  3.15it/s]
 47%|████▋     | 2693/5758 [15:18<16:00,  3.19it/s]
                                                   

 47%|████▋     | 2693/5758 [15:18<16:00,  3.19it/s]
 47%|████▋     | 2694/5758 [15:18<15:54,  3.21it/s]
                                                   

 47%|████▋     | 2694/5758 [15:18<15:54,  3.21it/s]
 47%|████▋     | 2695/5758 [15:19<15:42,  3.25it/s]
                                                   

 47%|████▋     | 2695/5758 [15:19<15:42,  3.25it/s]
 47%|████▋     | 2696/5758 [15:19<15:37,  3.27it/s]
                                                   

 47%|████▋     | 2696/5758 [15:19<15:37,  3.27it/s]
 47%|████▋     | 2697/5758 [15:19<16:41,  3.06it/s]
                                                   

 47%|████▋     | 2697/5758 [15:19<16:41,  3.06it/s]
 47%|████▋     | 2698/5758 [15:20<16:31,  3.09it/s]
                                                   

 47%|████▋     | 2698/5758 [15:20<16:31,  3.09it/s]
 47%|████▋     | 2699/5758 [15:20<15:57,  3.19it/s]
                                                   

 47%|████▋     | 2699/5758 [15:20<15:57,  3.19it/s]
 47%|████▋     | 2700/5758 [15:20<16:07,  3.16it/s]
                                                   

 47%|████▋     | 2700/5758 [15:20<16:07,  3.16it/s]
 47%|████▋     | 2701/5758 [15:21<15:54,  3.20it/s]
                                                   

 47%|████▋     | 2701/5758 [15:21<15:54,  3.20it/s]
 47%|████▋     | 2702/5758 [15:21<16:40,  3.05it/s]
                                                   

 47%|████▋     | 2702/5758 [15:21<16:40,  3.05it/s]
 47%|████▋     | 2703/5758 [15:21<15:58,  3.19it/s]
                                                   

 47%|████▋     | 2703/5758 [15:21<15:58,  3.19it/s]
 47%|████▋     | 2704/5758 [15:22<16:53,  3.01it/s]
                                                   

 47%|████▋     | 2704/5758 [15:22<16:53,  3.01it/s]
 47%|████▋     | 2705/5758 [15:22<16:19,  3.12it/s]
                                                   

 47%|████▋     | 2705/5758 [15:22<16:19,  3.12it/s]
 47%|████▋     | 2706/5758 [15:22<19:02,  2.67it/s]
                                                   

 47%|████▋     | 2706/5758 [15:22<19:02,  2.67it/s]
 47%|████▋     | 2707/5758 [15:23<18:11,  2.80it/s]
                                                   

 47%|████▋     | 2707/5758 [15:23<18:11,  2.80it/s]
 47%|████▋     | 2708/5758 [15:23<17:23,  2.92it/s]
                                                   

 47%|████▋     | 2708/5758 [15:23<17:23,  2.92it/s]
 47%|████▋     | 2709/5758 [15:23<16:53,  3.01it/s]
                                                   

 47%|████▋     | 2709/5758 [15:23<16:53,  3.01it/s]
 47%|████▋     | 2710/5758 [15:24<16:21,  3.11it/s]
                                                   

 47%|████▋     | 2710/5758 [15:24<16:21,  3.11it/s]
 47%|████▋     | 2711/5758 [15:24<15:57,  3.18it/s]
                                                   

 47%|████▋     | 2711/5758 [15:24<15:57,  3.18it/s]
 47%|████▋     | 2712/5758 [15:24<17:09,  2.96it/s]
                                                   

 47%|████▋     | 2712/5758 [15:24<17:09,  2.96it/s]
 47%|████▋     | 2713/5758 [15:25<18:08,  2.80it/s]
                                                   

 47%|████▋     | 2713/5758 [15:25<18:08,  2.80it/s]
 47%|████▋     | 2714/5758 [15:25<18:30,  2.74it/s]
                                                   

 47%|████▋     | 2714/5758 [15:25<18:30,  2.74it/s]
 47%|████▋     | 2715/5758 [15:25<17:44,  2.86it/s]
                                                   

 47%|████▋     | 2715/5758 [15:25<17:44,  2.86it/s]
 47%|████▋     | 2716/5758 [15:26<17:26,  2.91it/s]
                                                   

 47%|████▋     | 2716/5758 [15:26<17:26,  2.91it/s]
 47%|████▋     | 2717/5758 [15:26<17:31,  2.89it/s]
                                                   

 47%|████▋     | 2717/5758 [15:26<17:31,  2.89it/s]
 47%|████▋     | 2718/5758 [15:26<16:55,  2.99it/s]
                                                   

 47%|████▋     | 2718/5758 [15:26<16:55,  2.99it/s]
 47%|████▋     | 2719/5758 [15:27<16:19,  3.10it/s]
                                                   

 47%|████▋     | 2719/5758 [15:27<16:19,  3.10it/s]
 47%|████▋     | 2720/5758 [15:27<15:54,  3.18it/s]
                                                   

 47%|████▋     | 2720/5758 [15:27<15:54,  3.18it/s]
 47%|████▋     | 2721/5758 [15:27<15:37,  3.24it/s]
                                                   
{'loss': '1.747', 'grad_norm': '23.88', 'learning_rate': '5.372e-06', 'num_tokens': '1.765e+06', 'mean_token_accuracy': '0.5944', 'epoch': '0.463'}
+{'loss': '1.144', 'grad_norm': '19.12', 'learning_rate': '5.37e-06', 'num_tokens': '1.766e+06', 'mean_token_accuracy': '0.7363', 'epoch': '0.4632'}
+{'loss': '1.611', 'grad_norm': '14.06', 'learning_rate': '5.368e-06', 'num_tokens': '1.766e+06', 'mean_token_accuracy': '0.6468', 'epoch': '0.4634'}
+{'loss': '0.629', 'grad_norm': '7.75', 'learning_rate': '5.366e-06', 'num_tokens': '1.768e+06', 'mean_token_accuracy': '0.8156', 'epoch': '0.4635'}
+{'loss': '0.8422', 'grad_norm': '23.62', 'learning_rate': '5.365e-06', 'num_tokens': '1.768e+06', 'mean_token_accuracy': '0.8217', 'epoch': '0.4637'}
+{'loss': '0.9684', 'grad_norm': '14.81', 'learning_rate': '5.363e-06', 'num_tokens': '1.768e+06', 'mean_token_accuracy': '0.7566', 'epoch': '0.4639'}
+{'loss': '1.597', 'grad_norm': '21.88', 'learning_rate': '5.361e-06', 'num_tokens': '1.769e+06', 'mean_token_accuracy': '0.6723', 'epoch': '0.4641'}
+{'loss': '1.449', 'grad_norm': '8.5', 'learning_rate': '5.359e-06', 'num_tokens': '1.77e+06', 'mean_token_accuracy': '0.6616', 'epoch': '0.4642'}
+{'loss': '1.086', 'grad_norm': '12', 'learning_rate': '5.358e-06', 'num_tokens': '1.771e+06', 'mean_token_accuracy': '0.6799', 'epoch': '0.4644'}
+{'loss': '1.674', 'grad_norm': '7.688', 'learning_rate': '5.356e-06', 'num_tokens': '1.773e+06', 'mean_token_accuracy': '0.6156', 'epoch': '0.4646'}
+{'loss': '0.6273', 'grad_norm': '7.219', 'learning_rate': '5.354e-06', 'num_tokens': '1.774e+06', 'mean_token_accuracy': '0.8031', 'epoch': '0.4647'}
+{'loss': '1.722', 'grad_norm': '42.25', 'learning_rate': '5.353e-06', 'num_tokens': '1.774e+06', 'mean_token_accuracy': '0.6053', 'epoch': '0.4649'}
+{'loss': '1.389', 'grad_norm': '30.75', 'learning_rate': '5.351e-06', 'num_tokens': '1.774e+06', 'mean_token_accuracy': '0.7449', 'epoch': '0.4651'}
+{'loss': '1.755', 'grad_norm': '22.88', 'learning_rate': '5.349e-06', 'num_tokens': '1.775e+06', 'mean_token_accuracy': '0.652', 'epoch': '0.4653'}
+{'loss': '0.91', 'grad_norm': '12.94', 'learning_rate': '5.347e-06', 'num_tokens': '1.775e+06', 'mean_token_accuracy': '0.737', 'epoch': '0.4654'}
+{'loss': '1.086', 'grad_norm': '5.969', 'learning_rate': '5.346e-06', 'num_tokens': '1.778e+06', 'mean_token_accuracy': '0.7569', 'epoch': '0.4656'}
+{'loss': '1.245', 'grad_norm': '12.12', 'learning_rate': '5.344e-06', 'num_tokens': '1.778e+06', 'mean_token_accuracy': '0.6657', 'epoch': '0.4658'}
+{'loss': '1.174', 'grad_norm': '5.844', 'learning_rate': '5.342e-06', 'num_tokens': '1.782e+06', 'mean_token_accuracy': '0.7205', 'epoch': '0.466'}
+{'loss': '1.437', 'grad_norm': '29.88', 'learning_rate': '5.34e-06', 'num_tokens': '1.782e+06', 'mean_token_accuracy': '0.7296', 'epoch': '0.4661'}
+{'loss': '1.978', 'grad_norm': '11.25', 'learning_rate': '5.339e-06', 'num_tokens': '1.783e+06', 'mean_token_accuracy': '0.6138', 'epoch': '0.4663'}
+{'loss': '1.078', 'grad_norm': '7.312', 'learning_rate': '5.337e-06', 'num_tokens': '1.784e+06', 'mean_token_accuracy': '0.7612', 'epoch': '0.4665'}
+{'loss': '1.617', 'grad_norm': '22.12', 'learning_rate': '5.335e-06', 'num_tokens': '1.785e+06', 'mean_token_accuracy': '0.6686', 'epoch': '0.4667'}
+{'loss': '2.316', 'grad_norm': '28.75', 'learning_rate': '5.333e-06', 'num_tokens': '1.785e+06', 'mean_token_accuracy': '0.5426', 'epoch': '0.4668'}
+{'loss': '0.6203', 'grad_norm': '6.125', 'learning_rate': '5.332e-06', 'num_tokens': '1.787e+06', 'mean_token_accuracy': '0.8198', 'epoch': '0.467'}
+{'loss': '2.255', 'grad_norm': '22.88', 'learning_rate': '5.33e-06', 'num_tokens': '1.787e+06', 'mean_token_accuracy': '0.5749', 'epoch': '0.4672'}
+{'loss': '0.2268', 'grad_norm': '24', 'learning_rate': '5.328e-06', 'num_tokens': '1.787e+06', 'mean_token_accuracy': '0.9368', 'epoch': '0.4673'}
+{'loss': '1.648', 'grad_norm': '27.75', 'learning_rate': '5.327e-06', 'num_tokens': '1.787e+06', 'mean_token_accuracy': '0.6456', 'epoch': '0.4675'}
+{'loss': '1.902', 'grad_norm': '14.88', 'learning_rate': '5.325e-06', 'num_tokens': '1.788e+06', 'mean_token_accuracy': '0.6008', 'epoch': '0.4677'}
+{'loss': '0.6556', 'grad_norm': '8.5', 'learning_rate': '5.323e-06', 'num_tokens': '1.789e+06', 'mean_token_accuracy': '0.7936', 'epoch': '0.4679'}
+{'loss': '0.9349', 'grad_norm': '11.94', 'learning_rate': '5.321e-06', 'num_tokens': '1.79e+06', 'mean_token_accuracy': '0.7139', 'epoch': '0.468'}
+{'loss': '1.679', 'grad_norm': '16.88', 'learning_rate': '5.32e-06', 'num_tokens': '1.79e+06', 'mean_token_accuracy': '0.6524', 'epoch': '0.4682'}
+{'loss': '1.3', 'grad_norm': '9.062', 'learning_rate': '5.318e-06', 'num_tokens': '1.792e+06', 'mean_token_accuracy': '0.7286', 'epoch': '0.4684'}
+{'loss': '1.051', 'grad_norm': '11', 'learning_rate': '5.316e-06', 'num_tokens': '1.792e+06', 'mean_token_accuracy': '0.7123', 'epoch': '0.4686'}
+{'loss': '0.8756', 'grad_norm': '13.5', 'learning_rate': '5.314e-06', 'num_tokens': '1.793e+06', 'mean_token_accuracy': '0.752', 'epoch': '0.4687'}
+{'loss': '1.258', 'grad_norm': '12.75', 'learning_rate': '5.313e-06', 'num_tokens': '1.794e+06', 'mean_token_accuracy': '0.6582', 'epoch': '0.4689'}
+{'loss': '1.056', 'grad_norm': '12.06', 'learning_rate': '5.311e-06', 'num_tokens': '1.794e+06', 'mean_token_accuracy': '0.7037', 'epoch': '0.4691'}
+{'loss': '1.855', 'grad_norm': '17.75', 'learning_rate': '5.309e-06', 'num_tokens': '1.795e+06', 'mean_token_accuracy': '0.5958', 'epoch': '0.4693'}
+{'loss': '1.198', 'grad_norm': '56.75', 'learning_rate': '5.307e-06', 'num_tokens': '1.795e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.4694'}
+{'loss': '1.948', 'grad_norm': '9.375', 'learning_rate': '5.306e-06', 'num_tokens': '1.796e+06', 'mean_token_accuracy': '0.5549', 'epoch': '0.4696'}
+{'loss': '1.106', 'grad_norm': '13.44', 'learning_rate': '5.304e-06', 'num_tokens': '1.797e+06', 'mean_token_accuracy': '0.6878', 'epoch': '0.4698'}
+{'loss': '0.7226', 'grad_norm': '10.94', 'learning_rate': '5.302e-06', 'num_tokens': '1.797e+06', 'mean_token_accuracy': '0.8238', 'epoch': '0.47'}
+{'loss': '2.408', 'grad_norm': '13.94', 'learning_rate': '5.3e-06', 'num_tokens': '1.798e+06', 'mean_token_accuracy': '0.5177', 'epoch': '0.4701'}
+{'loss': '1.136', 'grad_norm': '12.12', 'learning_rate': '5.299e-06', 'num_tokens': '1.799e+06', 'mean_token_accuracy': '0.6837', 'epoch': '0.4703'}
+{'loss': '2.197', 'grad_norm': '20.12', 'learning_rate': '5.297e-06', 'num_tokens': '1.799e+06', 'mean_token_accuracy': '0.5658', 'epoch': '0.4705'}
+{'loss': '1.728', 'grad_norm': '20', 'learning_rate': '5.295e-06', 'num_tokens': '1.799e+06', 'mean_token_accuracy': '0.6739', 'epoch': '0.4706'}
+{'loss': '1.286', 'grad_norm': '9.188', 'learning_rate': '5.294e-06', 'num_tokens': '1.801e+06', 'mean_token_accuracy': '0.6906', 'epoch': '0.4708'}
+{'loss': '0.4642', 'grad_norm': '7.844', 'learning_rate': '5.292e-06', 'num_tokens': '1.802e+06', 'mean_token_accuracy': '0.8585', 'epoch': '0.471'}
+{'loss': '1.583', 'grad_norm': '41', 'learning_rate': '5.29e-06', 'num_tokens': '1.802e+06', 'mean_token_accuracy': '0.6571', 'epoch': '0.4712'}
+{'loss': '0.5807', 'grad_norm': '16', 'learning_rate': '5.288e-06', 'num_tokens': '1.802e+06', 'mean_token_accuracy': '0.8566', 'epoch': '0.4713'}
+{'loss': '1.322', 'grad_norm': '14.75', 'learning_rate': '5.287e-06', 'num_tokens': '1.803e+06', 'mean_token_accuracy': '0.6611', 'epoch': '0.4715'}
+{'loss': '1.34', 'grad_norm': '7.281', 'learning_rate': '5.285e-06', 'num_tokens': '1.805e+06', 'mean_token_accuracy': '0.6894', 'epoch': '0.4717'}
+{'loss': '1.838', 'grad_norm': '21.88', 'learning_rate': '5.283e-06', 'num_tokens': '1.805e+06', 'mean_token_accuracy': '0.6078', 'epoch': '0.4719'}
+{'loss': '1.284', 'grad_norm': '13.56', 'learning_rate': '5.281e-06', 'num_tokens': '1.806e+06', 'mean_token_accuracy': '0.6449', 'epoch': '0.472'}
+{'loss': '0.7815', 'grad_norm': '10.69', 'learning_rate': '5.28e-06', 'num_tokens': '1.806e+06', 'mean_token_accuracy': '0.8172', 'epoch': '0.4722'}
+{'loss': '0.7484', 'grad_norm': '17.12', 'learning_rate': '5.278e-06', 'num_tokens': '1.806e+06', 'mean_token_accuracy': '0.8268', 'epoch': '0.4724'}
+
 47%|████▋     | 2721/5758 [15:27<15:37,  3.24it/s]
 47%|████▋     | 2722/5758 [15:28<15:35,  3.25it/s]
                                                   

 47%|████▋     | 2722/5758 [15:28<15:35,  3.25it/s]
 47%|████▋     | 2723/5758 [15:28<16:44,  3.02it/s]
                                                   

 47%|████▋     | 2723/5758 [15:28<16:44,  3.02it/s]
 47%|████▋     | 2724/5758 [15:28<17:33,  2.88it/s]
                                                   

 47%|████▋     | 2724/5758 [15:28<17:33,  2.88it/s]
 47%|████▋     | 2725/5758 [15:29<18:05,  2.79it/s]
                                                   

 47%|████▋     | 2725/5758 [15:29<18:05,  2.79it/s]
 47%|████▋     | 2726/5758 [15:29<17:14,  2.93it/s]
                                                   

 47%|████▋     | 2726/5758 [15:29<17:14,  2.93it/s]
 47%|████▋     | 2727/5758 [15:29<16:46,  3.01it/s]
                                                   

 47%|████▋     | 2727/5758 [15:29<16:46,  3.01it/s]
 47%|████▋     | 2728/5758 [15:30<16:19,  3.09it/s]
                                                   

 47%|████▋     | 2728/5758 [15:30<16:19,  3.09it/s]
 47%|████▋     | 2729/5758 [15:30<16:11,  3.12it/s]
                                                   

 47%|████▋     | 2729/5758 [15:30<16:11,  3.12it/s]
 47%|████▋     | 2730/5758 [15:30<16:03,  3.14it/s]
                                                   

 47%|████▋     | 2730/5758 [15:30<16:03,  3.14it/s]
 47%|████▋     | 2731/5758 [15:31<16:16,  3.10it/s]
                                                   

 47%|████▋     | 2731/5758 [15:31<16:16,  3.10it/s]
 47%|████▋     | 2732/5758 [15:31<17:40,  2.85it/s]
                                                   

 47%|████▋     | 2732/5758 [15:31<17:40,  2.85it/s]
 47%|████▋     | 2733/5758 [15:31<18:12,  2.77it/s]
                                                   

 47%|████▋     | 2733/5758 [15:31<18:12,  2.77it/s]
 47%|████▋     | 2734/5758 [15:32<18:20,  2.75it/s]
                                                   

 47%|████▋     | 2734/5758 [15:32<18:20,  2.75it/s]
 47%|████▋     | 2735/5758 [15:32<17:40,  2.85it/s]
                                                   

 47%|████▋     | 2735/5758 [15:32<17:40,  2.85it/s]
 48%|████▊     | 2736/5758 [15:32<17:31,  2.87it/s]
                                                   

 48%|████▊     | 2736/5758 [15:32<17:31,  2.87it/s]
 48%|████▊     | 2737/5758 [15:33<17:05,  2.94it/s]
                                                   

 48%|████▊     | 2737/5758 [15:33<17:05,  2.94it/s]
 48%|████▊     | 2738/5758 [15:33<16:24,  3.07it/s]
                                                   

 48%|████▊     | 2738/5758 [15:33<16:24,  3.07it/s]
 48%|████▊     | 2739/5758 [15:33<17:21,  2.90it/s]
                                                   

 48%|████▊     | 2739/5758 [15:33<17:21,  2.90it/s]
 48%|████▊     | 2740/5758 [15:34<17:16,  2.91it/s]
                                                   

 48%|████▊     | 2740/5758 [15:34<17:16,  2.91it/s]
 48%|████▊     | 2741/5758 [15:34<16:46,  3.00it/s]
                                                   

 48%|████▊     | 2741/5758 [15:34<16:46,  3.00it/s]
 48%|████▊     | 2742/5758 [15:34<16:29,  3.05it/s]
                                                   

 48%|████▊     | 2742/5758 [15:34<16:29,  3.05it/s]
 48%|████▊     | 2743/5758 [15:35<16:45,  3.00it/s]
                                                   

 48%|████▊     | 2743/5758 [15:35<16:45,  3.00it/s]
 48%|████▊     | 2744/5758 [15:35<16:18,  3.08it/s]
                                                   

 48%|████▊     | 2744/5758 [15:35<16:18,  3.08it/s]
 48%|████▊     | 2745/5758 [15:35<15:46,  3.18it/s]
                                                   

 48%|████▊     | 2745/5758 [15:35<15:46,  3.18it/s]
 48%|████▊     | 2746/5758 [15:36<16:42,  3.00it/s]
                                                   

 48%|████▊     | 2746/5758 [15:36<16:42,  3.00it/s]
 48%|████▊     | 2747/5758 [15:36<16:32,  3.03it/s]
                                                   

 48%|████▊     | 2747/5758 [15:36<16:32,  3.03it/s]
 48%|████▊     | 2748/5758 [15:36<16:04,  3.12it/s]
                                                   

 48%|████▊     | 2748/5758 [15:36<16:04,  3.12it/s]
 48%|████▊     | 2749/5758 [15:37<16:19,  3.07it/s]
                                                   

 48%|████▊     | 2749/5758 [15:37<16:19,  3.07it/s]
 48%|████▊     | 2750/5758 [15:37<16:33,  3.03it/s]
                                                   

 48%|████▊     | 2750/5758 [15:37<16:33,  3.03it/s]
 48%|████▊     | 2751/5758 [15:37<16:41,  3.00it/s]
                                                   

 48%|████▊     | 2751/5758 [15:37<16:41,  3.00it/s]
 48%|████▊     | 2752/5758 [15:38<16:20,  3.07it/s]
                                                   

 48%|████▊     | 2752/5758 [15:38<16:20,  3.07it/s]
 48%|████▊     | 2753/5758 [15:38<16:03,  3.12it/s]
                                                   

 48%|████▊     | 2753/5758 [15:38<16:03,  3.12it/s]
 48%|████▊     | 2754/5758 [15:38<15:53,  3.15it/s]
                                                   

 48%|████▊     | 2754/5758 [15:38<15:53,  3.15it/s]
 48%|████▊     | 2755/5758 [15:39<16:45,  2.99it/s]
                                                   

 48%|████▊     | 2755/5758 [15:39<16:45,  2.99it/s]
 48%|████▊     | 2756/5758 [15:39<16:10,  3.09it/s]
                                                   

 48%|████▊     | 2756/5758 [15:39<16:10,  3.09it/s]
 48%|████▊     | 2757/5758 [15:39<16:50,  2.97it/s]
                                                   

 48%|████▊     | 2757/5758 [15:39<16:50,  2.97it/s]
 48%|████▊     | 2758/5758 [15:40<17:08,  2.92it/s]
                                                   

 48%|█���██▊     | 2758/5758 [15:40<17:08,  2.92it/s]
 48%|████▊     | 2759/5758 [15:40<18:30,  2.70it/s]
                                                   

 48%|████▊     | 2759/5758 [15:40<18:30,  2.70it/s]
 48%|████▊     | 2760/5758 [15:40<17:53,  2.79it/s]
                                                   

 48%|████▊     | 2760/5758 [15:40<17:53,  2.79it/s]
 48%|████▊     | 2761/5758 [15:41<17:08,  2.91it/s]
                                                   

 48%|████▊     | 2761/5758 [15:41<17:08,  2.91it/s]
 48%|████▊     | 2762/5758 [15:41<16:38,  3.00it/s]
                                                   

 48%|████▊     | 2762/5758 [15:41<16:38,  3.00it/s]
 48%|████▊     | 2763/5758 [15:41<16:57,  2.94it/s]
                                                   

 48%|████▊     | 2763/5758 [15:41<16:57,  2.94it/s]
 48%|████▊     | 2764/5758 [15:42<17:12,  2.90it/s]
                                                   

 48%|████▊     | 2764/5758 [15:42<17:12,  2.90it/s]
 48%|████▊     | 2765/5758 [15:42<16:37,  3.00it/s]
                                                   

 48%|████▊     | 2765/5758 [15:42<16:37,  3.00it/s]
 48%|████▊     | 2766/5758 [15:42<16:30,  3.02it/s]
                                                   

 48%|████▊     | 2766/5758 [15:42<16:30,  3.02it/s]
 48%|████▊     | 2767/5758 [15:43<17:14,  2.89it/s]
                                                   

 48%|████▊     | 2767/5758 [15:43<17:14,  2.89it/s]
 48%|████▊     | 2768/5758 [15:43<17:10,  2.90it/s]
                                                   

 48%|████▊     | 2768/5758 [15:43<17:10,  2.90it/s]
 48%|████▊     | 2769/5758 [15:43<16:44,  2.97it/s]
                                                   

 48%|████▊     | 2769/5758 [15:43<16:44,  2.97it/s]
 48%|████▊     | 2770/5758 [15:44<17:17,  2.88it/s]
                                                   

 48%|████▊     | 2770/5758 [15:44<17:17,  2.88it/s]
 48%|████▊     | 2771/5758 [15:44<16:49,  2.96it/s]
                                                   

 48%|████▊     | 2771/5758 [15:44<16:49,  2.96it/s]
 48%|████▊     | 2772/5758 [15:44<17:18,  2.88it/s]
                                                   

 48%|████▊     | 2772/5758 [15:44<17:18,  2.88it/s]
 48%|████▊     | 2773/5758 [15:45<16:43,  2.98it/s]
                                                   

 48%|████▊     | 2773/5758 [15:45<16:43,  2.98it/s]
 48%|████▊     | 2774/5758 [15:45<17:20,  2.87it/s]
                                                   

 48%|████▊     | 2774/5758 [15:45<17:20,  2.87it/s]
 48%|████▊     | 2775/5758 [15:45<16:30,  3.01it/s]
                                                   

 48%|████▊     | 2775/5758 [15:45<16:30,  3.01it/s]
 48%|████▊     | 2776/5758 [15:46<17:04,  2.91it/s]
                                                   
{'loss': '2.234', 'grad_norm': '18.12', 'learning_rate': '5.276e-06', 'num_tokens': '1.807e+06', 'mean_token_accuracy': '0.547', 'epoch': '0.4726'}
+{'loss': '1.169', 'grad_norm': '13.44', 'learning_rate': '5.274e-06', 'num_tokens': '1.808e+06', 'mean_token_accuracy': '0.6733', 'epoch': '0.4727'}
+{'loss': '1.002', 'grad_norm': '8.938', 'learning_rate': '5.273e-06', 'num_tokens': '1.809e+06', 'mean_token_accuracy': '0.7518', 'epoch': '0.4729'}
+{'loss': '1.214', 'grad_norm': '14.38', 'learning_rate': '5.271e-06', 'num_tokens': '1.81e+06', 'mean_token_accuracy': '0.7175', 'epoch': '0.4731'}
+{'loss': '1.966', 'grad_norm': '10.31', 'learning_rate': '5.269e-06', 'num_tokens': '1.811e+06', 'mean_token_accuracy': '0.5478', 'epoch': '0.4733'}
+{'loss': '1.499', 'grad_norm': '21', 'learning_rate': '5.267e-06', 'num_tokens': '1.811e+06', 'mean_token_accuracy': '0.6789', 'epoch': '0.4734'}
+{'loss': '0.5836', 'grad_norm': '34.25', 'learning_rate': '5.266e-06', 'num_tokens': '1.811e+06', 'mean_token_accuracy': '0.8481', 'epoch': '0.4736'}
+{'loss': '0.4016', 'grad_norm': '18.75', 'learning_rate': '5.264e-06', 'num_tokens': '1.811e+06', 'mean_token_accuracy': '0.8696', 'epoch': '0.4738'}
+{'loss': '0.9983', 'grad_norm': '9.688', 'learning_rate': '5.262e-06', 'num_tokens': '1.812e+06', 'mean_token_accuracy': '0.7797', 'epoch': '0.4739'}
+{'loss': '1.166', 'grad_norm': '12.5', 'learning_rate': '5.261e-06', 'num_tokens': '1.813e+06', 'mean_token_accuracy': '0.694', 'epoch': '0.4741'}
+{'loss': '1.649', 'grad_norm': '15', 'learning_rate': '5.259e-06', 'num_tokens': '1.813e+06', 'mean_token_accuracy': '0.678', 'epoch': '0.4743'}
+{'loss': '2.106', 'grad_norm': '35.25', 'learning_rate': '5.257e-06', 'num_tokens': '1.813e+06', 'mean_token_accuracy': '0.5962', 'epoch': '0.4745'}
+{'loss': '1.982', 'grad_norm': '17.5', 'learning_rate': '5.255e-06', 'num_tokens': '1.814e+06', 'mean_token_accuracy': '0.5565', 'epoch': '0.4746'}
+{'loss': '2.332', 'grad_norm': '8.688', 'learning_rate': '5.254e-06', 'num_tokens': '1.816e+06', 'mean_token_accuracy': '0.54', 'epoch': '0.4748'}
+{'loss': '0.7582', 'grad_norm': '14.44', 'learning_rate': '5.252e-06', 'num_tokens': '1.816e+06', 'mean_token_accuracy': '0.8218', 'epoch': '0.475'}
+{'loss': '1.364', 'grad_norm': '12.31', 'learning_rate': '5.25e-06', 'num_tokens': '1.817e+06', 'mean_token_accuracy': '0.6326', 'epoch': '0.4752'}
+{'loss': '1.035', 'grad_norm': '5.344', 'learning_rate': '5.248e-06', 'num_tokens': '1.82e+06', 'mean_token_accuracy': '0.7369', 'epoch': '0.4753'}
+{'loss': '1.947', 'grad_norm': '13.81', 'learning_rate': '5.247e-06', 'num_tokens': '1.821e+06', 'mean_token_accuracy': '0.555', 'epoch': '0.4755'}
+{'loss': '1.197', 'grad_norm': '15.44', 'learning_rate': '5.245e-06', 'num_tokens': '1.821e+06', 'mean_token_accuracy': '0.6367', 'epoch': '0.4757'}
+{'loss': '1.435', 'grad_norm': '10.06', 'learning_rate': '5.243e-06', 'num_tokens': '1.822e+06', 'mean_token_accuracy': '0.6934', 'epoch': '0.4759'}
+{'loss': '0.7411', 'grad_norm': '18.75', 'learning_rate': '5.241e-06', 'num_tokens': '1.823e+06', 'mean_token_accuracy': '0.8348', 'epoch': '0.476'}
+{'loss': '1.4', 'grad_norm': '18.75', 'learning_rate': '5.24e-06', 'num_tokens': '1.823e+06', 'mean_token_accuracy': '0.6813', 'epoch': '0.4762'}
+{'loss': '1.559', 'grad_norm': '13.81', 'learning_rate': '5.238e-06', 'num_tokens': '1.823e+06', 'mean_token_accuracy': '0.6644', 'epoch': '0.4764'}
+{'loss': '1.293', 'grad_norm': '13.88', 'learning_rate': '5.236e-06', 'num_tokens': '1.824e+06', 'mean_token_accuracy': '0.6519', 'epoch': '0.4766'}
+{'loss': '1.023', 'grad_norm': '12.31', 'learning_rate': '5.234e-06', 'num_tokens': '1.825e+06', 'mean_token_accuracy': '0.7061', 'epoch': '0.4767'}
+{'loss': '1.45', 'grad_norm': '20.25', 'learning_rate': '5.233e-06', 'num_tokens': '1.825e+06', 'mean_token_accuracy': '0.7153', 'epoch': '0.4769'}
+{'loss': '1.63', 'grad_norm': '18.38', 'learning_rate': '5.231e-06', 'num_tokens': '1.826e+06', 'mean_token_accuracy': '0.6364', 'epoch': '0.4771'}
+{'loss': '1.811', 'grad_norm': '15.75', 'learning_rate': '5.229e-06', 'num_tokens': '1.826e+06', 'mean_token_accuracy': '0.6349', 'epoch': '0.4772'}
+{'loss': '1.271', 'grad_norm': '11.81', 'learning_rate': '5.228e-06', 'num_tokens': '1.827e+06', 'mean_token_accuracy': '0.6586', 'epoch': '0.4774'}
+{'loss': '1.419', 'grad_norm': '12.19', 'learning_rate': '5.226e-06', 'num_tokens': '1.828e+06', 'mean_token_accuracy': '0.631', 'epoch': '0.4776'}
+{'loss': '1.108', 'grad_norm': '20.62', 'learning_rate': '5.224e-06', 'num_tokens': '1.828e+06', 'mean_token_accuracy': '0.7811', 'epoch': '0.4778'}
+{'loss': '1.074', 'grad_norm': '12.31', 'learning_rate': '5.222e-06', 'num_tokens': '1.829e+06', 'mean_token_accuracy': '0.6886', 'epoch': '0.4779'}
+{'loss': '1.381', 'grad_norm': '12.62', 'learning_rate': '5.221e-06', 'num_tokens': '1.83e+06', 'mean_token_accuracy': '0.6626', 'epoch': '0.4781'}
+{'loss': '0.6483', 'grad_norm': '11.56', 'learning_rate': '5.219e-06', 'num_tokens': '1.831e+06', 'mean_token_accuracy': '0.8231', 'epoch': '0.4783'}
+{'loss': '1.067', 'grad_norm': '14.12', 'learning_rate': '5.217e-06', 'num_tokens': '1.831e+06', 'mean_token_accuracy': '0.7203', 'epoch': '0.4785'}
+{'loss': '0.6575', 'grad_norm': '7.969', 'learning_rate': '5.215e-06', 'num_tokens': '1.832e+06', 'mean_token_accuracy': '0.8097', 'epoch': '0.4786'}
+{'loss': '1.082', 'grad_norm': '13.69', 'learning_rate': '5.214e-06', 'num_tokens': '1.833e+06', 'mean_token_accuracy': '0.7038', 'epoch': '0.4788'}
+{'loss': '1.829', 'grad_norm': '22.38', 'learning_rate': '5.212e-06', 'num_tokens': '1.833e+06', 'mean_token_accuracy': '0.6048', 'epoch': '0.479'}
+{'loss': '1.194', 'grad_norm': '12.56', 'learning_rate': '5.21e-06', 'num_tokens': '1.834e+06', 'mean_token_accuracy': '0.6611', 'epoch': '0.4792'}
+{'loss': '0.9507', 'grad_norm': '10.81', 'learning_rate': '5.208e-06', 'num_tokens': '1.835e+06', 'mean_token_accuracy': '0.7194', 'epoch': '0.4793'}
+{'loss': '1.152', 'grad_norm': '12.44', 'learning_rate': '5.207e-06', 'num_tokens': '1.836e+06', 'mean_token_accuracy': '0.6819', 'epoch': '0.4795'}
+{'loss': '0.9441', 'grad_norm': '22.75', 'learning_rate': '5.205e-06', 'num_tokens': '1.836e+06', 'mean_token_accuracy': '0.7947', 'epoch': '0.4797'}
+{'loss': '0.5742', 'grad_norm': '6.625', 'learning_rate': '5.203e-06', 'num_tokens': '1.838e+06', 'mean_token_accuracy': '0.8239', 'epoch': '0.4799'}
+{'loss': '2.22', 'grad_norm': '16.88', 'learning_rate': '5.201e-06', 'num_tokens': '1.838e+06', 'mean_token_accuracy': '0.5687', 'epoch': '0.48'}
+{'loss': '1.531', 'grad_norm': '24', 'learning_rate': '5.2e-06', 'num_tokens': '1.839e+06', 'mean_token_accuracy': '0.7167', 'epoch': '0.4802'}
+{'loss': '1.623', 'grad_norm': '14.06', 'learning_rate': '5.198e-06', 'num_tokens': '1.839e+06', 'mean_token_accuracy': '0.6241', 'epoch': '0.4804'}
+{'loss': '1.177', 'grad_norm': '14.06', 'learning_rate': '5.196e-06', 'num_tokens': '1.84e+06', 'mean_token_accuracy': '0.6706', 'epoch': '0.4805'}
+{'loss': '1.457', 'grad_norm': '12.06', 'learning_rate': '5.195e-06', 'num_tokens': '1.841e+06', 'mean_token_accuracy': '0.6112', 'epoch': '0.4807'}
+{'loss': '0.748', 'grad_norm': '9.188', 'learning_rate': '5.193e-06', 'num_tokens': '1.842e+06', 'mean_token_accuracy': '0.7624', 'epoch': '0.4809'}
+{'loss': '2.812', 'grad_norm': '30.38', 'learning_rate': '5.191e-06', 'num_tokens': '1.842e+06', 'mean_token_accuracy': '0.5143', 'epoch': '0.4811'}
+{'loss': '0.9632', 'grad_norm': '20.62', 'learning_rate': '5.189e-06', 'num_tokens': '1.843e+06', 'mean_token_accuracy': '0.8482', 'epoch': '0.4812'}
+{'loss': '0.753', 'grad_norm': '12.75', 'learning_rate': '5.188e-06', 'num_tokens': '1.843e+06', 'mean_token_accuracy': '0.8023', 'epoch': '0.4814'}
+{'loss': '2.606', 'grad_norm': '9.5', 'learning_rate': '5.186e-06', 'num_tokens': '1.844e+06', 'mean_token_accuracy': '0.4957', 'epoch': '0.4816'}
+{'loss': '0.6102', 'grad_norm': '9.688', 'learning_rate': '5.184e-06', 'num_tokens': '1.846e+06', 'mean_token_accuracy': '0.8016', 'epoch': '0.4818'}
+{'loss': '1.36', 'grad_norm': '14.81', 'learning_rate': '5.182e-06', 'num_tokens': '1.846e+06', 'mean_token_accuracy': '0.6341', 'epoch': '0.4819'}
+
 48%|████▊     | 2776/5758 [15:46<17:04,  2.91it/s]
 48%|████▊     | 2777/5758 [15:46<16:37,  2.99it/s]
                                                   

 48%|████▊     | 2777/5758 [15:46<16:37,  2.99it/s]
 48%|████▊     | 2778/5758 [15:46<16:51,  2.95it/s]
                                                   

 48%|████▊     | 2778/5758 [15:46<16:51,  2.95it/s]
 48%|████▊     | 2779/5758 [15:47<16:09,  3.07it/s]
                                                   

 48%|████▊     | 2779/5758 [15:47<16:09,  3.07it/s]
 48%|████▊     | 2780/5758 [15:47<16:51,  2.94it/s]
                                                   

 48%|████▊     | 2780/5758 [15:47<16:51,  2.94it/s]
 48%|████▊     | 2781/5758 [15:47<16:34,  2.99it/s]
                                                   

 48%|████▊     | 2781/5758 [15:47<16:34,  2.99it/s]
 48%|████▊     | 2782/5758 [15:48<16:07,  3.08it/s]
                                                   

 48%|████▊     | 2782/5758 [15:48<16:07,  3.08it/s]
 48%|████▊     | 2783/5758 [15:48<17:05,  2.90it/s]
                                                   

 48%|████▊     | 2783/5758 [15:48<17:05,  2.90it/s]
 48%|████▊     | 2784/5758 [15:48<16:40,  2.97it/s]
                                                   

 48%|████▊     | 2784/5758 [15:48<16:40,  2.97it/s]
 48%|████▊     | 2785/5758 [15:49<16:01,  3.09it/s]
                                                   

 48%|████▊     | 2785/5758 [15:49<16:01,  3.09it/s]
 48%|████▊     | 2786/5758 [15:49<16:41,  2.97it/s]
                                                   

 48%|████▊     | 2786/5758 [15:49<16:41,  2.97it/s]
 48%|████▊     | 2787/5758 [15:50<18:31,  2.67it/s]
                                                   

 48%|████▊     | 2787/5758 [15:50<18:31,  2.67it/s]
 48%|████▊     | 2788/5758 [15:50<17:26,  2.84it/s]
                                                   

 48%|████▊     | 2788/5758 [15:50<17:26,  2.84it/s]
 48%|████▊     | 2789/5758 [15:50<17:55,  2.76it/s]
                                                   

 48%|████▊     | 2789/5758 [15:50<17:55,  2.76it/s]
 48%|████▊     | 2790/5758 [15:51<17:16,  2.86it/s]
                                                   

 48%|████▊     | 2790/5758 [15:51<17:16,  2.86it/s]
 48%|████▊     | 2791/5758 [15:51<17:16,  2.86it/s]
                                                   

 48%|████▊     | 2791/5758 [15:51<17:16,  2.86it/s]
 48%|████▊     | 2792/5758 [15:51<16:41,  2.96it/s]
                                                   

 48%|████▊     | 2792/5758 [15:51<16:41,  2.96it/s]
 49%|████▊     | 2793/5758 [15:52<16:14,  3.04it/s]
                                                   

 49%|████▊     | 2793/5758 [15:52<16:14,  3.04it/s]
 49%|████▊     | 2794/5758 [15:52<17:01,  2.90it/s]
                                                   

 49%|████▊     | 2794/5758 [15:52<17:01,  2.90it/s]
 49%|████▊     | 2795/5758 [15:52<16:34,  2.98it/s]
                                                   

 49%|████▊     | 2795/5758 [15:52<16:34,  2.98it/s]
 49%|████▊     | 2796/5758 [15:53<15:54,  3.10it/s]
                                                   

 49%|████▊     | 2796/5758 [15:53<15:54,  3.10it/s]
 49%|████▊     | 2797/5758 [15:53<16:38,  2.97it/s]
                                                   

 49%|████▊     | 2797/5758 [15:53<16:38,  2.97it/s]
 49%|████▊     | 2798/5758 [15:53<16:53,  2.92it/s]
                                                   

 49%|████▊     | 2798/5758 [15:53<16:53,  2.92it/s]
 49%|████▊     | 2799/5758 [15:54<16:53,  2.92it/s]
                                                   

 49%|████▊     | 2799/5758 [15:54<16:53,  2.92it/s]
 49%|████▊     | 2800/5758 [15:54<17:09,  2.87it/s]
                                                   

 49%|████▊     | 2800/5758 [15:54<17:09,  2.87it/s]
 49%|████▊     | 2801/5758 [15:54<16:30,  2.99it/s]
                                                   

 49%|████▊     | 2801/5758 [15:54<16:30,  2.99it/s]
 49%|████▊     | 2802/5758 [15:55<15:47,  3.12it/s]
                                                   

 49%|████▊     | 2802/5758 [15:55<15:47,  3.12it/s]
 49%|████▊     | 2803/5758 [15:55<15:51,  3.11it/s]
                                                   

 49%|████▊     | 2803/5758 [15:55<15:51,  3.11it/s]
 49%|████▊     | 2804/5758 [15:55<15:23,  3.20it/s]
                                                   

 49%|████▊     | 2804/5758 [15:55<15:23,  3.20it/s]
 49%|████▊     | 2805/5758 [15:56<16:02,  3.07it/s]
                                                   

 49%|████▊     | 2805/5758 [15:56<16:02,  3.07it/s]
 49%|████▊     | 2806/5758 [15:56<16:09,  3.04it/s]
                                                   

 49%|████▊     | 2806/5758 [15:56<16:09,  3.04it/s]
 49%|████▊     | 2807/5758 [15:56<15:56,  3.09it/s]
                                                   

 49%|████▊     | 2807/5758 [15:56<15:56,  3.09it/s]
 49%|████▉     | 2808/5758 [15:57<16:52,  2.91it/s]
                                                   

 49%|████▉     | 2808/5758 [15:57<16:52,  2.91it/s]
 49%|████▉     | 2809/5758 [15:57<16:51,  2.91it/s]
                                                   

 49%|████▉     | 2809/5758 [15:57<16:51,  2.91it/s]
 49%|████▉     | 2810/5758 [15:57<17:50,  2.75it/s]
                                                   

 49%|████▉     | 2810/5758 [15:57<17:50,  2.75it/s]
 49%|████▉     | 2811/5758 [15:58<17:16,  2.84it/s]
                                                   

 49%|████▉     | 2811/5758 [15:58<17:16,  2.84it/s]
 49%|████▉     | 2812/5758 [15:58<17:36,  2.79it/s]
                                                   

 49%|████▉     | 2812/5758 [15:58<17:36,  2.79it/s]
 49%|████▉     | 2813/5758 [15:58<17:04,  2.87it/s]
                                                   

 49%|████▉     | 2813/5758 [15:58<17:04,  2.87it/s]
 49%|████▉     | 2814/5758 [15:59<17:10,  2.86it/s]
                                                   

 49%|████▉     | 2814/5758 [15:59<17:10,  2.86it/s]
 49%|████▉     | 2815/5758 [15:59<19:22,  2.53it/s]
                                                   

 49%|████▉     | 2815/5758 [15:59<19:22,  2.53it/s]
 49%|████▉     | 2816/5758 [16:00<18:24,  2.66it/s]
                                                   

 49%|████▉     | 2816/5758 [16:00<18:24,  2.66it/s]
 49%|████▉     | 2817/5758 [16:00<17:18,  2.83it/s]
                                                   

 49%|████▉     | 2817/5758 [16:00<17:18,  2.83it/s]
 49%|████▉     | 2818/5758 [16:00<16:33,  2.96it/s]
                                                   

 49%|████▉     | 2818/5758 [16:00<16:33,  2.96it/s]
 49%|████▉     | 2819/5758 [16:00<15:52,  3.08it/s]
                                                   

 49%|████▉     | 2819/5758 [16:00<15:52,  3.08it/s]
 49%|████▉     | 2820/5758 [16:01<16:40,  2.94it/s]
                                                   

 49%|████▉     | 2820/5758 [16:01<16:40,  2.94it/s]
 49%|████▉     | 2821/5758 [16:01<16:43,  2.93it/s]
                                                   

 49%|████▉     | 2821/5758 [16:01<16:43,  2.93it/s]
 49%|████▉     | 2822/5758 [16:01<16:14,  3.01it/s]
                                                   

 49%|████▉     | 2822/5758 [16:01<16:14,  3.01it/s]
 49%|████▉     | 2823/5758 [16:02<17:18,  2.83it/s]
                                                   

 49%|████▉     | 2823/5758 [16:02<17:18,  2.83it/s]
 49%|████▉     | 2824/5758 [16:02<16:57,  2.88it/s]
                                                   

 49%|████▉     | 2824/5758 [16:02<16:57,  2.88it/s]
 49%|████▉     | 2825/5758 [16:03<16:18,  3.00it/s]
                                                   

 49%|████▉     | 2825/5758 [16:03<16:18,  3.00it/s]
 49%|████▉     | 2826/5758 [16:03<15:37,  3.13it/s]
                                                   

 49%|████▉     | 2826/5758 [16:03<15:37,  3.13it/s]
 49%|████▉     | 2827/5758 [16:03<16:03,  3.04it/s]
                                                   

 49%|████▉     | 2827/5758 [16:03<16:03,  3.04it/s]
 49%|████▉     | 2828/5758 [16:04<16:38,  2.94it/s]
                                                   

 49%|████▉     | 2828/5758 [16:04<16:38,  2.94it/s]
 49%|████▉     | 2829/5758 [16:04<16:35,  2.94it/s]
                                                   

 49%|████▉     | 2829/5758 [16:04<16:35,  2.94it/s]
 49%|████▉     | 2830/5758 [16:04<16:27,  2.97it/s]
                                                   

 49%|████▉     | 2830/5758 [16:04<16:27,  2.97it/s]
 49%|████▉     | 2831/5758 [16:04<15:58,  3.05it/s]
                                                   
{'loss': '1.454', 'grad_norm': '10.38', 'learning_rate': '5.181e-06', 'num_tokens': '1.847e+06', 'mean_token_accuracy': '0.5985', 'epoch': '0.4821'}
+{'loss': '0.534', 'grad_norm': '6.781', 'learning_rate': '5.179e-06', 'num_tokens': '1.849e+06', 'mean_token_accuracy': '0.841', 'epoch': '0.4823'}
+{'loss': '1.689', 'grad_norm': '9.562', 'learning_rate': '5.177e-06', 'num_tokens': '1.85e+06', 'mean_token_accuracy': '0.6432', 'epoch': '0.4825'}
+{'loss': '0.6389', 'grad_norm': '34', 'learning_rate': '5.175e-06', 'num_tokens': '1.85e+06', 'mean_token_accuracy': '0.8382', 'epoch': '0.4826'}
+{'loss': '1.237', 'grad_norm': '18.5', 'learning_rate': '5.174e-06', 'num_tokens': '1.85e+06', 'mean_token_accuracy': '0.7313', 'epoch': '0.4828'}
+{'loss': '1.629', 'grad_norm': '16.88', 'learning_rate': '5.172e-06', 'num_tokens': '1.85e+06', 'mean_token_accuracy': '0.6469', 'epoch': '0.483'}
+{'loss': '1.759', 'grad_norm': '15.25', 'learning_rate': '5.17e-06', 'num_tokens': '1.851e+06', 'mean_token_accuracy': '0.5484', 'epoch': '0.4832'}
+{'loss': '1.894', 'grad_norm': '16.38', 'learning_rate': '5.168e-06', 'num_tokens': '1.852e+06', 'mean_token_accuracy': '0.6024', 'epoch': '0.4833'}
+{'loss': '2.213', 'grad_norm': '12.12', 'learning_rate': '5.167e-06', 'num_tokens': '1.852e+06', 'mean_token_accuracy': '0.5671', 'epoch': '0.4835'}
+{'loss': '0.5551', 'grad_norm': '7.406', 'learning_rate': '5.165e-06', 'num_tokens': '1.854e+06', 'mean_token_accuracy': '0.8333', 'epoch': '0.4837'}
+{'loss': '1.791', 'grad_norm': '14.56', 'learning_rate': '5.163e-06', 'num_tokens': '1.854e+06', 'mean_token_accuracy': '0.6355', 'epoch': '0.4838'}
+{'loss': '1.148', 'grad_norm': '11.31', 'learning_rate': '5.162e-06', 'num_tokens': '1.855e+06', 'mean_token_accuracy': '0.6671', 'epoch': '0.484'}
+{'loss': '0.7914', 'grad_norm': '8.938', 'learning_rate': '5.16e-06', 'num_tokens': '1.856e+06', 'mean_token_accuracy': '0.7544', 'epoch': '0.4842'}
+{'loss': '1.784', 'grad_norm': '17.88', 'learning_rate': '5.158e-06', 'num_tokens': '1.857e+06', 'mean_token_accuracy': '0.6572', 'epoch': '0.4844'}
+{'loss': '0.976', 'grad_norm': '7.125', 'learning_rate': '5.156e-06', 'num_tokens': '1.858e+06', 'mean_token_accuracy': '0.7548', 'epoch': '0.4845'}
+{'loss': '0.8408', 'grad_norm': '17.75', 'learning_rate': '5.155e-06', 'num_tokens': '1.858e+06', 'mean_token_accuracy': '0.8494', 'epoch': '0.4847'}
+{'loss': '1.432', 'grad_norm': '13.5', 'learning_rate': '5.153e-06', 'num_tokens': '1.859e+06', 'mean_token_accuracy': '0.642', 'epoch': '0.4849'}
+{'loss': '1.306', 'grad_norm': '24.75', 'learning_rate': '5.151e-06', 'num_tokens': '1.859e+06', 'mean_token_accuracy': '0.7386', 'epoch': '0.4851'}
+{'loss': '0.6465', 'grad_norm': '14.06', 'learning_rate': '5.149e-06', 'num_tokens': '1.86e+06', 'mean_token_accuracy': '0.8455', 'epoch': '0.4852'}
+{'loss': '1.852', 'grad_norm': '28', 'learning_rate': '5.148e-06', 'num_tokens': '1.86e+06', 'mean_token_accuracy': '0.663', 'epoch': '0.4854'}
+{'loss': '1.259', 'grad_norm': '13.38', 'learning_rate': '5.146e-06', 'num_tokens': '1.861e+06', 'mean_token_accuracy': '0.6611', 'epoch': '0.4856'}
+{'loss': '2.047', 'grad_norm': '18.62', 'learning_rate': '5.144e-06', 'num_tokens': '1.861e+06', 'mean_token_accuracy': '0.603', 'epoch': '0.4858'}
+{'loss': '1.675', 'grad_norm': '15.56', 'learning_rate': '5.142e-06', 'num_tokens': '1.861e+06', 'mean_token_accuracy': '0.6424', 'epoch': '0.4859'}
+{'loss': '1.358', 'grad_norm': '15.06', 'learning_rate': '5.141e-06', 'num_tokens': '1.862e+06', 'mean_token_accuracy': '0.7165', 'epoch': '0.4861'}
+{'loss': '1.943', 'grad_norm': '13.75', 'learning_rate': '5.139e-06', 'num_tokens': '1.862e+06', 'mean_token_accuracy': '0.5825', 'epoch': '0.4863'}
+{'loss': '1.138', 'grad_norm': '12.88', 'learning_rate': '5.137e-06', 'num_tokens': '1.863e+06', 'mean_token_accuracy': '0.6716', 'epoch': '0.4865'}
+{'loss': '0.6802', 'grad_norm': '33.75', 'learning_rate': '5.135e-06', 'num_tokens': '1.863e+06', 'mean_token_accuracy': '0.7941', 'epoch': '0.4866'}
+{'loss': '1.213', 'grad_norm': '12.44', 'learning_rate': '5.134e-06', 'num_tokens': '1.864e+06', 'mean_token_accuracy': '0.6634', 'epoch': '0.4868'}
+{'loss': '1.767', 'grad_norm': '15', 'learning_rate': '5.132e-06', 'num_tokens': '1.864e+06', 'mean_token_accuracy': '0.5795', 'epoch': '0.487'}
+{'loss': '2.282', 'grad_norm': '19.38', 'learning_rate': '5.13e-06', 'num_tokens': '1.865e+06', 'mean_token_accuracy': '0.5173', 'epoch': '0.4871'}
+{'loss': '0.5285', 'grad_norm': '22.88', 'learning_rate': '5.129e-06', 'num_tokens': '1.865e+06', 'mean_token_accuracy': '0.8698', 'epoch': '0.4873'}
+{'loss': '0.7045', 'grad_norm': '20.25', 'learning_rate': '5.127e-06', 'num_tokens': '1.865e+06', 'mean_token_accuracy': '0.8333', 'epoch': '0.4875'}
+{'loss': '1.903', 'grad_norm': '17', 'learning_rate': '5.125e-06', 'num_tokens': '1.866e+06', 'mean_token_accuracy': '0.5852', 'epoch': '0.4877'}
+{'loss': '1.6', 'grad_norm': '23.88', 'learning_rate': '5.123e-06', 'num_tokens': '1.866e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.4878'}
+{'loss': '0.3755', 'grad_norm': '19', 'learning_rate': '5.122e-06', 'num_tokens': '1.866e+06', 'mean_token_accuracy': '0.8973', 'epoch': '0.488'}
+{'loss': '2.18', 'grad_norm': '12.62', 'learning_rate': '5.12e-06', 'num_tokens': '1.867e+06', 'mean_token_accuracy': '0.5731', 'epoch': '0.4882'}
+{'loss': '1.955', 'grad_norm': '13.19', 'learning_rate': '5.118e-06', 'num_tokens': '1.868e+06', 'mean_token_accuracy': '0.593', 'epoch': '0.4884'}
+{'loss': '1.203', 'grad_norm': '11.62', 'learning_rate': '5.116e-06', 'num_tokens': '1.868e+06', 'mean_token_accuracy': '0.6992', 'epoch': '0.4885'}
+{'loss': '1.07', 'grad_norm': '14.69', 'learning_rate': '5.115e-06', 'num_tokens': '1.869e+06', 'mean_token_accuracy': '0.6982', 'epoch': '0.4887'}
+{'loss': '1.514', 'grad_norm': '30.25', 'learning_rate': '5.113e-06', 'num_tokens': '1.869e+06', 'mean_token_accuracy': '0.6911', 'epoch': '0.4889'}
+{'loss': '1.451', 'grad_norm': '11.88', 'learning_rate': '5.111e-06', 'num_tokens': '1.87e+06', 'mean_token_accuracy': '0.6218', 'epoch': '0.4891'}
+{'loss': '1.089', 'grad_norm': '12.44', 'learning_rate': '5.109e-06', 'num_tokens': '1.871e+06', 'mean_token_accuracy': '0.6988', 'epoch': '0.4892'}
+{'loss': '1.231', 'grad_norm': '6.219', 'learning_rate': '5.108e-06', 'num_tokens': '1.873e+06', 'mean_token_accuracy': '0.6896', 'epoch': '0.4894'}
+{'loss': '1.891', 'grad_norm': '13.69', 'learning_rate': '5.106e-06', 'num_tokens': '1.874e+06', 'mean_token_accuracy': '0.5921', 'epoch': '0.4896'}
+{'loss': '1.185', 'grad_norm': '10', 'learning_rate': '5.104e-06', 'num_tokens': '1.875e+06', 'mean_token_accuracy': '0.6933', 'epoch': '0.4898'}
+{'loss': '1.687', 'grad_norm': '10.38', 'learning_rate': '5.102e-06', 'num_tokens': '1.876e+06', 'mean_token_accuracy': '0.6337', 'epoch': '0.4899'}
+{'loss': '0.8229', 'grad_norm': '19.62', 'learning_rate': '5.101e-06', 'num_tokens': '1.876e+06', 'mean_token_accuracy': '0.7705', 'epoch': '0.4901'}
+{'loss': '2.409', 'grad_norm': '33.75', 'learning_rate': '5.099e-06', 'num_tokens': '1.876e+06', 'mean_token_accuracy': '0.5931', 'epoch': '0.4903'}
+{'loss': '1.261', 'grad_norm': '7.344', 'learning_rate': '5.097e-06', 'num_tokens': '1.878e+06', 'mean_token_accuracy': '0.7217', 'epoch': '0.4904'}
+{'loss': '0.7108', 'grad_norm': '31.88', 'learning_rate': '5.096e-06', 'num_tokens': '1.878e+06', 'mean_token_accuracy': '0.8312', 'epoch': '0.4906'}
+{'loss': '0.5844', 'grad_norm': '8.062', 'learning_rate': '5.094e-06', 'num_tokens': '1.879e+06', 'mean_token_accuracy': '0.8198', 'epoch': '0.4908'}
+{'loss': '2.04', 'grad_norm': '9.438', 'learning_rate': '5.092e-06', 'num_tokens': '1.88e+06', 'mean_token_accuracy': '0.5518', 'epoch': '0.491'}
+{'loss': '1.274', 'grad_norm': '15.56', 'learning_rate': '5.09e-06', 'num_tokens': '1.881e+06', 'mean_token_accuracy': '0.6698', 'epoch': '0.4911'}
+{'loss': '1.123', 'grad_norm': '11.31', 'learning_rate': '5.089e-06', 'num_tokens': '1.882e+06', 'mean_token_accuracy': '0.6852', 'epoch': '0.4913'}
+{'loss': '1.445', 'grad_norm': '11.56', 'learning_rate': '5.087e-06', 'num_tokens': '1.882e+06', 'mean_token_accuracy': '0.6128', 'epoch': '0.4915'}
+
 49%|████▉     | 2831/5758 [16:04<15:58,  3.05it/s]
 49%|████▉     | 2832/5758 [16:05<16:10,  3.02it/s]
                                                   

 49%|████▉     | 2832/5758 [16:05<16:10,  3.02it/s]
 49%|████▉     | 2833/5758 [16:05<16:48,  2.90it/s]
                                                   

 49%|████▉     | 2833/5758 [16:05<16:48,  2.90it/s]
 49%|████▉     | 2834/5758 [16:06<16:33,  2.94it/s]
                                                   

 49%|████▉     | 2834/5758 [16:06<16:33,  2.94it/s]
 49%|████▉     | 2835/5758 [16:06<16:34,  2.94it/s]
                                                   

 49%|████▉     | 2835/5758 [16:06<16:34,  2.94it/s]
 49%|████▉     | 2836/5758 [16:06<17:09,  2.84it/s]
                                                   

 49%|████▉     | 2836/5758 [16:06<17:09,  2.84it/s]
 49%|████▉     | 2837/5758 [16:07<16:49,  2.89it/s]
                                                   

 49%|████▉     | 2837/5758 [16:07<16:49,  2.89it/s]
 49%|████▉     | 2838/5758 [16:07<16:18,  2.98it/s]
                                                   

 49%|████▉     | 2838/5758 [16:07<16:18,  2.98it/s]
 49%|████▉     | 2839/5758 [16:07<17:01,  2.86it/s]
                                                   

 49%|████▉     | 2839/5758 [16:07<17:01,  2.86it/s]
 49%|████▉     | 2840/5758 [16:08<18:46,  2.59it/s]
                                                   

 49%|████▉     | 2840/5758 [16:08<18:46,  2.59it/s]
 49%|████▉     | 2841/5758 [16:08<18:23,  2.64it/s]
                                                   

 49%|████▉     | 2841/5758 [16:08<18:23,  2.64it/s]
 49%|████▉     | 2842/5758 [16:08<18:05,  2.69it/s]
                                                   

 49%|████▉     | 2842/5758 [16:08<18:05,  2.69it/s]
 49%|████▉     | 2843/5758 [16:09<18:15,  2.66it/s]
                                                   

 49%|████▉     | 2843/5758 [16:09<18:15,  2.66it/s]
 49%|████▉     | 2844/5758 [16:09<17:00,  2.86it/s]
                                                   

 49%|████▉     | 2844/5758 [16:09<17:00,  2.86it/s]
 49%|████▉     | 2845/5758 [16:09<16:33,  2.93it/s]
                                                   

 49%|████▉     | 2845/5758 [16:09<16:33,  2.93it/s]
 49%|████▉     | 2846/5758 [16:10<16:36,  2.92it/s]
                                                   

 49%|████▉     | 2846/5758 [16:10<16:36,  2.92it/s]
 49%|████▉     | 2847/5758 [16:10<16:31,  2.94it/s]
                                                   

 49%|████▉     | 2847/5758 [16:10<16:31,  2.94it/s]
 49%|████▉     | 2848/5758 [16:10<16:16,  2.98it/s]
                                                   

 49%|████▉     | 2848/5758 [16:10<16:16,  2.98it/s]
 49%|████▉     | 2849/5758 [16:11<16:20,  2.97it/s]
                                                   

 49%|████▉     | 2849/5758 [16:11<16:20,  2.97it/s]
 49%|████▉     | 2850/5758 [16:11<16:43,  2.90it/s]
                                                   

 49%|████▉     | 2850/5758 [16:11<16:43,  2.90it/s]
 50%|████▉     | 2851/5758 [16:12<17:10,  2.82it/s]
                                                   

 50%|████▉     | 2851/5758 [16:12<17:10,  2.82it/s]
 50%|████▉     | 2852/5758 [16:12<16:46,  2.89it/s]
                                                   

 50%|████▉     | 2852/5758 [16:12<16:46,  2.89it/s]
 50%|████▉     | 2853/5758 [16:12<16:05,  3.01it/s]
                                                   

 50%|████▉     | 2853/5758 [16:12<16:05,  3.01it/s]
 50%|████▉     | 2854/5758 [16:12<15:45,  3.07it/s]
                                                   

 50%|████▉     | 2854/5758 [16:12<15:45,  3.07it/s]
 50%|████▉     | 2855/5758 [16:13<16:26,  2.94it/s]
                                                   

 50%|████▉     | 2855/5758 [16:13<16:26,  2.94it/s]
 50%|████▉     | 2856/5758 [16:13<15:49,  3.06it/s]
                                                   

 50%|████▉     | 2856/5758 [16:13<15:49,  3.06it/s]
 50%|████▉     | 2857/5758 [16:14<16:29,  2.93it/s]
                                                   

 50%|████▉     | 2857/5758 [16:14<16:29,  2.93it/s]
 50%|████▉     | 2858/5758 [16:14<16:31,  2.93it/s]
                                                   

 50%|████▉     | 2858/5758 [16:14<16:31,  2.93it/s]
 50%|████▉     | 2859/5758 [16:14<15:52,  3.04it/s]
                                                   

 50%|████▉     | 2859/5758 [16:14<15:52,  3.04it/s]
 50%|████▉     | 2860/5758 [16:15<16:17,  2.96it/s]
                                                   

 50%|████▉     | 2860/5758 [16:15<16:17,  2.96it/s]
 50%|████▉     | 2861/5758 [16:15<16:17,  2.96it/s]
                                                   

 50%|████▉     | 2861/5758 [16:15<16:17,  2.96it/s]
 50%|████▉     | 2862/5758 [16:15<16:54,  2.85it/s]
                                                   

 50%|████▉     | 2862/5758 [16:15<16:54,  2.85it/s]
 50%|████▉     | 2863/5758 [16:16<16:54,  2.85it/s]
                                                   

 50%|████▉     | 2863/5758 [16:16<16:54,  2.85it/s]
 50%|████▉     | 2864/5758 [16:16<16:04,  3.00it/s]
                                                   

 50%|████▉     | 2864/5758 [16:16<16:04,  3.00it/s]
 50%|████▉     | 2865/5758 [16:16<15:39,  3.08it/s]
                                                   

 50%|████▉     | 2865/5758 [16:16<15:39,  3.08it/s]
 50%|████▉     | 2866/5758 [16:17<17:20,  2.78it/s]
                                                   

 50%|████▉     | 2866/5758 [16:17<17:20,  2.78it/s]
 50%|████▉     | 2867/5758 [16:17<17:16,  2.79it/s]
                                                   

 50%|████▉     | 2867/5758 [16:17<17:16,  2.79it/s]
 50%|████▉     | 2868/5758 [16:17<16:23,  2.94it/s]
                                                   

 50%|████▉     | 2868/5758 [16:17<16:23,  2.94it/s]
 50%|████▉     | 2869/5758 [16:18<15:55,  3.02it/s]
                                                   

 50%|████▉     | 2869/5758 [16:18<15:55,  3.02it/s]
 50%|████▉     | 2870/5758 [16:18<15:45,  3.05it/s]
                                                   

 50%|████▉     | 2870/5758 [16:18<15:45,  3.05it/s]
 50%|████▉     | 2871/5758 [16:18<15:26,  3.12it/s]
                                                   

 50%|████▉     | 2871/5758 [16:18<15:26,  3.12it/s]
 50%|████▉     | 2872/5758 [16:19<16:29,  2.92it/s]
                                                   

 50%|████▉     | 2872/5758 [16:19<16:29,  2.92it/s]
 50%|████▉     | 2873/5758 [16:19<17:23,  2.76it/s]
                                                   

 50%|████▉     | 2873/5758 [16:19<17:23,  2.76it/s]
 50%|████▉     | 2874/5758 [16:19<16:32,  2.91it/s]
                                                   

 50%|████▉     | 2874/5758 [16:19<16:32,  2.91it/s]
 50%|████▉     | 2875/5758 [16:20<16:13,  2.96it/s]
                                                   

 50%|████▉     | 2875/5758 [16:20<16:13,  2.96it/s]
 50%|████▉     | 2876/5758 [16:20<16:18,  2.95it/s]
                                                   

 50%|████▉     | 2876/5758 [16:20<16:18,  2.95it/s]
 50%|████▉     | 2877/5758 [16:20<16:20,  2.94it/s]
                                                   

 50%|████▉     | 2877/5758 [16:20<16:20,  2.94it/s]
 50%|████▉     | 2878/5758 [16:21<15:48,  3.04it/s]
                                                   

 50%|████▉     | 2878/5758 [16:21<15:48,  3.04it/s]
 50%|█████     | 2879/5758 [16:21<15:47,  3.04it/s]
                                                   

 50%|█████     | 2879/5758 [16:21<15:47,  3.04it/s]
 50%|█████     | 2880/5758 [16:21<16:08,  2.97it/s]
                                                   

 50%|█████     | 2880/5758 [16:21<16:08,  2.97it/s]
 50%|█████     | 2881/5758 [16:22<15:53,  3.02it/s]
                                                   

 50%|█████     | 2881/5758 [16:22<15:53,  3.02it/s]
 50%|█████     | 2882/5758 [16:22<15:38,  3.07it/s]
                                                   

 50%|█████     | 2882/5758 [16:22<15:38,  3.07it/s]
 50%|█████     | 2883/5758 [16:22<15:57,  3.00it/s]
                                                   

 50%|█████     | 2883/5758 [16:22<15:57,  3.00it/s]
 50%|█████     | 2884/5758 [16:23<15:17,  3.13it/s]
                                                   

 50%|█████     | 2884/5758 [16:23<15:17,  3.13it/s]
 50%|█████     | 2885/5758 [16:23<16:05,  2.98it/s]
                                                   

 50%|█████     | 2885/5758 [16:23<16:05,  2.98it/s]
 50%|█████     | 2886/5758 [16:23<16:11,  2.96it/s]
                                                   
{'loss': '1.717', 'grad_norm': '12.06', 'learning_rate': '5.085e-06', 'num_tokens': '1.883e+06', 'mean_token_accuracy': '0.6285', 'epoch': '0.4917'}
+{'loss': '0.7964', 'grad_norm': '12.25', 'learning_rate': '5.083e-06', 'num_tokens': '1.884e+06', 'mean_token_accuracy': '0.8249', 'epoch': '0.4918'}
+{'loss': '1.903', 'grad_norm': '16.88', 'learning_rate': '5.082e-06', 'num_tokens': '1.884e+06', 'mean_token_accuracy': '0.6029', 'epoch': '0.492'}
+{'loss': '1.305', 'grad_norm': '11.56', 'learning_rate': '5.08e-06', 'num_tokens': '1.885e+06', 'mean_token_accuracy': '0.6546', 'epoch': '0.4922'}
+{'loss': '1.024', 'grad_norm': '17', 'learning_rate': '5.078e-06', 'num_tokens': '1.885e+06', 'mean_token_accuracy': '0.7629', 'epoch': '0.4924'}
+{'loss': '1.327', 'grad_norm': '14.5', 'learning_rate': '5.076e-06', 'num_tokens': '1.886e+06', 'mean_token_accuracy': '0.6613', 'epoch': '0.4925'}
+{'loss': '1.054', 'grad_norm': '19.75', 'learning_rate': '5.075e-06', 'num_tokens': '1.886e+06', 'mean_token_accuracy': '0.7692', 'epoch': '0.4927'}
+{'loss': '1.711', 'grad_norm': '17.38', 'learning_rate': '5.073e-06', 'num_tokens': '1.886e+06', 'mean_token_accuracy': '0.6392', 'epoch': '0.4929'}
+{'loss': '1.771', 'grad_norm': '17.12', 'learning_rate': '5.071e-06', 'num_tokens': '1.887e+06', 'mean_token_accuracy': '0.642', 'epoch': '0.4931'}
+{'loss': '2.103', 'grad_norm': '11', 'learning_rate': '5.069e-06', 'num_tokens': '1.888e+06', 'mean_token_accuracy': '0.5799', 'epoch': '0.4932'}
+{'loss': '1.766', 'grad_norm': '14.44', 'learning_rate': '5.068e-06', 'num_tokens': '1.888e+06', 'mean_token_accuracy': '0.6329', 'epoch': '0.4934'}
+{'loss': '1.52', 'grad_norm': '7.531', 'learning_rate': '5.066e-06', 'num_tokens': '1.89e+06', 'mean_token_accuracy': '0.6718', 'epoch': '0.4936'}
+{'loss': '0.8892', 'grad_norm': '32.75', 'learning_rate': '5.064e-06', 'num_tokens': '1.89e+06', 'mean_token_accuracy': '0.8031', 'epoch': '0.4937'}
+{'loss': '1.07', 'grad_norm': '12.44', 'learning_rate': '5.063e-06', 'num_tokens': '1.891e+06', 'mean_token_accuracy': '0.7117', 'epoch': '0.4939'}
+{'loss': '1.931', 'grad_norm': '48', 'learning_rate': '5.061e-06', 'num_tokens': '1.891e+06', 'mean_token_accuracy': '0.5747', 'epoch': '0.4941'}
+{'loss': '2.252', 'grad_norm': '15.25', 'learning_rate': '5.059e-06', 'num_tokens': '1.892e+06', 'mean_token_accuracy': '0.5614', 'epoch': '0.4943'}
+{'loss': '1.78', 'grad_norm': '23.12', 'learning_rate': '5.057e-06', 'num_tokens': '1.892e+06', 'mean_token_accuracy': '0.6498', 'epoch': '0.4944'}
+{'loss': '1.283', 'grad_norm': '11.62', 'learning_rate': '5.056e-06', 'num_tokens': '1.893e+06', 'mean_token_accuracy': '0.7146', 'epoch': '0.4946'}
+{'loss': '1.601', 'grad_norm': '27.25', 'learning_rate': '5.054e-06', 'num_tokens': '1.893e+06', 'mean_token_accuracy': '0.6727', 'epoch': '0.4948'}
+{'loss': '1.012', 'grad_norm': '6.438', 'learning_rate': '5.052e-06', 'num_tokens': '1.895e+06', 'mean_token_accuracy': '0.7543', 'epoch': '0.495'}
+{'loss': '2.06', 'grad_norm': '18', 'learning_rate': '5.05e-06', 'num_tokens': '1.895e+06', 'mean_token_accuracy': '0.5774', 'epoch': '0.4951'}
+{'loss': '1.664', 'grad_norm': '21.88', 'learning_rate': '5.049e-06', 'num_tokens': '1.896e+06', 'mean_token_accuracy': '0.6452', 'epoch': '0.4953'}
+{'loss': '0.6948', 'grad_norm': '42.75', 'learning_rate': '5.047e-06', 'num_tokens': '1.896e+06', 'mean_token_accuracy': '0.8243', 'epoch': '0.4955'}
+{'loss': '0.9939', 'grad_norm': '11.69', 'learning_rate': '5.045e-06', 'num_tokens': '1.896e+06', 'mean_token_accuracy': '0.7191', 'epoch': '0.4957'}
+{'loss': '1.035', 'grad_norm': '13.56', 'learning_rate': '5.043e-06', 'num_tokens': '1.897e+06', 'mean_token_accuracy': '0.7546', 'epoch': '0.4958'}
+{'loss': '1.224', 'grad_norm': '28.75', 'learning_rate': '5.042e-06', 'num_tokens': '1.897e+06', 'mean_token_accuracy': '0.717', 'epoch': '0.496'}
+{'loss': '0.5413', 'grad_norm': '8.312', 'learning_rate': '5.04e-06', 'num_tokens': '1.898e+06', 'mean_token_accuracy': '0.8408', 'epoch': '0.4962'}
+{'loss': '1.66', 'grad_norm': '12.62', 'learning_rate': '5.038e-06', 'num_tokens': '1.899e+06', 'mean_token_accuracy': '0.5792', 'epoch': '0.4964'}
+{'loss': '1.345', 'grad_norm': '31', 'learning_rate': '5.036e-06', 'num_tokens': '1.899e+06', 'mean_token_accuracy': '0.7134', 'epoch': '0.4965'}
+{'loss': '1.119', 'grad_norm': '5.781', 'learning_rate': '5.035e-06', 'num_tokens': '1.903e+06', 'mean_token_accuracy': '0.7116', 'epoch': '0.4967'}
+{'loss': '2.094', 'grad_norm': '9.75', 'learning_rate': '5.033e-06', 'num_tokens': '1.904e+06', 'mean_token_accuracy': '0.5674', 'epoch': '0.4969'}
+{'loss': '1.571', 'grad_norm': '16.88', 'learning_rate': '5.031e-06', 'num_tokens': '1.905e+06', 'mean_token_accuracy': '0.6791', 'epoch': '0.497'}
+{'loss': '1.34', 'grad_norm': '11.06', 'learning_rate': '5.03e-06', 'num_tokens': '1.906e+06', 'mean_token_accuracy': '0.6505', 'epoch': '0.4972'}
+{'loss': '1.274', 'grad_norm': '17.38', 'learning_rate': '5.028e-06', 'num_tokens': '1.906e+06', 'mean_token_accuracy': '0.7247', 'epoch': '0.4974'}
+{'loss': '0.7241', 'grad_norm': '16.62', 'learning_rate': '5.026e-06', 'num_tokens': '1.906e+06', 'mean_token_accuracy': '0.8211', 'epoch': '0.4976'}
+{'loss': '1.367', 'grad_norm': '18.38', 'learning_rate': '5.024e-06', 'num_tokens': '1.907e+06', 'mean_token_accuracy': '0.6947', 'epoch': '0.4977'}
+{'loss': '1.433', 'grad_norm': '26.12', 'learning_rate': '5.023e-06', 'num_tokens': '1.907e+06', 'mean_token_accuracy': '0.7672', 'epoch': '0.4979'}
+{'loss': '2.106', 'grad_norm': '15.62', 'learning_rate': '5.021e-06', 'num_tokens': '1.908e+06', 'mean_token_accuracy': '0.5292', 'epoch': '0.4981'}
+{'loss': '1.562', 'grad_norm': '13.62', 'learning_rate': '5.019e-06', 'num_tokens': '1.908e+06', 'mean_token_accuracy': '0.6006', 'epoch': '0.4983'}
+{'loss': '1.26', 'grad_norm': '11.94', 'learning_rate': '5.017e-06', 'num_tokens': '1.909e+06', 'mean_token_accuracy': '0.6416', 'epoch': '0.4984'}
+{'loss': '1.168', 'grad_norm': '7.812', 'learning_rate': '5.016e-06', 'num_tokens': '1.91e+06', 'mean_token_accuracy': '0.7258', 'epoch': '0.4986'}
+{'loss': '2.284', 'grad_norm': '18.12', 'learning_rate': '5.014e-06', 'num_tokens': '1.911e+06', 'mean_token_accuracy': '0.5397', 'epoch': '0.4988'}
+{'loss': '0.9534', 'grad_norm': '27', 'learning_rate': '5.012e-06', 'num_tokens': '1.911e+06', 'mean_token_accuracy': '0.8014', 'epoch': '0.499'}
+{'loss': '1.213', 'grad_norm': '13.69', 'learning_rate': '5.01e-06', 'num_tokens': '1.912e+06', 'mean_token_accuracy': '0.6748', 'epoch': '0.4991'}
+{'loss': '0.8866', 'grad_norm': '10.75', 'learning_rate': '5.009e-06', 'num_tokens': '1.913e+06', 'mean_token_accuracy': '0.7616', 'epoch': '0.4993'}
+{'loss': '1.189', 'grad_norm': '10.69', 'learning_rate': '5.007e-06', 'num_tokens': '1.914e+06', 'mean_token_accuracy': '0.6975', 'epoch': '0.4995'}
+{'loss': '1.684', 'grad_norm': '22', 'learning_rate': '5.005e-06', 'num_tokens': '1.914e+06', 'mean_token_accuracy': '0.6712', 'epoch': '0.4997'}
+{'loss': '1.007', 'grad_norm': '12', 'learning_rate': '5.003e-06', 'num_tokens': '1.914e+06', 'mean_token_accuracy': '0.7254', 'epoch': '0.4998'}
+{'loss': '0.7271', 'grad_norm': '7.344', 'learning_rate': '5.002e-06', 'num_tokens': '1.916e+06', 'mean_token_accuracy': '0.78', 'epoch': '0.5'}
+{'loss': '1.547', 'grad_norm': '18.75', 'learning_rate': '5e-06', 'num_tokens': '1.917e+06', 'mean_token_accuracy': '0.687', 'epoch': '0.5002'}
+{'loss': '1.099', 'grad_norm': '11.56', 'learning_rate': '4.998e-06', 'num_tokens': '1.917e+06', 'mean_token_accuracy': '0.6976', 'epoch': '0.5003'}
+{'loss': '1.128', 'grad_norm': '12', 'learning_rate': '4.997e-06', 'num_tokens': '1.918e+06', 'mean_token_accuracy': '0.6947', 'epoch': '0.5005'}
+{'loss': '2.158', 'grad_norm': '10.75', 'learning_rate': '4.995e-06', 'num_tokens': '1.919e+06', 'mean_token_accuracy': '0.5231', 'epoch': '0.5007'}
+{'loss': '0.8141', 'grad_norm': '12.38', 'learning_rate': '4.993e-06', 'num_tokens': '1.92e+06', 'mean_token_accuracy': '0.7579', 'epoch': '0.5009'}
+{'loss': '0.949', 'grad_norm': '6.25', 'learning_rate': '4.991e-06', 'num_tokens': '1.922e+06', 'mean_token_accuracy': '0.759', 'epoch': '0.501'}
+
 50%|█████     | 2886/5758 [16:23<16:11,  2.96it/s]
 50%|█████     | 2887/5758 [16:24<15:47,  3.03it/s]
                                                   

 50%|█████     | 2887/5758 [16:24<15:47,  3.03it/s]
 50%|█████     | 2888/5758 [16:24<15:22,  3.11it/s]
                                                   

 50%|█████     | 2888/5758 [16:24<15:22,  3.11it/s]
 50%|█████     | 2889/5758 [16:24<14:56,  3.20it/s]
                                                   

 50%|█████     | 2889/5758 [16:24<14:56,  3.20it/s]
 50%|█████     | 2890/5758 [16:25<15:15,  3.13it/s]
                                                   

 50%|█████     | 2890/5758 [16:25<15:15,  3.13it/s]
 50%|█████     | 2891/5758 [16:25<14:51,  3.21it/s]
                                                   

 50%|█████     | 2891/5758 [16:25<14:51,  3.21it/s]
 50%|█████     | 2892/5758 [16:25<17:40,  2.70it/s]
                                                   

 50%|█████     | 2892/5758 [16:25<17:40,  2.70it/s]
 50%|█████     | 2893/5758 [16:26<17:07,  2.79it/s]
                                                   

 50%|█████     | 2893/5758 [16:26<17:07,  2.79it/s]
 50%|█████     | 2894/5758 [16:26<16:32,  2.89it/s]
                                                   

 50%|█████     | 2894/5758 [16:26<16:32,  2.89it/s]
 50%|█████     | 2895/5758 [16:26<16:03,  2.97it/s]
                                                   

 50%|█████     | 2895/5758 [16:26<16:03,  2.97it/s]
 50%|█████     | 2896/5758 [16:27<16:40,  2.86it/s]
                                                   

 50%|█████     | 2896/5758 [16:27<16:40,  2.86it/s]
 50%|█████     | 2897/5758 [16:27<17:08,  2.78it/s]
                                                   

 50%|█████     | 2897/5758 [16:27<17:08,  2.78it/s]
 50%|█████     | 2898/5758 [16:27<16:14,  2.94it/s]
                                                   

 50%|█████     | 2898/5758 [16:27<16:14,  2.94it/s]
 50%|█████     | 2899/5758 [16:28<16:25,  2.90it/s]
                                                   

 50%|█████     | 2899/5758 [16:28<16:25,  2.90it/s]
 50%|█████     | 2900/5758 [16:28<15:43,  3.03it/s]
                                                   

 50%|█████     | 2900/5758 [16:28<15:43,  3.03it/s]
 50%|█████     | 2901/5758 [16:28<15:45,  3.02it/s]
                                                   

 50%|█████     | 2901/5758 [16:28<15:45,  3.02it/s]
 50%|█████     | 2902/5758 [16:29<15:58,  2.98it/s]
                                                   

 50%|█████     | 2902/5758 [16:29<15:58,  2.98it/s]
 50%|█████     | 2903/5758 [16:29<16:31,  2.88it/s]
                                                   

 50%|█████     | 2903/5758 [16:29<16:31,  2.88it/s]
 50%|█████     | 2904/5758 [16:29<17:04,  2.79it/s]
                                                   

 50%|█████     | 2904/5758 [16:29<17:04,  2.79it/s]
 50%|█████     | 2905/5758 [16:30<16:42,  2.85it/s]
                                                   

 50%|█████     | 2905/5758 [16:30<16:42,  2.85it/s]
 50%|█████     | 2906/5758 [16:30<16:31,  2.88it/s]
                                                   

 50%|█████     | 2906/5758 [16:30<16:31,  2.88it/s]
 50%|█████     | 2907/5758 [16:31<17:10,  2.77it/s]
                                                   

 50%|█████     | 2907/5758 [16:31<17:10,  2.77it/s]
 51%|█████     | 2908/5758 [16:31<16:19,  2.91it/s]
                                                   

 51%|█████     | 2908/5758 [16:31<16:19,  2.91it/s]
 51%|█████     | 2909/5758 [16:31<16:29,  2.88it/s]
                                                   

 51%|█████     | 2909/5758 [16:31<16:29,  2.88it/s]
 51%|█████     | 2910/5758 [16:32<16:40,  2.85it/s]
                                                   

 51%|█████     | 2910/5758 [16:32<16:40,  2.85it/s]
 51%|█████     | 2911/5758 [16:32<16:13,  2.92it/s]
                                                   

 51%|█████     | 2911/5758 [16:32<16:13,  2.92it/s]
 51%|█████     | 2912/5758 [16:32<15:59,  2.97it/s]
                                                   

 51%|█████     | 2912/5758 [16:32<15:59,  2.97it/s]
 51%|█████     | 2913/5758 [16:33<15:40,  3.02it/s]
                                                   

 51%|█████     | 2913/5758 [16:33<15:40,  3.02it/s]
 51%|█████     | 2914/5758 [16:33<15:20,  3.09it/s]
                                                   

 51%|█████     | 2914/5758 [16:33<15:20,  3.09it/s]
 51%|█████     | 2915/5758 [16:33<16:03,  2.95it/s]
                                                   

 51%|█████     | 2915/5758 [16:33<16:03,  2.95it/s]
 51%|█████     | 2916/5758 [16:34<16:30,  2.87it/s]
                                                   

 51%|█████     | 2916/5758 [16:34<16:30,  2.87it/s]
 51%|█████     | 2917/5758 [16:34<16:05,  2.94it/s]
                                                   

 51%|█████     | 2917/5758 [16:34<16:05,  2.94it/s]
 51%|█████     | 2918/5758 [16:34<18:07,  2.61it/s]
                                                   

 51%|█████     | 2918/5758 [16:34<18:07,  2.61it/s]
 51%|█████     | 2919/5758 [16:35<17:53,  2.65it/s]
                                                   

 51%|█████     | 2919/5758 [16:35<17:53,  2.65it/s]
 51%|█████     | 2920/5758 [16:35<17:46,  2.66it/s]
                                                   

 51%|█████     | 2920/5758 [16:35<17:46,  2.66it/s]
 51%|█████     | 2921/5758 [16:35<16:45,  2.82it/s]
                                                   

 51%|█████     | 2921/5758 [16:35<16:45,  2.82it/s]
 51%|█████     | 2922/5758 [16:36<16:11,  2.92it/s]
                                                   

 51%|█████     | 2922/5758 [16:36<16:11,  2.92it/s]
 51%|█████     | 2923/5758 [16:36<16:47,  2.81it/s]
                                                   

 51%|█████     | 2923/5758 [16:36<16:47,  2.81it/s]
 51%|█████     | 2924/5758 [16:36<15:59,  2.95it/s]
                                                   

 51%|█████     | 2924/5758 [16:36<15:59,  2.95it/s]
 51%|█████     | 2925/5758 [16:37<16:11,  2.92it/s]
                                                   

 51%|█████     | 2925/5758 [16:37<16:11,  2.92it/s]
 51%|█████     | 2926/5758 [16:37<16:10,  2.92it/s]
                                                   

 51%|█████     | 2926/5758 [16:37<16:10,  2.92it/s]
 51%|█████     | 2927/5758 [16:37<15:41,  3.01it/s]
                                                   

 51%|█████     | 2927/5758 [16:37<15:41,  3.01it/s]
 51%|█████     | 2928/5758 [16:38<15:32,  3.03it/s]
                                                   

 51%|█████     | 2928/5758 [16:38<15:32,  3.03it/s]
 51%|█████     | 2929/5758 [16:38<16:00,  2.95it/s]
                                                   

 51%|█████     | 2929/5758 [16:38<16:00,  2.95it/s]
 51%|█████     | 2930/5758 [16:38<15:46,  2.99it/s]
                                                   

 51%|█████     | 2930/5758 [16:38<15:46,  2.99it/s]
 51%|█████     | 2931/5758 [16:39<15:15,  3.09it/s]
                                                   

 51%|█████     | 2931/5758 [16:39<15:15,  3.09it/s]
 51%|█████     | 2932/5758 [16:39<15:07,  3.11it/s]
                                                   

 51%|█████     | 2932/5758 [16:39<15:07,  3.11it/s]
 51%|█████     | 2933/5758 [16:39<14:56,  3.15it/s]
                                                   

 51%|█████     | 2933/5758 [16:39<14:56,  3.15it/s]
 51%|█████     | 2934/5758 [16:40<15:03,  3.13it/s]
                                                   

 51%|█████     | 2934/5758 [16:40<15:03,  3.13it/s]
 51%|█████     | 2935/5758 [16:40<15:12,  3.09it/s]
                                                   

 51%|█████     | 2935/5758 [16:40<15:12,  3.09it/s]
 51%|█████     | 2936/5758 [16:40<15:17,  3.08it/s]
                                                   

 51%|█████     | 2936/5758 [16:40<15:17,  3.08it/s]
 51%|█████     | 2937/5758 [16:41<15:02,  3.13it/s]
                                                   

 51%|███���█     | 2937/5758 [16:41<15:02,  3.13it/s]
 51%|█████     | 2938/5758 [16:41<14:36,  3.22it/s]
                                                   

 51%|█████     | 2938/5758 [16:41<14:36,  3.22it/s]
 51%|█████     | 2939/5758 [16:41<15:01,  3.13it/s]
                                                   

 51%|█████     | 2939/5758 [16:41<15:01,  3.13it/s]
 51%|█████     | 2940/5758 [16:42<15:22,  3.05it/s]
                                                   

 51%|█████     | 2940/5758 [16:42<15:22,  3.05it/s]
 51%|█████     | 2941/5758 [16:42<15:08,  3.10it/s]
                                                   
{'loss': '0.6495', 'grad_norm': '30.5', 'learning_rate': '4.99e-06', 'num_tokens': '1.922e+06', 'mean_token_accuracy': '0.8431', 'epoch': '0.5012'}
+{'loss': '1.985', 'grad_norm': '11.38', 'learning_rate': '4.988e-06', 'num_tokens': '1.922e+06', 'mean_token_accuracy': '0.6113', 'epoch': '0.5014'}
+{'loss': '1.373', 'grad_norm': '13.25', 'learning_rate': '4.986e-06', 'num_tokens': '1.923e+06', 'mean_token_accuracy': '0.6336', 'epoch': '0.5016'}
+{'loss': '0.5088', 'grad_norm': '30.5', 'learning_rate': '4.984e-06', 'num_tokens': '1.923e+06', 'mean_token_accuracy': '0.8507', 'epoch': '0.5017'}
+{'loss': '1.142', 'grad_norm': '14.56', 'learning_rate': '4.983e-06', 'num_tokens': '1.924e+06', 'mean_token_accuracy': '0.7546', 'epoch': '0.5019'}
+{'loss': '1.425', 'grad_norm': '34.5', 'learning_rate': '4.981e-06', 'num_tokens': '1.924e+06', 'mean_token_accuracy': '0.6711', 'epoch': '0.5021'}
+{'loss': '1.643', 'grad_norm': '12.25', 'learning_rate': '4.979e-06', 'num_tokens': '1.924e+06', 'mean_token_accuracy': '0.661', 'epoch': '0.5023'}
+{'loss': '0.7681', 'grad_norm': '29.88', 'learning_rate': '4.977e-06', 'num_tokens': '1.924e+06', 'mean_token_accuracy': '0.7935', 'epoch': '0.5024'}
+{'loss': '1.23', 'grad_norm': '12.62', 'learning_rate': '4.976e-06', 'num_tokens': '1.925e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.5026'}
+{'loss': '0.7362', 'grad_norm': '21.25', 'learning_rate': '4.974e-06', 'num_tokens': '1.925e+06', 'mean_token_accuracy': '0.8171', 'epoch': '0.5028'}
+{'loss': '1.875', 'grad_norm': '14', 'learning_rate': '4.972e-06', 'num_tokens': '1.926e+06', 'mean_token_accuracy': '0.6359', 'epoch': '0.503'}
+{'loss': '1.494', 'grad_norm': '16.75', 'learning_rate': '4.97e-06', 'num_tokens': '1.926e+06', 'mean_token_accuracy': '0.641', 'epoch': '0.5031'}
+{'loss': '1.554', 'grad_norm': '12.94', 'learning_rate': '4.969e-06', 'num_tokens': '1.927e+06', 'mean_token_accuracy': '0.6743', 'epoch': '0.5033'}
+{'loss': '1.339', 'grad_norm': '11.19', 'learning_rate': '4.967e-06', 'num_tokens': '1.928e+06', 'mean_token_accuracy': '0.6308', 'epoch': '0.5035'}
+{'loss': '2.302', 'grad_norm': '17.88', 'learning_rate': '4.965e-06', 'num_tokens': '1.929e+06', 'mean_token_accuracy': '0.5535', 'epoch': '0.5036'}
+{'loss': '1.799', 'grad_norm': '13.94', 'learning_rate': '4.964e-06', 'num_tokens': '1.929e+06', 'mean_token_accuracy': '0.6181', 'epoch': '0.5038'}
+{'loss': '1.235', 'grad_norm': '11.5', 'learning_rate': '4.962e-06', 'num_tokens': '1.93e+06', 'mean_token_accuracy': '0.6681', 'epoch': '0.504'}
+{'loss': '1.785', 'grad_norm': '13.94', 'learning_rate': '4.96e-06', 'num_tokens': '1.931e+06', 'mean_token_accuracy': '0.6133', 'epoch': '0.5042'}
+{'loss': '0.4168', 'grad_norm': '23.62', 'learning_rate': '4.958e-06', 'num_tokens': '1.931e+06', 'mean_token_accuracy': '0.8992', 'epoch': '0.5043'}
+{'loss': '0.7421', 'grad_norm': '13.38', 'learning_rate': '4.957e-06', 'num_tokens': '1.931e+06', 'mean_token_accuracy': '0.8061', 'epoch': '0.5045'}
+{'loss': '1.551', 'grad_norm': '12.5', 'learning_rate': '4.955e-06', 'num_tokens': '1.932e+06', 'mean_token_accuracy': '0.6102', 'epoch': '0.5047'}
+{'loss': '1.314', 'grad_norm': '20', 'learning_rate': '4.953e-06', 'num_tokens': '1.933e+06', 'mean_token_accuracy': '0.7276', 'epoch': '0.5049'}
+{'loss': '2.07', 'grad_norm': '24.88', 'learning_rate': '4.951e-06', 'num_tokens': '1.933e+06', 'mean_token_accuracy': '0.6196', 'epoch': '0.505'}
+{'loss': '1.956', 'grad_norm': '14.69', 'learning_rate': '4.95e-06', 'num_tokens': '1.933e+06', 'mean_token_accuracy': '0.5866', 'epoch': '0.5052'}
+{'loss': '0.8749', 'grad_norm': '11.88', 'learning_rate': '4.948e-06', 'num_tokens': '1.934e+06', 'mean_token_accuracy': '0.7691', 'epoch': '0.5054'}
+{'loss': '0.5863', 'grad_norm': '35', 'learning_rate': '4.946e-06', 'num_tokens': '1.934e+06', 'mean_token_accuracy': '0.8621', 'epoch': '0.5056'}
+{'loss': '1.544', 'grad_norm': '17.5', 'learning_rate': '4.944e-06', 'num_tokens': '1.934e+06', 'mean_token_accuracy': '0.6496', 'epoch': '0.5057'}
+{'loss': '1.851', 'grad_norm': '18', 'learning_rate': '4.943e-06', 'num_tokens': '1.935e+06', 'mean_token_accuracy': '0.6115', 'epoch': '0.5059'}
+{'loss': '1.09', 'grad_norm': '11.75', 'learning_rate': '4.941e-06', 'num_tokens': '1.935e+06', 'mean_token_accuracy': '0.6853', 'epoch': '0.5061'}
+{'loss': '0.6945', 'grad_norm': '18.75', 'learning_rate': '4.939e-06', 'num_tokens': '1.936e+06', 'mean_token_accuracy': '0.853', 'epoch': '0.5063'}
+{'loss': '2.146', 'grad_norm': '19.88', 'learning_rate': '4.937e-06', 'num_tokens': '1.936e+06', 'mean_token_accuracy': '0.5634', 'epoch': '0.5064'}
+{'loss': '1.566', 'grad_norm': '9.5', 'learning_rate': '4.936e-06', 'num_tokens': '1.937e+06', 'mean_token_accuracy': '0.671', 'epoch': '0.5066'}
+{'loss': '1.179', 'grad_norm': '5.562', 'learning_rate': '4.934e-06', 'num_tokens': '1.942e+06', 'mean_token_accuracy': '0.7009', 'epoch': '0.5068'}
+{'loss': '1.398', 'grad_norm': '11', 'learning_rate': '4.932e-06', 'num_tokens': '1.943e+06', 'mean_token_accuracy': '0.665', 'epoch': '0.5069'}
+{'loss': '1.25', 'grad_norm': '14.75', 'learning_rate': '4.931e-06', 'num_tokens': '1.943e+06', 'mean_token_accuracy': '0.6711', 'epoch': '0.5071'}
+{'loss': '0.5418', 'grad_norm': '8.812', 'learning_rate': '4.929e-06', 'num_tokens': '1.944e+06', 'mean_token_accuracy': '0.8314', 'epoch': '0.5073'}
+{'loss': '1.108', 'grad_norm': '12.25', 'learning_rate': '4.927e-06', 'num_tokens': '1.945e+06', 'mean_token_accuracy': '0.6872', 'epoch': '0.5075'}
+{'loss': '1.062', 'grad_norm': '26.75', 'learning_rate': '4.925e-06', 'num_tokens': '1.945e+06', 'mean_token_accuracy': '0.7557', 'epoch': '0.5076'}
+{'loss': '0.8648', 'grad_norm': '12.06', 'learning_rate': '4.924e-06', 'num_tokens': '1.946e+06', 'mean_token_accuracy': '0.7545', 'epoch': '0.5078'}
+{'loss': '1.273', 'grad_norm': '7.719', 'learning_rate': '4.922e-06', 'num_tokens': '1.948e+06', 'mean_token_accuracy': '0.7044', 'epoch': '0.508'}
+{'loss': '2.455', 'grad_norm': '14.88', 'learning_rate': '4.92e-06', 'num_tokens': '1.948e+06', 'mean_token_accuracy': '0.5419', 'epoch': '0.5082'}
+{'loss': '1.26', 'grad_norm': '14.62', 'learning_rate': '4.918e-06', 'num_tokens': '1.949e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.5083'}
+{'loss': '0.7311', 'grad_norm': '15.69', 'learning_rate': '4.917e-06', 'num_tokens': '1.949e+06', 'mean_token_accuracy': '0.8034', 'epoch': '0.5085'}
+{'loss': '1.824', 'grad_norm': '7.719', 'learning_rate': '4.915e-06', 'num_tokens': '1.952e+06', 'mean_token_accuracy': '0.5913', 'epoch': '0.5087'}
+{'loss': '0.5927', 'grad_norm': '7.219', 'learning_rate': '4.913e-06', 'num_tokens': '1.953e+06', 'mean_token_accuracy': '0.8243', 'epoch': '0.5089'}
+{'loss': '1.449', 'grad_norm': '12.56', 'learning_rate': '4.911e-06', 'num_tokens': '1.954e+06', 'mean_token_accuracy': '0.683', 'epoch': '0.509'}
+{'loss': '1.892', 'grad_norm': '18.12', 'learning_rate': '4.91e-06', 'num_tokens': '1.954e+06', 'mean_token_accuracy': '0.6202', 'epoch': '0.5092'}
+{'loss': '1.271', 'grad_norm': '26.75', 'learning_rate': '4.908e-06', 'num_tokens': '1.954e+06', 'mean_token_accuracy': '0.7414', 'epoch': '0.5094'}
+{'loss': '1.967', 'grad_norm': '13.19', 'learning_rate': '4.906e-06', 'num_tokens': '1.955e+06', 'mean_token_accuracy': '0.5865', 'epoch': '0.5096'}
+{'loss': '0.951', 'grad_norm': '20.62', 'learning_rate': '4.904e-06', 'num_tokens': '1.955e+06', 'mean_token_accuracy': '0.7799', 'epoch': '0.5097'}
+{'loss': '1.39', 'grad_norm': '21.38', 'learning_rate': '4.903e-06', 'num_tokens': '1.956e+06', 'mean_token_accuracy': '0.6792', 'epoch': '0.5099'}
+{'loss': '1.447', 'grad_norm': '12.81', 'learning_rate': '4.901e-06', 'num_tokens': '1.956e+06', 'mean_token_accuracy': '0.6322', 'epoch': '0.5101'}
+{'loss': '0.8734', 'grad_norm': '41.75', 'learning_rate': '4.899e-06', 'num_tokens': '1.956e+06', 'mean_token_accuracy': '0.8143', 'epoch': '0.5102'}
+{'loss': '1.895', 'grad_norm': '19', 'learning_rate': '4.898e-06', 'num_tokens': '1.957e+06', 'mean_token_accuracy': '0.5889', 'epoch': '0.5104'}
+{'loss': '0.7531', 'grad_norm': '30.25', 'learning_rate': '4.896e-06', 'num_tokens': '1.957e+06', 'mean_token_accuracy': '0.8252', 'epoch': '0.5106'}
+
 51%|█████     | 2941/5758 [16:42<15:08,  3.10it/s]
 51%|█████     | 2942/5758 [16:42<14:51,  3.16it/s]
                                                   

 51%|█████     | 2942/5758 [16:42<14:51,  3.16it/s]
 51%|█████     | 2943/5758 [16:43<15:14,  3.08it/s]
                                                   

 51%|█████     | 2943/5758 [16:43<15:14,  3.08it/s]
 51%|█████     | 2944/5758 [16:43<16:46,  2.80it/s]
                                                   

 51%|█████     | 2944/5758 [16:43<16:46,  2.80it/s]
 51%|█████     | 2945/5758 [16:43<15:54,  2.95it/s]
                                                   

 51%|█████     | 2945/5758 [16:43<15:54,  2.95it/s]
 51%|█████     | 2946/5758 [16:44<15:16,  3.07it/s]
                                                   

 51%|█████     | 2946/5758 [16:44<15:16,  3.07it/s]
 51%|█████     | 2947/5758 [16:44<14:54,  3.14it/s]
                                                   

 51%|█████     | 2947/5758 [16:44<14:54,  3.14it/s]
 51%|█████     | 2948/5758 [16:44<14:49,  3.16it/s]
                                                   

 51%|█████     | 2948/5758 [16:44<14:49,  3.16it/s]
 51%|█████     | 2949/5758 [16:45<14:45,  3.17it/s]
                                                   

 51%|█████     | 2949/5758 [16:45<14:45,  3.17it/s]
 51%|█████     | 2950/5758 [16:45<15:20,  3.05it/s]
                                                   

 51%|█████     | 2950/5758 [16:45<15:20,  3.05it/s]
 51%|█████▏    | 2951/5758 [16:45<15:34,  3.00it/s]
                                                   

 51%|█████▏    | 2951/5758 [16:45<15:34,  3.00it/s]
 51%|█████▏    | 2952/5758 [16:46<15:40,  2.98it/s]
                                                   

 51%|█████▏    | 2952/5758 [16:46<15:40,  2.98it/s]
 51%|█████▏    | 2953/5758 [16:46<15:47,  2.96it/s]
                                                   

 51%|█████▏    | 2953/5758 [16:46<15:47,  2.96it/s]
 51%|█████▏    | 2954/5758 [16:46<16:07,  2.90it/s]
                                                   

 51%|█████▏    | 2954/5758 [16:46<16:07,  2.90it/s]
 51%|█████▏    | 2955/5758 [16:47<16:35,  2.82it/s]
                                                   

 51%|█████▏    | 2955/5758 [16:47<16:35,  2.82it/s]
 51%|█████▏    | 2956/5758 [16:47<16:33,  2.82it/s]
                                                   

 51%|█████▏    | 2956/5758 [16:47<16:33,  2.82it/s]
 51%|█████▏    | 2957/5758 [16:47<17:02,  2.74it/s]
                                                   

 51%|█████▏    | 2957/5758 [16:47<17:02,  2.74it/s]
 51%|█████▏    | 2958/5758 [16:48<16:28,  2.83it/s]
                                                   

 51%|█████▏    | 2958/5758 [16:48<16:28,  2.83it/s]
 51%|█████▏    | 2959/5758 [16:48<16:22,  2.85it/s]
                                                   

 51%|█████▏    | 2959/5758 [16:48<16:22,  2.85it/s]
 51%|█████▏    | 2960/5758 [16:48<16:01,  2.91it/s]
                                                   

 51%|█████▏    | 2960/5758 [16:48<16:01,  2.91it/s]
 51%|█████▏    | 2961/5758 [16:49<15:32,  3.00it/s]
                                                   

 51%|█████▏    | 2961/5758 [16:49<15:32,  3.00it/s]
 51%|█████▏    | 2962/5758 [16:49<16:17,  2.86it/s]
                                                   

 51%|█████▏    | 2962/5758 [16:49<16:17,  2.86it/s]
 51%|█████▏    | 2963/5758 [16:49<16:31,  2.82it/s]
                                                   

 51%|█████▏    | 2963/5758 [16:49<16:31,  2.82it/s]
 51%|█████▏    | 2964/5758 [16:50<15:57,  2.92it/s]
                                                   

 51%|█████▏    | 2964/5758 [16:50<15:57,  2.92it/s]
 51%|█████▏    | 2965/5758 [16:50<15:46,  2.95it/s]
                                                   

 51%|█████▏    | 2965/5758 [16:50<15:46,  2.95it/s]
 52%|█████▏    | 2966/5758 [16:50<15:20,  3.03it/s]
                                                   

 52%|█████▏    | 2966/5758 [16:50<15:20,  3.03it/s]
 52%|█████▏    | 2967/5758 [16:51<14:59,  3.10it/s]
                                                   

 52%|█████▏    | 2967/5758 [16:51<14:59,  3.10it/s]
 52%|█████▏    | 2968/5758 [16:51<14:49,  3.14it/s]
                                                   

 52%|█████▏    | 2968/5758 [16:51<14:49,  3.14it/s]
 52%|█████▏    | 2969/5758 [16:51<15:19,  3.03it/s]
                                                   

 52%|█████▏    | 2969/5758 [16:51<15:19,  3.03it/s]
 52%|█████▏    | 2970/5758 [16:52<15:08,  3.07it/s]
                                                   

 52%|█████▏    | 2970/5758 [16:52<15:08,  3.07it/s]
 52%|█████▏    | 2971/5758 [16:52<16:39,  2.79it/s]
                                                   

 52%|█████▏    | 2971/5758 [16:52<16:39,  2.79it/s]
 52%|█████▏    | 2972/5758 [16:52<16:01,  2.90it/s]
                                                   

 52%|█████▏    | 2972/5758 [16:52<16:01,  2.90it/s]
 52%|█████▏    | 2973/5758 [16:53<16:08,  2.88it/s]
                                                   

 52%|█████▏    | 2973/5758 [16:53<16:08,  2.88it/s]
 52%|█████▏    | 2974/5758 [16:53<15:47,  2.94it/s]
                                                   

 52%|█████▏    | 2974/5758 [16:53<15:47,  2.94it/s]
 52%|█████▏    | 2975/5758 [16:53<15:13,  3.05it/s]
                                                   

 52%|█████▏    | 2975/5758 [16:53<15:13,  3.05it/s]
 52%|█████▏    | 2976/5758 [16:54<14:51,  3.12it/s]
                                                   

 52%|█████▏    | 2976/5758 [16:54<14:51,  3.12it/s]
 52%|█████▏    | 2977/5758 [16:54<14:37,  3.17it/s]
                                                   

 52%|█████▏    | 2977/5758 [16:54<14:37,  3.17it/s]
 52%|█████▏    | 2978/5758 [16:54<15:11,  3.05it/s]
                                                   

 52%|█████▏    | 2978/5758 [16:54<15:11,  3.05it/s]
 52%|█████▏    | 2979/5758 [16:55<15:23,  3.01it/s]
                                                   

 52%|█████▏    | 2979/5758 [16:55<15:23,  3.01it/s]
 52%|█████▏    | 2980/5758 [16:55<14:53,  3.11it/s]
                                                   

 52%|█████▏    | 2980/5758 [16:55<14:53,  3.11it/s]
 52%|█████▏    | 2981/5758 [16:55<15:24,  3.00it/s]
                                                   

 52%|█████▏    | 2981/5758 [16:55<15:24,  3.00it/s]
 52%|█████▏    | 2982/5758 [16:56<15:18,  3.02it/s]
                                                   

 52%|█████▏    | 2982/5758 [16:56<15:18,  3.02it/s]
 52%|█████▏    | 2983/5758 [16:56<15:08,  3.05it/s]
                                                   

 52%|█████▏    | 2983/5758 [16:56<15:08,  3.05it/s]
 52%|█████▏    | 2984/5758 [16:56<15:57,  2.90it/s]
                                                   

 52%|█████▏    | 2984/5758 [16:56<15:57,  2.90it/s]
 52%|█████▏    | 2985/5758 [16:57<16:09,  2.86it/s]
                                                   

 52%|█████▏    | 2985/5758 [16:57<16:09,  2.86it/s]
 52%|█████▏    | 2986/5758 [16:57<15:29,  2.98it/s]
                                                   

 52%|█████▏    | 2986/5758 [16:57<15:29,  2.98it/s]
 52%|█████▏    | 2987/5758 [16:57<14:54,  3.10it/s]
                                                   

 52%|█████▏    | 2987/5758 [16:57<14:54,  3.10it/s]
 52%|█████▏    | 2988/5758 [16:58<14:43,  3.13it/s]
                                                   

 52%|█████▏    | 2988/5758 [16:58<14:43,  3.13it/s]
 52%|█████▏    | 2989/5758 [16:58<15:26,  2.99it/s]
                                                   

 52%|█████▏    | 2989/5758 [16:58<15:26,  2.99it/s]
 52%|█████▏    | 2990/5758 [16:58<15:08,  3.05it/s]
                                                   

 52%|█████▏    | 2990/5758 [16:58<15:08,  3.05it/s]
 52%|█████▏    | 2991/5758 [16:59<15:51,  2.91it/s]
                                                   

 52%|█████▏    | 2991/5758 [16:59<15:51,  2.91it/s]
 52%|█████▏    | 2992/5758 [16:59<15:33,  2.96it/s]
                                                   

 52%|█████▏    | 2992/5758 [16:59<15:33,  2.96it/s]
 52%|█████▏    | 2993/5758 [16:59<14:53,  3.09it/s]
                                                   

 52%|█████▏    | 2993/5758 [16:59<14:53,  3.09it/s]
 52%|█████▏    | 2994/5758 [17:00<14:38,  3.15it/s]
                                                   

 52%|█████▏    | 2994/5758 [17:00<14:38,  3.15it/s]
 52%|█████▏    | 2995/5758 [17:00<14:21,  3.21it/s]
                                                   

 52%|█████▏    | 2995/5758 [17:00<14:21,  3.21it/s]
 52%|█████▏    | 2996/5758 [17:00<14:48,  3.11it/s]
                                                   
{'loss': '1.257', 'grad_norm': '12.25', 'learning_rate': '4.894e-06', 'num_tokens': '1.958e+06', 'mean_token_accuracy': '0.6508', 'epoch': '0.5108'}
+{'loss': '0.8091', 'grad_norm': '22.5', 'learning_rate': '4.892e-06', 'num_tokens': '1.958e+06', 'mean_token_accuracy': '0.8036', 'epoch': '0.5109'}
+{'loss': '0.5882', 'grad_norm': '6.219', 'learning_rate': '4.891e-06', 'num_tokens': '1.96e+06', 'mean_token_accuracy': '0.8274', 'epoch': '0.5111'}
+{'loss': '1.549', 'grad_norm': '7.531', 'learning_rate': '4.889e-06', 'num_tokens': '1.961e+06', 'mean_token_accuracy': '0.6813', 'epoch': '0.5113'}
+{'loss': '0.8745', 'grad_norm': '11.81', 'learning_rate': '4.887e-06', 'num_tokens': '1.962e+06', 'mean_token_accuracy': '0.7558', 'epoch': '0.5115'}
+{'loss': '1.988', 'grad_norm': '50.75', 'learning_rate': '4.885e-06', 'num_tokens': '1.962e+06', 'mean_token_accuracy': '0.6111', 'epoch': '0.5116'}
+{'loss': '1.373', 'grad_norm': '13.81', 'learning_rate': '4.884e-06', 'num_tokens': '1.963e+06', 'mean_token_accuracy': '0.638', 'epoch': '0.5118'}
+{'loss': '0.9798', 'grad_norm': '27', 'learning_rate': '4.882e-06', 'num_tokens': '1.963e+06', 'mean_token_accuracy': '0.7637', 'epoch': '0.512'}
+{'loss': '0.6215', 'grad_norm': '8.062', 'learning_rate': '4.88e-06', 'num_tokens': '1.964e+06', 'mean_token_accuracy': '0.8086', 'epoch': '0.5122'}
+{'loss': '1.163', 'grad_norm': '14.62', 'learning_rate': '4.878e-06', 'num_tokens': '1.965e+06', 'mean_token_accuracy': '0.7303', 'epoch': '0.5123'}
+{'loss': '1.015', 'grad_norm': '14.38', 'learning_rate': '4.877e-06', 'num_tokens': '1.965e+06', 'mean_token_accuracy': '0.7677', 'epoch': '0.5125'}
+{'loss': '1.026', 'grad_norm': '6.969', 'learning_rate': '4.875e-06', 'num_tokens': '1.967e+06', 'mean_token_accuracy': '0.7392', 'epoch': '0.5127'}
+{'loss': '2.092', 'grad_norm': '16.12', 'learning_rate': '4.873e-06', 'num_tokens': '1.967e+06', 'mean_token_accuracy': '0.5828', 'epoch': '0.5129'}
+{'loss': '1.987', 'grad_norm': '20.75', 'learning_rate': '4.871e-06', 'num_tokens': '1.968e+06', 'mean_token_accuracy': '0.5919', 'epoch': '0.513'}
+{'loss': '1.945', 'grad_norm': '13.62', 'learning_rate': '4.87e-06', 'num_tokens': '1.968e+06', 'mean_token_accuracy': '0.605', 'epoch': '0.5132'}
+{'loss': '0.9145', 'grad_norm': '18.5', 'learning_rate': '4.868e-06', 'num_tokens': '1.969e+06', 'mean_token_accuracy': '0.7755', 'epoch': '0.5134'}
+{'loss': '1.275', 'grad_norm': '14.38', 'learning_rate': '4.866e-06', 'num_tokens': '1.969e+06', 'mean_token_accuracy': '0.6537', 'epoch': '0.5135'}
+{'loss': '1.226', 'grad_norm': '11.69', 'learning_rate': '4.865e-06', 'num_tokens': '1.97e+06', 'mean_token_accuracy': '0.661', 'epoch': '0.5137'}
+{'loss': '1.908', 'grad_norm': '10.12', 'learning_rate': '4.863e-06', 'num_tokens': '1.971e+06', 'mean_token_accuracy': '0.628', 'epoch': '0.5139'}
+{'loss': '1.295', 'grad_norm': '12.12', 'learning_rate': '4.861e-06', 'num_tokens': '1.972e+06', 'mean_token_accuracy': '0.6247', 'epoch': '0.5141'}
+{'loss': '1.737', 'grad_norm': '23.62', 'learning_rate': '4.859e-06', 'num_tokens': '1.972e+06', 'mean_token_accuracy': '0.6747', 'epoch': '0.5142'}
+{'loss': '1.286', 'grad_norm': '13.88', 'learning_rate': '4.858e-06', 'num_tokens': '1.972e+06', 'mean_token_accuracy': '0.6558', 'epoch': '0.5144'}
+{'loss': '0.7304', 'grad_norm': '26.38', 'learning_rate': '4.856e-06', 'num_tokens': '1.973e+06', 'mean_token_accuracy': '0.7966', 'epoch': '0.5146'}
+{'loss': '2.065', 'grad_norm': '15.12', 'learning_rate': '4.854e-06', 'num_tokens': '1.973e+06', 'mean_token_accuracy': '0.5496', 'epoch': '0.5148'}
+{'loss': '1.327', 'grad_norm': '12.5', 'learning_rate': '4.852e-06', 'num_tokens': '1.974e+06', 'mean_token_accuracy': '0.635', 'epoch': '0.5149'}
+{'loss': '1.441', 'grad_norm': '13.62', 'learning_rate': '4.851e-06', 'num_tokens': '1.975e+06', 'mean_token_accuracy': '0.6223', 'epoch': '0.5151'}
+{'loss': '1.112', 'grad_norm': '14', 'learning_rate': '4.849e-06', 'num_tokens': '1.976e+06', 'mean_token_accuracy': '0.6865', 'epoch': '0.5153'}
+{'loss': '2.056', 'grad_norm': '12.19', 'learning_rate': '4.847e-06', 'num_tokens': '1.976e+06', 'mean_token_accuracy': '0.5836', 'epoch': '0.5155'}
+{'loss': '0.8906', 'grad_norm': '15.5', 'learning_rate': '4.845e-06', 'num_tokens': '1.977e+06', 'mean_token_accuracy': '0.7971', 'epoch': '0.5156'}
+{'loss': '0.8859', 'grad_norm': '30.25', 'learning_rate': '4.844e-06', 'num_tokens': '1.977e+06', 'mean_token_accuracy': '0.8', 'epoch': '0.5158'}
+{'loss': '0.8633', 'grad_norm': '9.875', 'learning_rate': '4.842e-06', 'num_tokens': '1.978e+06', 'mean_token_accuracy': '0.7978', 'epoch': '0.516'}
+{'loss': '1.492', 'grad_norm': '11.69', 'learning_rate': '4.84e-06', 'num_tokens': '1.978e+06', 'mean_token_accuracy': '0.6704', 'epoch': '0.5162'}
+{'loss': '2.179', 'grad_norm': '20.25', 'learning_rate': '4.838e-06', 'num_tokens': '1.978e+06', 'mean_token_accuracy': '0.5892', 'epoch': '0.5163'}
+{'loss': '1.088', 'grad_norm': '11', 'learning_rate': '4.837e-06', 'num_tokens': '1.979e+06', 'mean_token_accuracy': '0.6994', 'epoch': '0.5165'}
+{'loss': '1.331', 'grad_norm': '13.12', 'learning_rate': '4.835e-06', 'num_tokens': '1.98e+06', 'mean_token_accuracy': '0.63', 'epoch': '0.5167'}
+{'loss': '0.6451', 'grad_norm': '11.62', 'learning_rate': '4.833e-06', 'num_tokens': '1.981e+06', 'mean_token_accuracy': '0.8257', 'epoch': '0.5168'}
+{'loss': '0.4789', 'grad_norm': '31.5', 'learning_rate': '4.832e-06', 'num_tokens': '1.981e+06', 'mean_token_accuracy': '0.8701', 'epoch': '0.517'}
+{'loss': '2.069', 'grad_norm': '19.12', 'learning_rate': '4.83e-06', 'num_tokens': '1.981e+06', 'mean_token_accuracy': '0.5714', 'epoch': '0.5172'}
+{'loss': '0.6491', 'grad_norm': '12.06', 'learning_rate': '4.828e-06', 'num_tokens': '1.982e+06', 'mean_token_accuracy': '0.8455', 'epoch': '0.5174'}
+{'loss': '1.272', 'grad_norm': '13.62', 'learning_rate': '4.826e-06', 'num_tokens': '1.982e+06', 'mean_token_accuracy': '0.653', 'epoch': '0.5175'}
+{'loss': '0.9862', 'grad_norm': '4.906', 'learning_rate': '4.825e-06', 'num_tokens': '1.986e+06', 'mean_token_accuracy': '0.7299', 'epoch': '0.5177'}
+{'loss': '1.535', 'grad_norm': '15.81', 'learning_rate': '4.823e-06', 'num_tokens': '1.987e+06', 'mean_token_accuracy': '0.6609', 'epoch': '0.5179'}
+{'loss': '1.323', 'grad_norm': '24.25', 'learning_rate': '4.821e-06', 'num_tokens': '1.987e+06', 'mean_token_accuracy': '0.7031', 'epoch': '0.5181'}
+{'loss': '1.993', 'grad_norm': '16.62', 'learning_rate': '4.819e-06', 'num_tokens': '1.988e+06', 'mean_token_accuracy': '0.6051', 'epoch': '0.5182'}
+{'loss': '1.003', 'grad_norm': '5.125', 'learning_rate': '4.818e-06', 'num_tokens': '1.992e+06', 'mean_token_accuracy': '0.7289', 'epoch': '0.5184'}
+{'loss': '1.032', 'grad_norm': '11.94', 'learning_rate': '4.816e-06', 'num_tokens': '1.992e+06', 'mean_token_accuracy': '0.7151', 'epoch': '0.5186'}
+{'loss': '1.041', 'grad_norm': '13', 'learning_rate': '4.814e-06', 'num_tokens': '1.993e+06', 'mean_token_accuracy': '0.7069', 'epoch': '0.5188'}
+{'loss': '1.411', 'grad_norm': '13.12', 'learning_rate': '4.812e-06', 'num_tokens': '1.994e+06', 'mean_token_accuracy': '0.6397', 'epoch': '0.5189'}
+{'loss': '1.183', 'grad_norm': '12.38', 'learning_rate': '4.811e-06', 'num_tokens': '1.994e+06', 'mean_token_accuracy': '0.7673', 'epoch': '0.5191'}
+{'loss': '2.014', 'grad_norm': '11.5', 'learning_rate': '4.809e-06', 'num_tokens': '1.995e+06', 'mean_token_accuracy': '0.5836', 'epoch': '0.5193'}
+{'loss': '1.012', 'grad_norm': '15.75', 'learning_rate': '4.807e-06', 'num_tokens': '1.995e+06', 'mean_token_accuracy': '0.7899', 'epoch': '0.5195'}
+{'loss': '1.435', 'grad_norm': '12.25', 'learning_rate': '4.805e-06', 'num_tokens': '1.996e+06', 'mean_token_accuracy': '0.6247', 'epoch': '0.5196'}
+{'loss': '1.518', 'grad_norm': '10.12', 'learning_rate': '4.804e-06', 'num_tokens': '1.997e+06', 'mean_token_accuracy': '0.6461', 'epoch': '0.5198'}
+{'loss': '0.9007', 'grad_norm': '12.62', 'learning_rate': '4.802e-06', 'num_tokens': '1.998e+06', 'mean_token_accuracy': '0.7237', 'epoch': '0.52'}
+{'loss': '1.038', 'grad_norm': '8.312', 'learning_rate': '4.8e-06', 'num_tokens': '1.999e+06', 'mean_token_accuracy': '0.7704', 'epoch': '0.5201'}
+
 52%|█████▏    | 2996/5758 [17:00<14:48,  3.11it/s]
 52%|█████▏    | 2997/5758 [17:01<14:56,  3.08it/s]
                                                   

 52%|█████▏    | 2997/5758 [17:01<14:56,  3.08it/s]
 52%|█████▏    | 2998/5758 [17:01<15:18,  3.00it/s]
                                                   

 52%|█████▏    | 2998/5758 [17:01<15:18,  3.00it/s]
 52%|█████▏    | 2999/5758 [17:01<17:20,  2.65it/s]
                                                   

 52%|█████▏    | 2999/5758 [17:01<17:20,  2.65it/s]
 52%|█████▏    | 3000/5758 [17:02<17:16,  2.66it/s]
                                                   

 52%|█████▏    | 3000/5758 [17:02<17:16,  2.66it/s]{'loss': '1.177', 'grad_norm': '11.06', 'learning_rate': '4.799e-06', 'num_tokens': '2e+06', 'mean_token_accuracy': '0.69', 'epoch': '0.5203'}
+{'loss': '1.618', 'grad_norm': '7.75', 'learning_rate': '4.797e-06', 'num_tokens': '2.002e+06', 'mean_token_accuracy': '0.6447', 'epoch': '0.5205'}
+{'loss': '1.177', 'grad_norm': '6.906', 'learning_rate': '4.795e-06', 'num_tokens': '2.004e+06', 'mean_token_accuracy': '0.7163', 'epoch': '0.5207'}
+{'loss': '1.174', 'grad_norm': '19.25', 'learning_rate': '4.793e-06', 'num_tokens': '2.004e+06', 'mean_token_accuracy': '0.7581', 'epoch': '0.5208'}
+{'loss': '1.749', 'grad_norm': '48.75', 'learning_rate': '4.792e-06', 'num_tokens': '2.004e+06', 'mean_token_accuracy': '0.6803', 'epoch': '0.521'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.60it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.59it/s]
+
 52%|█████▏    | 3001/5758 [17:04<41:10,  1.12it/s]
                                                   

 52%|█████▏    | 3001/5758 [17:04<41:10,  1.12it/s]
 52%|█████▏    | 3002/5758 [17:04<34:00,  1.35it/s]
                                                   

 52%|█████▏    | 3002/5758 [17:04<34:00,  1.35it/s]
 52%|█████▏    | 3003/5758 [17:05<28:26,  1.61it/s]
                                                   

 52%|█████▏    | 3003/5758 [17:05<28:26,  1.61it/s]
 52%|█████▏    | 3004/5758 [17:05<24:50,  1.85it/s]
                                                   

 52%|█████▏    | 3004/5758 [17:05<24:50,  1.85it/s]
 52%|█████▏    | 3005/5758 [17:05<21:46,  2.11it/s]
                                                   

 52%|█████▏    | 3005/5758 [17:05<21:46,  2.11it/s]
 52%|█████▏    | 3006/5758 [17:06<19:34,  2.34it/s]
                                                   

 52%|█████▏    | 3006/5758 [17:06<19:34,  2.34it/s]
 52%|█████▏    | 3007/5758 [17:06<17:44,  2.59it/s]
                                                   

 52%|█████▏    | 3007/5758 [17:06<17:44,  2.59it/s]
 52%|█████▏    | 3008/5758 [17:06<16:32,  2.77it/s]
                                                   

 52%|█████▏    | 3008/5758 [17:06<16:32,  2.77it/s]
 52%|█████▏    | 3009/5758 [17:07<16:13,  2.82it/s]
                                                   

 52%|█████▏    | 3009/5758 [17:07<16:13,  2.82it/s]
 52%|█████▏    | 3010/5758 [17:07<15:24,  2.97it/s]
                                                   

 52%|█████▏    | 3010/5758 [17:07<15:24,  2.97it/s]
 52%|█████▏    | 3011/5758 [17:07<15:19,  2.99it/s]
                                                   

 52%|█████▏    | 3011/5758 [17:07<15:19,  2.99it/s]
 52%|█████▏    | 3012/5758 [17:08<15:11,  3.01it/s]
                                                   

 52%|█████▏    | 3012/5758 [17:08<15:11,  3.01it/s]
 52%|█████▏    | 3013/5758 [17:08<15:49,  2.89it/s]
                                                   

 52%|█████▏    | 3013/5758 [17:08<15:49,  2.89it/s]
 52%|█████▏    | 3014/5758 [17:08<15:24,  2.97it/s]
                                                   

 52%|█████▏    | 3014/5758 [17:08<15:24,  2.97it/s]
 52%|█████▏    | 3015/5758 [17:09<15:37,  2.93it/s]
                                                   

 52%|█████▏    | 3015/5758 [17:09<15:37,  2.93it/s]
 52%|█████▏    | 3016/5758 [17:09<15:16,  2.99it/s]
                                                   

 52%|█████▏    | 3016/5758 [17:09<15:16,  2.99it/s]
 52%|█████▏    | 3017/5758 [17:09<16:08,  2.83it/s]
                                                   

 52%|█████▏    | 3017/5758 [17:09<16:08,  2.83it/s]
 52%|█████▏    | 3018/5758 [17:10<16:03,  2.84it/s]
                                                   

 52%|█████▏    | 3018/5758 [17:10<16:03,  2.84it/s]
 52%|█████▏    | 3019/5758 [17:10<15:24,  2.96it/s]
                                                   

 52%|█████▏    | 3019/5758 [17:10<15:24,  2.96it/s]
 52%|█████▏    | 3020/5758 [17:10<14:52,  3.07it/s]
                                                   

 52%|█████▏    | 3020/5758 [17:10<14:52,  3.07it/s]
 52%|█████▏    | 3021/5758 [17:11<15:44,  2.90it/s]
                                                   

 52%|█████▏    | 3021/5758 [17:11<15:44,  2.90it/s]
 52%|█████▏    | 3022/5758 [17:11<15:24,  2.96it/s]
                                                   

 52%|█████▏    | 3022/5758 [17:11<15:24,  2.96it/s]
 53%|█████▎    | 3023/5758 [17:11<15:17,  2.98it/s]
                                                   

 53%|█████▎    | 3023/5758 [17:11<15:17,  2.98it/s]
 53%|█████▎    | 3024/5758 [17:12<15:39,  2.91it/s]
                                                   

 53%|█████▎    | 3024/5758 [17:12<15:39,  2.91it/s]
 53%|█████▎    | 3025/5758 [17:12<17:48,  2.56it/s]
                                                   

 53%|█████▎    | 3025/5758 [17:12<17:48,  2.56it/s]
 53%|█████▎    | 3026/5758 [17:13<18:07,  2.51it/s]
                                                   

 53%|█████▎    | 3026/5758 [17:13<18:07,  2.51it/s]
 53%|█████▎    | 3027/5758 [17:13<18:42,  2.43it/s]
                                                   

 53%|█████▎    | 3027/5758 [17:13<18:42,  2.43it/s]
 53%|█████▎    | 3028/5758 [17:13<18:20,  2.48it/s]
                                                   

 53%|█████▎    | 3028/5758 [17:13<18:20,  2.48it/s]
 53%|█████▎    | 3029/5758 [17:14<17:02,  2.67it/s]
                                                   

 53%|█████▎    | 3029/5758 [17:14<17:02,  2.67it/s]
 53%|█████▎    | 3030/5758 [17:14<16:24,  2.77it/s]
                                                   

 53%|█████▎    | 3030/5758 [17:14<16:24,  2.77it/s]
 53%|█████▎    | 3031/5758 [17:14<15:32,  2.92it/s]
                                                   

 53%|█████▎    | 3031/5758 [17:14<15:32,  2.92it/s]
 53%|█████▎    | 3032/5758 [17:15<16:04,  2.83it/s]
                                                   

 53%|█████▎    | 3032/5758 [17:15<16:04,  2.83it/s]
 53%|█████▎    | 3033/5758 [17:15<15:12,  2.99it/s]
                                                   

 53%|█████▎    | 3033/5758 [17:15<15:12,  2.99it/s]
 53%|█████▎    | 3034/5758 [17:15<14:55,  3.04it/s]
                                                   

 53%|█████▎    | 3034/5758 [17:15<14:55,  3.04it/s]
 53%|█████▎    | 3035/5758 [17:16<14:45,  3.07it/s]
                                                   

 53%|█████▎    | 3035/5758 [17:16<14:45,  3.07it/s]
 53%|█████▎    | 3036/5758 [17:16<14:59,  3.03it/s]
                                                   

 53%|█████▎    | 3036/5758 [17:16<14:59,  3.03it/s]
 53%|█████▎    | 3037/5758 [17:16<15:11,  2.99it/s]
                                                   

 53%|█████▎    | 3037/5758 [17:16<15:11,  2.99it/s]
 53%|█████▎    | 3038/5758 [17:17<15:39,  2.89it/s]
                                                   

 53%|█████▎    | 3038/5758 [17:17<15:39,  2.89it/s]
 53%|█████▎    | 3039/5758 [17:17<15:04,  3.00it/s]
                                                   

 53%|█████���    | 3039/5758 [17:17<15:04,  3.00it/s]
 53%|█████▎    | 3040/5758 [17:17<14:42,  3.08it/s]
                                                   

 53%|█████▎    | 3040/5758 [17:17<14:42,  3.08it/s]
 53%|█████▎    | 3041/5758 [17:18<14:46,  3.06it/s]
                                                   

 53%|█████▎    | 3041/5758 [17:18<14:46,  3.06it/s]
 53%|█████▎    | 3042/5758 [17:18<15:20,  2.95it/s]
                                                   

 53%|█████▎    | 3042/5758 [17:18<15:20,  2.95it/s]
 53%|█████▎    | 3043/5758 [17:18<14:43,  3.07it/s]
                                                   

 53%|█████▎    | 3043/5758 [17:18<14:43,  3.07it/s]
 53%|█████▎    | 3044/5758 [17:19<15:28,  2.92it/s]
                                                   

 53%|█████▎    | 3044/5758 [17:19<15:28,  2.92it/s]
 53%|█████▎    | 3045/5758 [17:19<14:49,  3.05it/s]
                                                   

 53%|█████▎    | 3045/5758 [17:19<14:49,  3.05it/s]
 53%|█████▎    | 3046/5758 [17:19<15:23,  2.94it/s]
                                                   

 53%|█████▎    | 3046/5758 [17:19<15:23,  2.94it/s]
 53%|█████▎    | 3047/5758 [17:20<15:24,  2.93it/s]
                                                   

 53%|█████▎    | 3047/5758 [17:20<15:24,  2.93it/s]
 53%|█████▎    | 3048/5758 [17:20<15:01,  3.01it/s]
                                                   

 53%|█████▎    | 3048/5758 [17:20<15:01,  3.01it/s]
 53%|█████▎    | 3049/5758 [17:20<15:11,  2.97it/s]
                                                   

 53%|█████▎    | 3049/5758 [17:20<15:11,  2.97it/s]
 53%|█████▎    | 3050/5758 [17:21<15:40,  2.88it/s]
                                                   

 53%|█████▎    | 3050/5758 [17:21<15:40,  2.88it/s]
 53%|█████▎    | 3051/5758 [17:21<15:02,  3.00it/s]
                                                   

 53%|█████▎    | 3051/5758 [17:21<15:02,  3.00it/s]
 53%|█████▎    | 3052/5758 [17:21<14:38,  3.08it/s]
                                                   

 53%|█████▎    | 3052/5758 [17:21<14:38,  3.08it/s]
 53%|█████▎    | 3053/5758 [17:22<15:44,  2.86it/s]
                                                   

 53%|█████▎    | 3053/5758 [17:22<15:44,  2.86it/s]
 53%|█████▎    | 3054/5758 [17:22<16:00,  2.82it/s]
                                                   

 53%|█████▎    | 3054/5758 [17:22<16:00,  2.82it/s]
 53%|█████▎    | 3055/5758 [17:22<15:29,  2.91it/s]
                                                   

 53%|█████▎    | 3055/5758 [17:22<15:29,  2.91it/s]
 53%|█████▎    | 3056/5758 [17:23<15:02,  2.99it/s]
                                                   
{'loss': '0.4847', 'grad_norm': '6.5', 'learning_rate': '4.79e-06', 'num_tokens': '2.006e+06', 'mean_token_accuracy': '0.8542', 'epoch': '0.5212'}
+{'loss': '0.938', 'grad_norm': '13.19', 'learning_rate': '4.788e-06', 'num_tokens': '2.007e+06', 'mean_token_accuracy': '0.7339', 'epoch': '0.5214'}
+{'loss': '0.6564', 'grad_norm': '19.12', 'learning_rate': '4.786e-06', 'num_tokens': '2.007e+06', 'mean_token_accuracy': '0.8565', 'epoch': '0.5215'}
+{'loss': '1.935', 'grad_norm': '19.75', 'learning_rate': '4.785e-06', 'num_tokens': '2.007e+06', 'mean_token_accuracy': '0.6136', 'epoch': '0.5217'}
+{'loss': '0.6807', 'grad_norm': '8.062', 'learning_rate': '4.783e-06', 'num_tokens': '2.009e+06', 'mean_token_accuracy': '0.803', 'epoch': '0.5219'}
+{'loss': '1.162', 'grad_norm': '12.31', 'learning_rate': '4.781e-06', 'num_tokens': '2.01e+06', 'mean_token_accuracy': '0.6804', 'epoch': '0.5221'}
+{'loss': '0.7948', 'grad_norm': '21.5', 'learning_rate': '4.779e-06', 'num_tokens': '2.01e+06', 'mean_token_accuracy': '0.825', 'epoch': '0.5222'}
+{'loss': '1.838', 'grad_norm': '19.88', 'learning_rate': '4.778e-06', 'num_tokens': '2.01e+06', 'mean_token_accuracy': '0.6269', 'epoch': '0.5224'}
+{'loss': '2.423', 'grad_norm': '10.38', 'learning_rate': '4.776e-06', 'num_tokens': '2.011e+06', 'mean_token_accuracy': '0.5248', 'epoch': '0.5226'}
+{'loss': '1.119', 'grad_norm': '5.75', 'learning_rate': '4.774e-06', 'num_tokens': '2.013e+06', 'mean_token_accuracy': '0.7443', 'epoch': '0.5228'}
+{'loss': '0.7845', 'grad_norm': '25.5', 'learning_rate': '4.772e-06', 'num_tokens': '2.014e+06', 'mean_token_accuracy': '0.8551', 'epoch': '0.5229'}
+{'loss': '1.835', 'grad_norm': '11.12', 'learning_rate': '4.771e-06', 'num_tokens': '2.014e+06', 'mean_token_accuracy': '0.606', 'epoch': '0.5231'}
+{'loss': '1.324', 'grad_norm': '10.88', 'learning_rate': '4.769e-06', 'num_tokens': '2.015e+06', 'mean_token_accuracy': '0.6814', 'epoch': '0.5233'}
+{'loss': '0.7612', 'grad_norm': '20.5', 'learning_rate': '4.767e-06', 'num_tokens': '2.015e+06', 'mean_token_accuracy': '0.8205', 'epoch': '0.5234'}
+{'loss': '1.674', 'grad_norm': '10.69', 'learning_rate': '4.766e-06', 'num_tokens': '2.016e+06', 'mean_token_accuracy': '0.619', 'epoch': '0.5236'}
+{'loss': '1.113', 'grad_norm': '13.56', 'learning_rate': '4.764e-06', 'num_tokens': '2.017e+06', 'mean_token_accuracy': '0.6748', 'epoch': '0.5238'}
+{'loss': '1.793', 'grad_norm': '14', 'learning_rate': '4.762e-06', 'num_tokens': '2.018e+06', 'mean_token_accuracy': '0.6212', 'epoch': '0.524'}
+{'loss': '0.5846', 'grad_norm': '9.625', 'learning_rate': '4.76e-06', 'num_tokens': '2.018e+06', 'mean_token_accuracy': '0.8268', 'epoch': '0.5241'}
+{'loss': '1.078', 'grad_norm': '11.5', 'learning_rate': '4.759e-06', 'num_tokens': '2.019e+06', 'mean_token_accuracy': '0.7017', 'epoch': '0.5243'}
+{'loss': '1.698', 'grad_norm': '15.62', 'learning_rate': '4.757e-06', 'num_tokens': '2.02e+06', 'mean_token_accuracy': '0.6591', 'epoch': '0.5245'}
+{'loss': '1.53', 'grad_norm': '23.12', 'learning_rate': '4.755e-06', 'num_tokens': '2.02e+06', 'mean_token_accuracy': '0.6875', 'epoch': '0.5247'}
+{'loss': '1.413', 'grad_norm': '14.44', 'learning_rate': '4.753e-06', 'num_tokens': '2.021e+06', 'mean_token_accuracy': '0.706', 'epoch': '0.5248'}
+{'loss': '1.204', 'grad_norm': '17.62', 'learning_rate': '4.752e-06', 'num_tokens': '2.021e+06', 'mean_token_accuracy': '0.7543', 'epoch': '0.525'}
+{'loss': '1.795', 'grad_norm': '10.31', 'learning_rate': '4.75e-06', 'num_tokens': '2.022e+06', 'mean_token_accuracy': '0.633', 'epoch': '0.5252'}
+{'loss': '1.478', 'grad_norm': '11', 'learning_rate': '4.748e-06', 'num_tokens': '2.023e+06', 'mean_token_accuracy': '0.5991', 'epoch': '0.5254'}
+{'loss': '1.996', 'grad_norm': '37', 'learning_rate': '4.746e-06', 'num_tokens': '2.024e+06', 'mean_token_accuracy': '0.6597', 'epoch': '0.5255'}
+{'loss': '1.182', 'grad_norm': '5.125', 'learning_rate': '4.745e-06', 'num_tokens': '2.027e+06', 'mean_token_accuracy': '0.7138', 'epoch': '0.5257'}
+{'loss': '0.9915', 'grad_norm': '17.38', 'learning_rate': '4.743e-06', 'num_tokens': '2.028e+06', 'mean_token_accuracy': '0.7641', 'epoch': '0.5259'}
+{'loss': '1.422', 'grad_norm': '12.31', 'learning_rate': '4.741e-06', 'num_tokens': '2.028e+06', 'mean_token_accuracy': '0.6364', 'epoch': '0.5261'}
+{'loss': '1.385', 'grad_norm': '11.12', 'learning_rate': '4.739e-06', 'num_tokens': '2.029e+06', 'mean_token_accuracy': '0.6998', 'epoch': '0.5262'}
+{'loss': '0.7764', 'grad_norm': '16.25', 'learning_rate': '4.738e-06', 'num_tokens': '2.03e+06', 'mean_token_accuracy': '0.8179', 'epoch': '0.5264'}
+{'loss': '2.183', 'grad_norm': '19.25', 'learning_rate': '4.736e-06', 'num_tokens': '2.03e+06', 'mean_token_accuracy': '0.5526', 'epoch': '0.5266'}
+{'loss': '1.99', 'grad_norm': '16.62', 'learning_rate': '4.734e-06', 'num_tokens': '2.031e+06', 'mean_token_accuracy': '0.6009', 'epoch': '0.5267'}
+{'loss': '1.364', 'grad_norm': '12.69', 'learning_rate': '4.733e-06', 'num_tokens': '2.031e+06', 'mean_token_accuracy': '0.633', 'epoch': '0.5269'}
+{'loss': '1.326', 'grad_norm': '27.75', 'learning_rate': '4.731e-06', 'num_tokens': '2.032e+06', 'mean_token_accuracy': '0.7186', 'epoch': '0.5271'}
+{'loss': '2.082', 'grad_norm': '16.25', 'learning_rate': '4.729e-06', 'num_tokens': '2.032e+06', 'mean_token_accuracy': '0.621', 'epoch': '0.5273'}
+{'loss': '1.022', 'grad_norm': '32.75', 'learning_rate': '4.727e-06', 'num_tokens': '2.032e+06', 'mean_token_accuracy': '0.7941', 'epoch': '0.5274'}
+{'loss': '1.436', 'grad_norm': '20.75', 'learning_rate': '4.726e-06', 'num_tokens': '2.033e+06', 'mean_token_accuracy': '0.6937', 'epoch': '0.5276'}
+{'loss': '1.941', 'grad_norm': '30.75', 'learning_rate': '4.724e-06', 'num_tokens': '2.033e+06', 'mean_token_accuracy': '0.593', 'epoch': '0.5278'}
+{'loss': '0.7371', 'grad_norm': '9.938', 'learning_rate': '4.722e-06', 'num_tokens': '2.034e+06', 'mean_token_accuracy': '0.8029', 'epoch': '0.528'}
+{'loss': '0.959', 'grad_norm': '10.75', 'learning_rate': '4.72e-06', 'num_tokens': '2.034e+06', 'mean_token_accuracy': '0.752', 'epoch': '0.5281'}
+{'loss': '1.065', 'grad_norm': '14.69', 'learning_rate': '4.719e-06', 'num_tokens': '2.035e+06', 'mean_token_accuracy': '0.7116', 'epoch': '0.5283'}
+{'loss': '1.43', 'grad_norm': '13.94', 'learning_rate': '4.717e-06', 'num_tokens': '2.036e+06', 'mean_token_accuracy': '0.6165', 'epoch': '0.5285'}
+{'loss': '1.57', 'grad_norm': '18.38', 'learning_rate': '4.715e-06', 'num_tokens': '2.036e+06', 'mean_token_accuracy': '0.6696', 'epoch': '0.5287'}
+{'loss': '1.275', 'grad_norm': '9.375', 'learning_rate': '4.713e-06', 'num_tokens': '2.037e+06', 'mean_token_accuracy': '0.715', 'epoch': '0.5288'}
+{'loss': '0.8765', 'grad_norm': '13.62', 'learning_rate': '4.712e-06', 'num_tokens': '2.038e+06', 'mean_token_accuracy': '0.7945', 'epoch': '0.529'}
+{'loss': '1.493', 'grad_norm': '11.56', 'learning_rate': '4.71e-06', 'num_tokens': '2.039e+06', 'mean_token_accuracy': '0.6121', 'epoch': '0.5292'}
+{'loss': '1.11', 'grad_norm': '11.81', 'learning_rate': '4.708e-06', 'num_tokens': '2.04e+06', 'mean_token_accuracy': '0.6831', 'epoch': '0.5294'}
+{'loss': '1.69', 'grad_norm': '17.75', 'learning_rate': '4.706e-06', 'num_tokens': '2.04e+06', 'mean_token_accuracy': '0.63', 'epoch': '0.5295'}
+{'loss': '1.008', 'grad_norm': '13.19', 'learning_rate': '4.705e-06', 'num_tokens': '2.041e+06', 'mean_token_accuracy': '0.7257', 'epoch': '0.5297'}
+{'loss': '0.4268', 'grad_norm': '22.12', 'learning_rate': '4.703e-06', 'num_tokens': '2.041e+06', 'mean_token_accuracy': '0.8994', 'epoch': '0.5299'}
+{'loss': '1.149', 'grad_norm': '38.75', 'learning_rate': '4.701e-06', 'num_tokens': '2.041e+06', 'mean_token_accuracy': '0.7595', 'epoch': '0.53'}
+{'loss': '0.7788', 'grad_norm': '20.62', 'learning_rate': '4.7e-06', 'num_tokens': '2.041e+06', 'mean_token_accuracy': '0.84', 'epoch': '0.5302'}
+{'loss': '2.192', 'grad_norm': '10.94', 'learning_rate': '4.698e-06', 'num_tokens': '2.042e+06', 'mean_token_accuracy': '0.5515', 'epoch': '0.5304'}
+{'loss': '1.008', 'grad_norm': '11.44', 'learning_rate': '4.696e-06', 'num_tokens': '2.043e+06', 'mean_token_accuracy': '0.6826', 'epoch': '0.5306'}
+
 53%|█████▎    | 3056/5758 [17:23<15:02,  2.99it/s]
 53%|█████▎    | 3057/5758 [17:23<14:35,  3.09it/s]
                                                   

 53%|█████▎    | 3057/5758 [17:23<14:35,  3.09it/s]
 53%|█████▎    | 3058/5758 [17:23<15:25,  2.92it/s]
                                                   

 53%|█████▎    | 3058/5758 [17:23<15:25,  2.92it/s]
 53%|█████▎    | 3059/5758 [17:24<15:12,  2.96it/s]
                                                   

 53%|█████▎    | 3059/5758 [17:24<15:12,  2.96it/s]
 53%|█████▎    | 3060/5758 [17:24<15:32,  2.89it/s]
                                                   

 53%|█████▎    | 3060/5758 [17:24<15:32,  2.89it/s]
 53%|█████▎    | 3061/5758 [17:24<14:42,  3.05it/s]
                                                   

 53%|█████▎    | 3061/5758 [17:24<14:42,  3.05it/s]
 53%|█████▎    | 3062/5758 [17:25<15:19,  2.93it/s]
                                                   

 53%|█████▎    | 3062/5758 [17:25<15:19,  2.93it/s]
 53%|█████▎    | 3063/5758 [17:25<14:46,  3.04it/s]
                                                   

 53%|█████▎    | 3063/5758 [17:25<14:46,  3.04it/s]
 53%|█████▎    | 3064/5758 [17:25<15:03,  2.98it/s]
                                                   

 53%|█████▎    | 3064/5758 [17:25<15:03,  2.98it/s]
 53%|█████▎    | 3065/5758 [17:26<15:33,  2.88it/s]
                                                   

 53%|█████▎    | 3065/5758 [17:26<15:33,  2.88it/s]
 53%|█████▎    | 3066/5758 [17:26<15:13,  2.95it/s]
                                                   

 53%|█████▎    | 3066/5758 [17:26<15:13,  2.95it/s]
 53%|█████▎    | 3067/5758 [17:26<14:49,  3.03it/s]
                                                   

 53%|█████▎    | 3067/5758 [17:26<14:49,  3.03it/s]
 53%|█████▎    | 3068/5758 [17:27<15:17,  2.93it/s]
                                                   

 53%|█████▎    | 3068/5758 [17:27<15:17,  2.93it/s]
 53%|█████▎    | 3069/5758 [17:27<15:19,  2.92it/s]
                                                   

 53%|█████▎    | 3069/5758 [17:27<15:19,  2.92it/s]
 53%|█████▎    | 3070/5758 [17:28<15:55,  2.81it/s]
                                                   

 53%|█████▎    | 3070/5758 [17:28<15:55,  2.81it/s]
 53%|█████▎    | 3071/5758 [17:28<15:09,  2.95it/s]
                                                   

 53%|█████▎    | 3071/5758 [17:28<15:09,  2.95it/s]
 53%|█████▎    | 3072/5758 [17:28<15:22,  2.91it/s]
                                                   

 53%|█████▎    | 3072/5758 [17:28<15:22,  2.91it/s]
 53%|█████▎    | 3073/5758 [17:29<15:16,  2.93it/s]
                                                   

 53%|█████▎    | 3073/5758 [17:29<15:16,  2.93it/s]
 53%|█████▎    | 3074/5758 [17:29<14:48,  3.02it/s]
                                                   

 53%|█████▎    | 3074/5758 [17:29<14:48,  3.02it/s]
 53%|█████▎    | 3075/5758 [17:29<14:24,  3.10it/s]
                                                   

 53%|█████▎    | 3075/5758 [17:29<14:24,  3.10it/s]
 53%|█████▎    | 3076/5758 [17:29<14:12,  3.15it/s]
                                                   

 53%|█████▎    | 3076/5758 [17:29<14:12,  3.15it/s]
 53%|█████▎    | 3077/5758 [17:30<14:06,  3.17it/s]
                                                   

 53%|█████▎    | 3077/5758 [17:30<14:06,  3.17it/s]
 53%|█████▎    | 3078/5758 [17:30<16:37,  2.69it/s]
                                                   

 53%|█████▎    | 3078/5758 [17:30<16:37,  2.69it/s]
 53%|█████▎    | 3079/5758 [17:31<16:06,  2.77it/s]
                                                   

 53%|█████▎    | 3079/5758 [17:31<16:06,  2.77it/s]
 53%|█████▎    | 3080/5758 [17:31<15:48,  2.82it/s]
                                                   

 53%|█████▎    | 3080/5758 [17:31<15:48,  2.82it/s]
 54%|█████▎    | 3081/5758 [17:31<15:11,  2.94it/s]
                                                   

 54%|█████▎    | 3081/5758 [17:31<15:11,  2.94it/s]
 54%|█████▎    | 3082/5758 [17:32<15:43,  2.84it/s]
                                                   

 54%|█████▎    | 3082/5758 [17:32<15:43,  2.84it/s]
 54%|█████▎    | 3083/5758 [17:32<15:01,  2.97it/s]
                                                   

 54%|█████▎    | 3083/5758 [17:32<15:01,  2.97it/s]
 54%|█████▎    | 3084/5758 [17:32<15:37,  2.85it/s]
                                                   

 54%|█████▎    | 3084/5758 [17:32<15:37,  2.85it/s]
 54%|█████▎    | 3085/5758 [17:33<15:47,  2.82it/s]
                                                   

 54%|█████▎    | 3085/5758 [17:33<15:47,  2.82it/s]
 54%|█████▎    | 3086/5758 [17:33<15:58,  2.79it/s]
                                                   

 54%|█████▎    | 3086/5758 [17:33<15:58,  2.79it/s]
 54%|█████▎    | 3087/5758 [17:33<15:44,  2.83it/s]
                                                   

 54%|█████▎    | 3087/5758 [17:33<15:44,  2.83it/s]
 54%|█████▎    | 3088/5758 [17:34<15:26,  2.88it/s]
                                                   

 54%|█████▎    | 3088/5758 [17:34<15:26,  2.88it/s]
 54%|█████▎    | 3089/5758 [17:34<14:59,  2.97it/s]
                                                   

 54%|█████▎    | 3089/5758 [17:34<14:59,  2.97it/s]
 54%|█████▎    | 3090/5758 [17:34<14:57,  2.97it/s]
                                                   

 54%|█████▎    | 3090/5758 [17:34<14:57,  2.97it/s]
 54%|█████▎    | 3091/5758 [17:35<14:43,  3.02it/s]
                                                   

 54%|█████▎    | 3091/5758 [17:35<14:43,  3.02it/s]
 54%|█████▎    | 3092/5758 [17:35<14:28,  3.07it/s]
                                                   

 54%|█████▎    | 3092/5758 [17:35<14:28,  3.07it/s]
 54%|█████▎    | 3093/5758 [17:35<15:53,  2.79it/s]
                                                   

 54%|█████▎    | 3093/5758 [17:35<15:53,  2.79it/s]
 54%|█████▎    | 3094/5758 [17:36<16:18,  2.72it/s]
                                                   

 54%|█████▎    | 3094/5758 [17:36<16:18,  2.72it/s]
 54%|█████▍    | 3095/5758 [17:36<15:22,  2.89it/s]
                                                   

 54%|█████▍    | 3095/5758 [17:36<15:22,  2.89it/s]
 54%|█████▍    | 3096/5758 [17:36<15:08,  2.93it/s]
                                                   

 54%|█████▍    | 3096/5758 [17:36<15:08,  2.93it/s]
 54%|█████▍    | 3097/5758 [17:37<14:47,  3.00it/s]
                                                   

 54%|█████▍    | 3097/5758 [17:37<14:47,  3.00it/s]
 54%|█████▍    | 3098/5758 [17:37<14:21,  3.09it/s]
                                                   

 54%|█████▍    | 3098/5758 [17:37<14:21,  3.09it/s]
 54%|█████▍    | 3099/5758 [17:37<14:26,  3.07it/s]
                                                   

 54%|█████▍    | 3099/5758 [17:37<14:26,  3.07it/s]
 54%|█████▍    | 3100/5758 [17:38<14:06,  3.14it/s]
                                                   

 54%|█████▍    | 3100/5758 [17:38<14:06,  3.14it/s]
 54%|█████▍    | 3101/5758 [17:38<13:53,  3.19it/s]
                                                   

 54%|█████▍    | 3101/5758 [17:38<13:53,  3.19it/s]
 54%|█████▍    | 3102/5758 [17:38<13:43,  3.22it/s]
                                                   

 54%|█████▍    | 3102/5758 [17:38<13:43,  3.22it/s]
 54%|█████▍    | 3103/5758 [17:39<14:12,  3.11it/s]
                                                   

 54%|█████▍    | 3103/5758 [17:39<14:12,  3.11it/s]
 54%|█████▍    | 3104/5758 [17:39<15:53,  2.78it/s]
                                                   

 54%|█████▍    | 3104/5758 [17:39<15:53,  2.78it/s]
 54%|█████▍    | 3105/5758 [17:39<15:51,  2.79it/s]
                                                   

 54%|█████▍    | 3105/5758 [17:39<15:51,  2.79it/s]
 54%|█████▍    | 3106/5758 [17:40<15:16,  2.89it/s]
                                                   

 54%|█████▍    | 3106/5758 [17:40<15:16,  2.89it/s]
 54%|█████▍    | 3107/5758 [17:40<14:57,  2.95it/s]
                                                   

 54%|█████▍    | 3107/5758 [17:40<14:57,  2.95it/s]
 54%|█████▍    | 3108/5758 [17:40<14:37,  3.02it/s]
                                                   

 54%|█████▍    | 3108/5758 [17:40<14:37,  3.02it/s]
 54%|█████▍    | 3109/5758 [17:41<14:12,  3.11it/s]
                                                   

 54%|█████▍    | 3109/5758 [17:41<14:12,  3.11it/s]
 54%|█████▍    | 3110/5758 [17:41<13:56,  3.17it/s]
                                                   

 54%|█████▍    | 3110/5758 [17:41<13:56,  3.17it/s]
 54%|█████▍    | 3111/5758 [17:41<13:40,  3.23it/s]
                                                   
{'loss': '1.377', 'grad_norm': '24.38', 'learning_rate': '4.694e-06', 'num_tokens': '2.043e+06', 'mean_token_accuracy': '0.6906', 'epoch': '0.5307'}
+{'loss': '1.245', 'grad_norm': '45.5', 'learning_rate': '4.693e-06', 'num_tokens': '2.043e+06', 'mean_token_accuracy': '0.7403', 'epoch': '0.5309'}
+{'loss': '1.018', 'grad_norm': '13.25', 'learning_rate': '4.691e-06', 'num_tokens': '2.044e+06', 'mean_token_accuracy': '0.7163', 'epoch': '0.5311'}
+{'loss': '1.437', 'grad_norm': '12.88', 'learning_rate': '4.689e-06', 'num_tokens': '2.044e+06', 'mean_token_accuracy': '0.6294', 'epoch': '0.5313'}
+{'loss': '1.514', 'grad_norm': '10.88', 'learning_rate': '4.687e-06', 'num_tokens': '2.046e+06', 'mean_token_accuracy': '0.5792', 'epoch': '0.5314'}
+{'loss': '1.237', 'grad_norm': '13.62', 'learning_rate': '4.686e-06', 'num_tokens': '2.046e+06', 'mean_token_accuracy': '0.6879', 'epoch': '0.5316'}
+{'loss': '2.132', 'grad_norm': '22.12', 'learning_rate': '4.684e-06', 'num_tokens': '2.046e+06', 'mean_token_accuracy': '0.6043', 'epoch': '0.5318'}
+{'loss': '0.5066', 'grad_norm': '19.25', 'learning_rate': '4.682e-06', 'num_tokens': '2.047e+06', 'mean_token_accuracy': '0.8659', 'epoch': '0.532'}
+{'loss': '0.958', 'grad_norm': '6.656', 'learning_rate': '4.68e-06', 'num_tokens': '2.049e+06', 'mean_token_accuracy': '0.7528', 'epoch': '0.5321'}
+{'loss': '1.607', 'grad_norm': '43.5', 'learning_rate': '4.679e-06', 'num_tokens': '2.049e+06', 'mean_token_accuracy': '0.7381', 'epoch': '0.5323'}
+{'loss': '1.379', 'grad_norm': '33.5', 'learning_rate': '4.677e-06', 'num_tokens': '2.049e+06', 'mean_token_accuracy': '0.7056', 'epoch': '0.5325'}
+{'loss': '1.11', 'grad_norm': '12', 'learning_rate': '4.675e-06', 'num_tokens': '2.05e+06', 'mean_token_accuracy': '0.6875', 'epoch': '0.5327'}
+{'loss': '1.287', 'grad_norm': '11.31', 'learning_rate': '4.673e-06', 'num_tokens': '2.051e+06', 'mean_token_accuracy': '0.6595', 'epoch': '0.5328'}
+{'loss': '2.099', 'grad_norm': '23.25', 'learning_rate': '4.672e-06', 'num_tokens': '2.051e+06', 'mean_token_accuracy': '0.6044', 'epoch': '0.533'}
+{'loss': '0.8861', 'grad_norm': '12.94', 'learning_rate': '4.67e-06', 'num_tokens': '2.051e+06', 'mean_token_accuracy': '0.7975', 'epoch': '0.5332'}
+{'loss': '1.34', 'grad_norm': '13', 'learning_rate': '4.668e-06', 'num_tokens': '2.052e+06', 'mean_token_accuracy': '0.6636', 'epoch': '0.5333'}
+{'loss': '0.391', 'grad_norm': '18.38', 'learning_rate': '4.667e-06', 'num_tokens': '2.052e+06', 'mean_token_accuracy': '0.8992', 'epoch': '0.5335'}
+{'loss': '1.186', 'grad_norm': '11.31', 'learning_rate': '4.665e-06', 'num_tokens': '2.053e+06', 'mean_token_accuracy': '0.6601', 'epoch': '0.5337'}
+{'loss': '1.154', 'grad_norm': '13', 'learning_rate': '4.663e-06', 'num_tokens': '2.054e+06', 'mean_token_accuracy': '0.686', 'epoch': '0.5339'}
+{'loss': '1.144', 'grad_norm': '13.12', 'learning_rate': '4.661e-06', 'num_tokens': '2.054e+06', 'mean_token_accuracy': '0.6997', 'epoch': '0.534'}
+{'loss': '0.9283', 'grad_norm': '23.12', 'learning_rate': '4.66e-06', 'num_tokens': '2.055e+06', 'mean_token_accuracy': '0.815', 'epoch': '0.5342'}
+{'loss': '0.4679', 'grad_norm': '46.75', 'learning_rate': '4.658e-06', 'num_tokens': '2.055e+06', 'mean_token_accuracy': '0.8642', 'epoch': '0.5344'}
+{'loss': '1.397', 'grad_norm': '25.12', 'learning_rate': '4.656e-06', 'num_tokens': '2.055e+06', 'mean_token_accuracy': '0.6643', 'epoch': '0.5346'}
+{'loss': '1.327', 'grad_norm': '11.12', 'learning_rate': '4.654e-06', 'num_tokens': '2.056e+06', 'mean_token_accuracy': '0.7013', 'epoch': '0.5347'}
+{'loss': '0.5348', 'grad_norm': '5.75', 'learning_rate': '4.653e-06', 'num_tokens': '2.058e+06', 'mean_token_accuracy': '0.8461', 'epoch': '0.5349'}
+{'loss': '1.217', 'grad_norm': '12', 'learning_rate': '4.651e-06', 'num_tokens': '2.058e+06', 'mean_token_accuracy': '0.693', 'epoch': '0.5351'}
+{'loss': '1.335', 'grad_norm': '11.19', 'learning_rate': '4.649e-06', 'num_tokens': '2.059e+06', 'mean_token_accuracy': '0.6438', 'epoch': '0.5353'}
+{'loss': '0.9562', 'grad_norm': '12.06', 'learning_rate': '4.647e-06', 'num_tokens': '2.06e+06', 'mean_token_accuracy': '0.7091', 'epoch': '0.5354'}
+{'loss': '1.87', 'grad_norm': '17.5', 'learning_rate': '4.646e-06', 'num_tokens': '2.06e+06', 'mean_token_accuracy': '0.5993', 'epoch': '0.5356'}
+{'loss': '1.248', 'grad_norm': '30.88', 'learning_rate': '4.644e-06', 'num_tokens': '2.061e+06', 'mean_token_accuracy': '0.7', 'epoch': '0.5358'}
+{'loss': '1.265', 'grad_norm': '5.094', 'learning_rate': '4.642e-06', 'num_tokens': '2.065e+06', 'mean_token_accuracy': '0.7126', 'epoch': '0.5359'}
+{'loss': '1.219', 'grad_norm': '13.06', 'learning_rate': '4.641e-06', 'num_tokens': '2.066e+06', 'mean_token_accuracy': '0.6606', 'epoch': '0.5361'}
+{'loss': '1.545', 'grad_norm': '13.81', 'learning_rate': '4.639e-06', 'num_tokens': '2.066e+06', 'mean_token_accuracy': '0.6069', 'epoch': '0.5363'}
+{'loss': '0.5078', 'grad_norm': '11.56', 'learning_rate': '4.637e-06', 'num_tokens': '2.067e+06', 'mean_token_accuracy': '0.8561', 'epoch': '0.5365'}
+{'loss': '1.554', 'grad_norm': '12.44', 'learning_rate': '4.635e-06', 'num_tokens': '2.068e+06', 'mean_token_accuracy': '0.5895', 'epoch': '0.5366'}
+{'loss': '0.5399', 'grad_norm': '7.062', 'learning_rate': '4.634e-06', 'num_tokens': '2.07e+06', 'mean_token_accuracy': '0.8354', 'epoch': '0.5368'}
+{'loss': '0.603', 'grad_norm': '14.62', 'learning_rate': '4.632e-06', 'num_tokens': '2.07e+06', 'mean_token_accuracy': '0.8659', 'epoch': '0.537'}
+{'loss': '1.891', 'grad_norm': '34', 'learning_rate': '4.63e-06', 'num_tokens': '2.07e+06', 'mean_token_accuracy': '0.6216', 'epoch': '0.5372'}
+{'loss': '1.624', 'grad_norm': '14.81', 'learning_rate': '4.628e-06', 'num_tokens': '2.071e+06', 'mean_token_accuracy': '0.6454', 'epoch': '0.5373'}
+{'loss': '1.162', 'grad_norm': '13.44', 'learning_rate': '4.627e-06', 'num_tokens': '2.071e+06', 'mean_token_accuracy': '0.663', 'epoch': '0.5375'}
+{'loss': '0.8584', 'grad_norm': '11.62', 'learning_rate': '4.625e-06', 'num_tokens': '2.072e+06', 'mean_token_accuracy': '0.7334', 'epoch': '0.5377'}
+{'loss': '0.6828', 'grad_norm': '12.06', 'learning_rate': '4.623e-06', 'num_tokens': '2.073e+06', 'mean_token_accuracy': '0.8093', 'epoch': '0.5379'}
+{'loss': '1.083', 'grad_norm': '13.75', 'learning_rate': '4.621e-06', 'num_tokens': '2.073e+06', 'mean_token_accuracy': '0.6802', 'epoch': '0.538'}
+{'loss': '1.345', 'grad_norm': '22', 'learning_rate': '4.62e-06', 'num_tokens': '2.074e+06', 'mean_token_accuracy': '0.7056', 'epoch': '0.5382'}
+{'loss': '1.511', 'grad_norm': '22.38', 'learning_rate': '4.618e-06', 'num_tokens': '2.074e+06', 'mean_token_accuracy': '0.6725', 'epoch': '0.5384'}
+{'loss': '0.66', 'grad_norm': '14.12', 'learning_rate': '4.616e-06', 'num_tokens': '2.074e+06', 'mean_token_accuracy': '0.8357', 'epoch': '0.5386'}
+{'loss': '1.05', 'grad_norm': '12.31', 'learning_rate': '4.614e-06', 'num_tokens': '2.075e+06', 'mean_token_accuracy': '0.7212', 'epoch': '0.5387'}
+{'loss': '0.6752', 'grad_norm': '13.75', 'learning_rate': '4.613e-06', 'num_tokens': '2.075e+06', 'mean_token_accuracy': '0.8195', 'epoch': '0.5389'}
+{'loss': '0.9096', 'grad_norm': '6.688', 'learning_rate': '4.611e-06', 'num_tokens': '2.077e+06', 'mean_token_accuracy': '0.7701', 'epoch': '0.5391'}
+{'loss': '1.972', 'grad_norm': '19.12', 'learning_rate': '4.609e-06', 'num_tokens': '2.078e+06', 'mean_token_accuracy': '0.5984', 'epoch': '0.5392'}
+{'loss': '1.117', 'grad_norm': '38.25', 'learning_rate': '4.608e-06', 'num_tokens': '2.078e+06', 'mean_token_accuracy': '0.7701', 'epoch': '0.5394'}
+{'loss': '0.8525', 'grad_norm': '38.25', 'learning_rate': '4.606e-06', 'num_tokens': '2.078e+06', 'mean_token_accuracy': '0.828', 'epoch': '0.5396'}
+{'loss': '1.256', 'grad_norm': '13.31', 'learning_rate': '4.604e-06', 'num_tokens': '2.079e+06', 'mean_token_accuracy': '0.6779', 'epoch': '0.5398'}
+{'loss': '1.137', 'grad_norm': '13.5', 'learning_rate': '4.602e-06', 'num_tokens': '2.079e+06', 'mean_token_accuracy': '0.6772', 'epoch': '0.5399'}
+{'loss': '0.9694', 'grad_norm': '22.38', 'learning_rate': '4.601e-06', 'num_tokens': '2.079e+06', 'mean_token_accuracy': '0.8103', 'epoch': '0.5401'}
+
 54%|█████▍    | 3111/5758 [17:41<13:40,  3.23it/s]
 54%|█████▍    | 3112/5758 [17:42<14:46,  2.99it/s]
                                                   

 54%|█████▍    | 3112/5758 [17:42<14:46,  2.99it/s]
 54%|█████▍    | 3113/5758 [17:42<14:30,  3.04it/s]
                                                   

 54%|█████▍    | 3113/5758 [17:42<14:30,  3.04it/s]
 54%|█████▍    | 3114/5758 [17:42<14:35,  3.02it/s]
                                                   

 54%|█████▍    | 3114/5758 [17:42<14:35,  3.02it/s]
 54%|█████▍    | 3115/5758 [17:43<14:09,  3.11it/s]
                                                   

 54%|█████▍    | 3115/5758 [17:43<14:09,  3.11it/s]
 54%|█████▍    | 3116/5758 [17:43<14:06,  3.12it/s]
                                                   

 54%|█████▍    | 3116/5758 [17:43<14:06,  3.12it/s]
 54%|█████▍    | 3117/5758 [17:43<14:54,  2.95it/s]
                                                   

 54%|█████▍    | 3117/5758 [17:43<14:54,  2.95it/s]
 54%|█████▍    | 3118/5758 [17:44<15:16,  2.88it/s]
                                                   

 54%|█████▍    | 3118/5758 [17:44<15:16,  2.88it/s]
 54%|█████▍    | 3119/5758 [17:44<15:18,  2.87it/s]
                                                   

 54%|█████▍    | 3119/5758 [17:44<15:18,  2.87it/s]
 54%|█████▍    | 3120/5758 [17:44<14:50,  2.96it/s]
                                                   

 54%|█████▍    | 3120/5758 [17:44<14:50,  2.96it/s]
 54%|█████▍    | 3121/5758 [17:45<14:47,  2.97it/s]
                                                   

 54%|█████▍    | 3121/5758 [17:45<14:47,  2.97it/s]
 54%|█████▍    | 3122/5758 [17:45<15:24,  2.85it/s]
                                                   

 54%|█████▍    | 3122/5758 [17:45<15:24,  2.85it/s]
 54%|█████▍    | 3123/5758 [17:45<15:14,  2.88it/s]
                                                   

 54%|█████▍    | 3123/5758 [17:45<15:14,  2.88it/s]
 54%|█████▍    | 3124/5758 [17:46<14:44,  2.98it/s]
                                                   

 54%|█████▍    | 3124/5758 [17:46<14:44,  2.98it/s]
 54%|█████▍    | 3125/5758 [17:46<15:35,  2.81it/s]
                                                   

 54%|█████▍    | 3125/5758 [17:46<15:35,  2.81it/s]
 54%|█████▍    | 3126/5758 [17:46<15:10,  2.89it/s]
                                                   

 54%|█████▍    | 3126/5758 [17:46<15:10,  2.89it/s]
 54%|█████▍    | 3127/5758 [17:47<14:23,  3.05it/s]
                                                   

 54%|█████▍    | 3127/5758 [17:47<14:23,  3.05it/s]
 54%|█████▍    | 3128/5758 [17:47<14:14,  3.08it/s]
                                                   

 54%|█████▍    | 3128/5758 [17:47<14:14,  3.08it/s]
 54%|█████▍    | 3129/5758 [17:47<15:00,  2.92it/s]
                                                   

 54%|█████▍    | 3129/5758 [17:47<15:00,  2.92it/s]
 54%|█████▍    | 3130/5758 [17:48<16:41,  2.62it/s]
                                                   

 54%|█████▍    | 3130/5758 [17:48<16:41,  2.62it/s]
 54%|█████▍    | 3131/5758 [17:48<15:40,  2.79it/s]
                                                   

 54%|█████▍    | 3131/5758 [17:48<15:40,  2.79it/s]
 54%|█████▍    | 3132/5758 [17:48<14:52,  2.94it/s]
                                                   

 54%|█████▍    | 3132/5758 [17:48<14:52,  2.94it/s]
 54%|█████▍    | 3133/5758 [17:49<14:25,  3.03it/s]
                                                   

 54%|█████▍    | 3133/5758 [17:49<14:25,  3.03it/s]
 54%|█████▍    | 3134/5758 [17:49<14:59,  2.92it/s]
                                                   

 54%|█████▍    | 3134/5758 [17:49<14:59,  2.92it/s]
 54%|█████▍    | 3135/5758 [17:50<15:36,  2.80it/s]
                                                   

 54%|█████▍    | 3135/5758 [17:50<15:36,  2.80it/s]
 54%|█████▍    | 3136/5758 [17:50<15:18,  2.86it/s]
                                                   

 54%|█████▍    | 3136/5758 [17:50<15:18,  2.86it/s]
 54%|█████▍    | 3137/5758 [17:50<14:41,  2.97it/s]
                                                   

 54%|█████▍    | 3137/5758 [17:50<14:41,  2.97it/s]
 54%|█████▍    | 3138/5758 [17:51<15:20,  2.85it/s]
                                                   

 54%|█████▍    | 3138/5758 [17:51<15:20,  2.85it/s]
 55%|█████▍    | 3139/5758 [17:51<15:17,  2.85it/s]
                                                   

 55%|█████▍    | 3139/5758 [17:51<15:17,  2.85it/s]
 55%|█████▍    | 3140/5758 [17:51<15:17,  2.85it/s]
                                                   

 55%|█████▍    | 3140/5758 [17:51<15:17,  2.85it/s]
 55%|█████▍    | 3141/5758 [17:52<15:08,  2.88it/s]
                                                   

 55%|█████▍    | 3141/5758 [17:52<15:08,  2.88it/s]
 55%|█████▍    | 3142/5758 [17:52<15:01,  2.90it/s]
                                                   

 55%|█████▍    | 3142/5758 [17:52<15:01,  2.90it/s]
 55%|█████▍    | 3143/5758 [17:52<14:33,  2.99it/s]
                                                   

 55%|█████▍    | 3143/5758 [17:52<14:33,  2.99it/s]
 55%|█████▍    | 3144/5758 [17:53<14:54,  2.92it/s]
                                                   

 55%|█████▍    | 3144/5758 [17:53<14:54,  2.92it/s]
 55%|█████▍    | 3145/5758 [17:53<15:05,  2.88it/s]
                                                   

 55%|█████▍    | 3145/5758 [17:53<15:05,  2.88it/s]
 55%|█████▍    | 3146/5758 [17:53<14:05,  3.09it/s]
                                                   

 55%|█████▍    | 3146/5758 [17:53<14:05,  3.09it/s]
 55%|███���█▍    | 3147/5758 [17:54<14:37,  2.97it/s]
                                                   

 55%|█████▍    | 3147/5758 [17:54<14:37,  2.97it/s]
 55%|█████▍    | 3148/5758 [17:54<14:33,  2.99it/s]
                                                   

 55%|█████▍    | 3148/5758 [17:54<14:33,  2.99it/s]
 55%|█████▍    | 3149/5758 [17:54<14:21,  3.03it/s]
                                                   

 55%|█████▍    | 3149/5758 [17:54<14:21,  3.03it/s]
 55%|█████▍    | 3150/5758 [17:55<13:52,  3.13it/s]
                                                   

 55%|█████▍    | 3150/5758 [17:55<13:52,  3.13it/s]
 55%|█████▍    | 3151/5758 [17:55<13:34,  3.20it/s]
                                                   

 55%|█████▍    | 3151/5758 [17:55<13:34,  3.20it/s]
 55%|█████▍    | 3152/5758 [17:55<13:14,  3.28it/s]
                                                   

 55%|█████▍    | 3152/5758 [17:55<13:14,  3.28it/s]
 55%|█████▍    | 3153/5758 [17:55<13:11,  3.29it/s]
                                                   

 55%|█████▍    | 3153/5758 [17:55<13:11,  3.29it/s]
 55%|█████▍    | 3154/5758 [17:56<13:00,  3.33it/s]
                                                   

 55%|█████▍    | 3154/5758 [17:56<13:00,  3.33it/s]
 55%|█████▍    | 3155/5758 [17:56<13:50,  3.13it/s]
                                                   

 55%|█████▍    | 3155/5758 [17:56<13:50,  3.13it/s]
 55%|█████▍    | 3156/5758 [17:57<15:03,  2.88it/s]
                                                   

 55%|█████▍    | 3156/5758 [17:57<15:03,  2.88it/s]
 55%|█████▍    | 3157/5758 [17:57<14:53,  2.91it/s]
                                                   

 55%|█████▍    | 3157/5758 [17:57<14:53,  2.91it/s]
 55%|█████▍    | 3158/5758 [17:57<14:31,  2.98it/s]
                                                   

 55%|█████▍    | 3158/5758 [17:57<14:31,  2.98it/s]
 55%|█████▍    | 3159/5758 [17:58<14:30,  2.99it/s]
                                                   

 55%|█████▍    | 3159/5758 [17:58<14:30,  2.99it/s]
 55%|█████▍    | 3160/5758 [17:58<14:35,  2.97it/s]
                                                   

 55%|█████▍    | 3160/5758 [17:58<14:35,  2.97it/s]
 55%|█████▍    | 3161/5758 [17:58<14:43,  2.94it/s]
                                                   

 55%|█████▍    | 3161/5758 [17:58<14:43,  2.94it/s]
 55%|█████▍    | 3162/5758 [17:59<14:29,  2.98it/s]
                                                   

 55%|█████▍    | 3162/5758 [17:59<14:29,  2.98it/s]
 55%|█████▍    | 3163/5758 [17:59<14:35,  2.96it/s]
                                                   

 55%|█████▍    | 3163/5758 [17:59<14:35,  2.96it/s]
 55%|█████▍    | 3164/5758 [17:59<14:10,  3.05it/s]
                                                   

 55%|█████▍    | 3164/5758 [17:59<14:10,  3.05it/s]
 55%|█████▍    | 3165/5758 [17:59<14:07,  3.06it/s]
                                                   

 55%|█████▍    | 3165/5758 [17:59<14:07,  3.06it/s]
 55%|█████▍    | 3166/5758 [18:00<13:51,  3.12it/s]
                                                   
{'loss': '1.284', 'grad_norm': '18.12', 'learning_rate': '4.599e-06', 'num_tokens': '2.08e+06', 'mean_token_accuracy': '0.7493', 'epoch': '0.5403'}
+{'loss': '1.287', 'grad_norm': '9.375', 'learning_rate': '4.597e-06', 'num_tokens': '2.081e+06', 'mean_token_accuracy': '0.7154', 'epoch': '0.5405'}
+{'loss': '1.358', 'grad_norm': '12.88', 'learning_rate': '4.595e-06', 'num_tokens': '2.082e+06', 'mean_token_accuracy': '0.6359', 'epoch': '0.5406'}
+{'loss': '0.7608', 'grad_norm': '10.69', 'learning_rate': '4.594e-06', 'num_tokens': '2.083e+06', 'mean_token_accuracy': '0.7692', 'epoch': '0.5408'}
+{'loss': '1.723', 'grad_norm': '11.56', 'learning_rate': '4.592e-06', 'num_tokens': '2.083e+06', 'mean_token_accuracy': '0.6295', 'epoch': '0.541'}
+{'loss': '1.294', 'grad_norm': '12.56', 'learning_rate': '4.59e-06', 'num_tokens': '2.084e+06', 'mean_token_accuracy': '0.6588', 'epoch': '0.5412'}
+{'loss': '1.34', 'grad_norm': '10.56', 'learning_rate': '4.588e-06', 'num_tokens': '2.085e+06', 'mean_token_accuracy': '0.6342', 'epoch': '0.5413'}
+{'loss': '1.553', 'grad_norm': '13.56', 'learning_rate': '4.587e-06', 'num_tokens': '2.086e+06', 'mean_token_accuracy': '0.6717', 'epoch': '0.5415'}
+{'loss': '1.968', 'grad_norm': '20.25', 'learning_rate': '4.585e-06', 'num_tokens': '2.086e+06', 'mean_token_accuracy': '0.6088', 'epoch': '0.5417'}
+{'loss': '1.084', 'grad_norm': '7.219', 'learning_rate': '4.583e-06', 'num_tokens': '2.088e+06', 'mean_token_accuracy': '0.7551', 'epoch': '0.5419'}
+{'loss': '2.292', 'grad_norm': '11.12', 'learning_rate': '4.581e-06', 'num_tokens': '2.089e+06', 'mean_token_accuracy': '0.5418', 'epoch': '0.542'}
+{'loss': '1.31', 'grad_norm': '14.19', 'learning_rate': '4.58e-06', 'num_tokens': '2.089e+06', 'mean_token_accuracy': '0.6447', 'epoch': '0.5422'}
+{'loss': '0.7269', 'grad_norm': '14.38', 'learning_rate': '4.578e-06', 'num_tokens': '2.09e+06', 'mean_token_accuracy': '0.8032', 'epoch': '0.5424'}
+{'loss': '0.8549', 'grad_norm': '22.5', 'learning_rate': '4.576e-06', 'num_tokens': '2.09e+06', 'mean_token_accuracy': '0.7838', 'epoch': '0.5425'}
+{'loss': '0.8225', 'grad_norm': '28.25', 'learning_rate': '4.575e-06', 'num_tokens': '2.09e+06', 'mean_token_accuracy': '0.7958', 'epoch': '0.5427'}
+{'loss': '1.18', 'grad_norm': '7.594', 'learning_rate': '4.573e-06', 'num_tokens': '2.091e+06', 'mean_token_accuracy': '0.7138', 'epoch': '0.5429'}
+{'loss': '0.5439', 'grad_norm': '8.188', 'learning_rate': '4.571e-06', 'num_tokens': '2.093e+06', 'mean_token_accuracy': '0.8276', 'epoch': '0.5431'}
+{'loss': '1.198', 'grad_norm': '7.344', 'learning_rate': '4.569e-06', 'num_tokens': '2.094e+06', 'mean_token_accuracy': '0.7199', 'epoch': '0.5432'}
+{'loss': '1.037', 'grad_norm': '12.62', 'learning_rate': '4.568e-06', 'num_tokens': '2.095e+06', 'mean_token_accuracy': '0.716', 'epoch': '0.5434'}
+{'loss': '1.215', 'grad_norm': '21.12', 'learning_rate': '4.566e-06', 'num_tokens': '2.095e+06', 'mean_token_accuracy': '0.7377', 'epoch': '0.5436'}
+{'loss': '1.806', 'grad_norm': '17.38', 'learning_rate': '4.564e-06', 'num_tokens': '2.096e+06', 'mean_token_accuracy': '0.6314', 'epoch': '0.5438'}
+{'loss': '1.018', 'grad_norm': '13.69', 'learning_rate': '4.562e-06', 'num_tokens': '2.096e+06', 'mean_token_accuracy': '0.73', 'epoch': '0.5439'}
+{'loss': '1.061', 'grad_norm': '12', 'learning_rate': '4.561e-06', 'num_tokens': '2.097e+06', 'mean_token_accuracy': '0.7166', 'epoch': '0.5441'}
+{'loss': '2.1', 'grad_norm': '19.75', 'learning_rate': '4.559e-06', 'num_tokens': '2.097e+06', 'mean_token_accuracy': '0.569', 'epoch': '0.5443'}
+{'loss': '1.131', 'grad_norm': '13.75', 'learning_rate': '4.557e-06', 'num_tokens': '2.098e+06', 'mean_token_accuracy': '0.6827', 'epoch': '0.5445'}
+{'loss': '0.7994', 'grad_norm': '14.19', 'learning_rate': '4.555e-06', 'num_tokens': '2.098e+06', 'mean_token_accuracy': '0.8239', 'epoch': '0.5446'}
+{'loss': '1.706', 'grad_norm': '19', 'learning_rate': '4.554e-06', 'num_tokens': '2.099e+06', 'mean_token_accuracy': '0.6284', 'epoch': '0.5448'}
+{'loss': '0.9017', 'grad_norm': '17.5', 'learning_rate': '4.552e-06', 'num_tokens': '2.099e+06', 'mean_token_accuracy': '0.7966', 'epoch': '0.545'}
+{'loss': '1.335', 'grad_norm': '10.75', 'learning_rate': '4.55e-06', 'num_tokens': '2.1e+06', 'mean_token_accuracy': '0.6581', 'epoch': '0.5452'}
+{'loss': '1.489', 'grad_norm': '35', 'learning_rate': '4.548e-06', 'num_tokens': '2.1e+06', 'mean_token_accuracy': '0.6916', 'epoch': '0.5453'}
+{'loss': '1.319', 'grad_norm': '21.5', 'learning_rate': '4.547e-06', 'num_tokens': '2.1e+06', 'mean_token_accuracy': '0.7265', 'epoch': '0.5455'}
+{'loss': '2.275', 'grad_norm': '15.06', 'learning_rate': '4.545e-06', 'num_tokens': '2.101e+06', 'mean_token_accuracy': '0.5235', 'epoch': '0.5457'}
+{'loss': '0.5217', 'grad_norm': '7.562', 'learning_rate': '4.543e-06', 'num_tokens': '2.102e+06', 'mean_token_accuracy': '0.8358', 'epoch': '0.5458'}
+{'loss': '1.409', 'grad_norm': '5', 'learning_rate': '4.542e-06', 'num_tokens': '2.106e+06', 'mean_token_accuracy': '0.6681', 'epoch': '0.546'}
+{'loss': '1.484', 'grad_norm': '19', 'learning_rate': '4.54e-06', 'num_tokens': '2.107e+06', 'mean_token_accuracy': '0.6474', 'epoch': '0.5462'}
+{'loss': '1.131', 'grad_norm': '40.75', 'learning_rate': '4.538e-06', 'num_tokens': '2.107e+06', 'mean_token_accuracy': '0.8571', 'epoch': '0.5464'}
+{'loss': '1.313', 'grad_norm': '13.94', 'learning_rate': '4.536e-06', 'num_tokens': '2.107e+06', 'mean_token_accuracy': '0.7239', 'epoch': '0.5465'}
+{'loss': '2.236', 'grad_norm': '24.5', 'learning_rate': '4.535e-06', 'num_tokens': '2.108e+06', 'mean_token_accuracy': '0.5654', 'epoch': '0.5467'}
+{'loss': '1.217', 'grad_norm': '17.12', 'learning_rate': '4.533e-06', 'num_tokens': '2.108e+06', 'mean_token_accuracy': '0.7031', 'epoch': '0.5469'}
+{'loss': '1.055', 'grad_norm': '13.75', 'learning_rate': '4.531e-06', 'num_tokens': '2.109e+06', 'mean_token_accuracy': '0.6986', 'epoch': '0.5471'}
+{'loss': '0.9698', 'grad_norm': '12.62', 'learning_rate': '4.529e-06', 'num_tokens': '2.109e+06', 'mean_token_accuracy': '0.7164', 'epoch': '0.5472'}
+{'loss': '1.127', 'grad_norm': '13.19', 'learning_rate': '4.528e-06', 'num_tokens': '2.11e+06', 'mean_token_accuracy': '0.6937', 'epoch': '0.5474'}
+{'loss': '1.759', 'grad_norm': '11.81', 'learning_rate': '4.526e-06', 'num_tokens': '2.111e+06', 'mean_token_accuracy': '0.6026', 'epoch': '0.5476'}
+{'loss': '1.029', 'grad_norm': '13.38', 'learning_rate': '4.524e-06', 'num_tokens': '2.111e+06', 'mean_token_accuracy': '0.7429', 'epoch': '0.5478'}
+{'loss': '1.089', 'grad_norm': '9.688', 'learning_rate': '4.522e-06', 'num_tokens': '2.112e+06', 'mean_token_accuracy': '0.7333', 'epoch': '0.5479'}
+{'loss': '1.133', 'grad_norm': '14.75', 'learning_rate': '4.521e-06', 'num_tokens': '2.113e+06', 'mean_token_accuracy': '0.6843', 'epoch': '0.5481'}
+{'loss': '0.6847', 'grad_norm': '20.62', 'learning_rate': '4.519e-06', 'num_tokens': '2.113e+06', 'mean_token_accuracy': '0.8317', 'epoch': '0.5483'}
+{'loss': '1.153', 'grad_norm': '11', 'learning_rate': '4.517e-06', 'num_tokens': '2.114e+06', 'mean_token_accuracy': '0.6781', 'epoch': '0.5485'}
+{'loss': '2.17', 'grad_norm': '17.12', 'learning_rate': '4.515e-06', 'num_tokens': '2.114e+06', 'mean_token_accuracy': '0.554', 'epoch': '0.5486'}
+{'loss': '1.076', 'grad_norm': '18.88', 'learning_rate': '4.514e-06', 'num_tokens': '2.115e+06', 'mean_token_accuracy': '0.7456', 'epoch': '0.5488'}
+{'loss': '2.48', 'grad_norm': '11.38', 'learning_rate': '4.512e-06', 'num_tokens': '2.116e+06', 'mean_token_accuracy': '0.5255', 'epoch': '0.549'}
+{'loss': '1.58', 'grad_norm': '18.75', 'learning_rate': '4.51e-06', 'num_tokens': '2.116e+06', 'mean_token_accuracy': '0.6384', 'epoch': '0.5491'}
+{'loss': '1.116', 'grad_norm': '20.75', 'learning_rate': '4.509e-06', 'num_tokens': '2.116e+06', 'mean_token_accuracy': '0.7716', 'epoch': '0.5493'}
+{'loss': '0.9058', 'grad_norm': '11.94', 'learning_rate': '4.507e-06', 'num_tokens': '2.117e+06', 'mean_token_accuracy': '0.7389', 'epoch': '0.5495'}
+{'loss': '1.246', 'grad_norm': '11.38', 'learning_rate': '4.505e-06', 'num_tokens': '2.118e+06', 'mean_token_accuracy': '0.6404', 'epoch': '0.5497'}
+
 55%|█████▍    | 3166/5758 [18:00<13:51,  3.12it/s]
 55%|█████▌    | 3167/5758 [18:00<13:49,  3.12it/s]
                                                   

 55%|█████▌    | 3167/5758 [18:00<13:49,  3.12it/s]
 55%|█████▌    | 3168/5758 [18:00<14:13,  3.03it/s]
                                                   

 55%|█████▌    | 3168/5758 [18:00<14:13,  3.03it/s]
 55%|█████▌    | 3169/5758 [18:01<14:02,  3.07it/s]
                                                   

 55%|█████▌    | 3169/5758 [18:01<14:02,  3.07it/s]
 55%|█████▌    | 3170/5758 [18:01<13:43,  3.14it/s]
                                                   

 55%|█████▌    | 3170/5758 [18:01<13:43,  3.14it/s]
 55%|█████▌    | 3171/5758 [18:01<13:23,  3.22it/s]
                                                   

 55%|█████▌    | 3171/5758 [18:01<13:23,  3.22it/s]
 55%|█████▌    | 3172/5758 [18:02<13:18,  3.24it/s]
                                                   

 55%|█████▌    | 3172/5758 [18:02<13:18,  3.24it/s]
 55%|█████▌    | 3173/5758 [18:02<13:12,  3.26it/s]
                                                   

 55%|█████▌    | 3173/5758 [18:02<13:12,  3.26it/s]
 55%|█████▌    | 3174/5758 [18:02<12:59,  3.32it/s]
                                                   

 55%|█████▌    | 3174/5758 [18:02<12:59,  3.32it/s]
 55%|█████▌    | 3175/5758 [18:03<13:11,  3.26it/s]
                                                   

 55%|█████▌    | 3175/5758 [18:03<13:11,  3.26it/s]
 55%|█████▌    | 3176/5758 [18:03<13:18,  3.23it/s]
                                                   

 55%|█████▌    | 3176/5758 [18:03<13:18,  3.23it/s]
 55%|█████▌    | 3177/5758 [18:03<13:30,  3.18it/s]
                                                   

 55%|█████▌    | 3177/5758 [18:03<13:30,  3.18it/s]
 55%|█████▌    | 3178/5758 [18:04<13:26,  3.20it/s]
                                                   

 55%|█████▌    | 3178/5758 [18:04<13:26,  3.20it/s]
 55%|█████▌    | 3179/5758 [18:04<13:34,  3.17it/s]
                                                   

 55%|█████▌    | 3179/5758 [18:04<13:34,  3.17it/s]
 55%|█████▌    | 3180/5758 [18:04<13:44,  3.13it/s]
                                                   

 55%|█████▌    | 3180/5758 [18:04<13:44,  3.13it/s]
 55%|█████▌    | 3181/5758 [18:05<13:55,  3.08it/s]
                                                   

 55%|█████▌    | 3181/5758 [18:05<13:55,  3.08it/s]
 55%|█████▌    | 3182/5758 [18:05<15:50,  2.71it/s]
                                                   

 55%|█████▌    | 3182/5758 [18:05<15:50,  2.71it/s]
 55%|█████▌    | 3183/5758 [18:05<15:42,  2.73it/s]
                                                   

 55%|█████▌    | 3183/5758 [18:05<15:42,  2.73it/s]
 55%|█████▌    | 3184/5758 [18:06<15:20,  2.80it/s]
                                                   

 55%|█████▌    | 3184/5758 [18:06<15:20,  2.80it/s]
 55%|█████▌    | 3185/5758 [18:06<14:44,  2.91it/s]
                                                   

 55%|█████▌    | 3185/5758 [18:06<14:44,  2.91it/s]
 55%|█████▌    | 3186/5758 [18:06<14:32,  2.95it/s]
                                                   

 55%|█████▌    | 3186/5758 [18:06<14:32,  2.95it/s]
 55%|█████▌    | 3187/5758 [18:07<14:15,  3.00it/s]
                                                   

 55%|█████▌    | 3187/5758 [18:07<14:15,  3.00it/s]
 55%|█████▌    | 3188/5758 [18:07<14:36,  2.93it/s]
                                                   

 55%|█████▌    | 3188/5758 [18:07<14:36,  2.93it/s]
 55%|█████▌    | 3189/5758 [18:07<14:47,  2.89it/s]
                                                   

 55%|█████▌    | 3189/5758 [18:07<14:47,  2.89it/s]
 55%|█████▌    | 3190/5758 [18:08<14:58,  2.86it/s]
                                                   

 55%|█████▌    | 3190/5758 [18:08<14:58,  2.86it/s]
 55%|█████▌    | 3191/5758 [18:08<14:30,  2.95it/s]
                                                   

 55%|█████▌    | 3191/5758 [18:08<14:30,  2.95it/s]
 55%|█████▌    | 3192/5758 [18:08<13:58,  3.06it/s]
                                                   

 55%|█████▌    | 3192/5758 [18:08<13:58,  3.06it/s]
 55%|█████▌    | 3193/5758 [18:09<13:54,  3.07it/s]
                                                   

 55%|█████▌    | 3193/5758 [18:09<13:54,  3.07it/s]
 55%|█████▌    | 3194/5758 [18:09<13:31,  3.16it/s]
                                                   

 55%|█████▌    | 3194/5758 [18:09<13:31,  3.16it/s]
 55%|█████▌    | 3195/5758 [18:09<13:46,  3.10it/s]
                                                   

 55%|█████▌    | 3195/5758 [18:09<13:46,  3.10it/s]
 56%|█████▌    | 3196/5758 [18:10<14:01,  3.04it/s]
                                                   

 56%|█████▌    | 3196/5758 [18:10<14:01,  3.04it/s]
 56%|█████▌    | 3197/5758 [18:10<14:39,  2.91it/s]
                                                   

 56%|█████▌    | 3197/5758 [18:10<14:39,  2.91it/s]
 56%|█████▌    | 3198/5758 [18:10<14:22,  2.97it/s]
                                                   

 56%|█████▌    | 3198/5758 [18:10<14:22,  2.97it/s]
 56%|█████▌    | 3199/5758 [18:11<14:57,  2.85it/s]
                                                   

 56%|█████▌    | 3199/5758 [18:11<14:57,  2.85it/s]
 56%|█████▌    | 3200/5758 [18:11<14:18,  2.98it/s]
                                                   

 56%|█████▌    | 3200/5758 [18:11<14:18,  2.98it/s]
 56%|█████▌    | 3201/5758 [18:11<13:52,  3.07it/s]
                                                   

 56%|█████▌    | 3201/5758 [18:11<13:52,  3.07it/s]
 56%|█████▌    | 3202/5758 [18:12<13:35,  3.13it/s]
                                                   

 56%|█████▌    | 3202/5758 [18:12<13:35,  3.13it/s]
 56%|█████▌    | 3203/5758 [18:12<13:13,  3.22it/s]
                                                   

 56%|█████▌    | 3203/5758 [18:12<13:13,  3.22it/s]
 56%|█████▌    | 3204/5758 [18:12<13:13,  3.22it/s]
                                                   

 56%|█████▌    | 3204/5758 [18:12<13:13,  3.22it/s]
 56%|█████▌    | 3205/5758 [18:13<13:08,  3.24it/s]
                                                   

 56%|█████▌    | 3205/5758 [18:13<13:08,  3.24it/s]
 56%|█████▌    | 3206/5758 [18:13<14:11,  3.00it/s]
                                                   

 56%|█████▌    | 3206/5758 [18:13<14:11,  3.00it/s]
 56%|█████▌    | 3207/5758 [18:13<13:53,  3.06it/s]
                                                   

 56%|█████▌    | 3207/5758 [18:13<13:53,  3.06it/s]
 56%|█████▌    | 3208/5758 [18:14<14:22,  2.96it/s]
                                                   

 56%|█████▌    | 3208/5758 [18:14<14:22,  2.96it/s]
 56%|█████▌    | 3209/5758 [18:14<15:32,  2.73it/s]
                                                   

 56%|█████▌    | 3209/5758 [18:14<15:32,  2.73it/s]
 56%|█████▌    | 3210/5758 [18:14<14:56,  2.84it/s]
                                                   

 56%|█████▌    | 3210/5758 [18:14<14:56,  2.84it/s]
 56%|█████▌    | 3211/5758 [18:15<14:23,  2.95it/s]
                                                   

 56%|█████▌    | 3211/5758 [18:15<14:23,  2.95it/s]
 56%|█████▌    | 3212/5758 [18:15<14:38,  2.90it/s]
                                                   

 56%|█████▌    | 3212/5758 [18:15<14:38,  2.90it/s]
 56%|█████▌    | 3213/5758 [18:15<14:05,  3.01it/s]
                                                   

 56%|█████▌    | 3213/5758 [18:15<14:05,  3.01it/s]
 56%|█████▌    | 3214/5758 [18:16<14:00,  3.03it/s]
                                                   

 56%|█████▌    | 3214/5758 [18:16<14:00,  3.03it/s]
 56%|█████▌    | 3215/5758 [18:16<14:24,  2.94it/s]
                                                   

 56%|█████▌    | 3215/5758 [18:16<14:24,  2.94it/s]
 56%|█████▌    | 3216/5758 [18:16<14:34,  2.91it/s]
                                                   

 56%|█████▌    | 3216/5758 [18:16<14:34,  2.91it/s]
 56%|█████▌    | 3217/5758 [18:17<14:42,  2.88it/s]
                                                   

 56%|█████▌    | 3217/5758 [18:17<14:42,  2.88it/s]
 56%|█████▌    | 3218/5758 [18:17<15:03,  2.81it/s]
                                                   

 56%|█████▌    | 3218/5758 [18:17<15:03,  2.81it/s]
 56%|█████▌    | 3219/5758 [18:17<15:29,  2.73it/s]
                                                   

 56%|█████▌    | 3219/5758 [18:17<15:29,  2.73it/s]
 56%|█████▌    | 3220/5758 [18:18<14:34,  2.90it/s]
                                                   

 56%|█████▌    | 3220/5758 [18:18<14:34,  2.90it/s]
 56%|█████▌    | 3221/5758 [18:18<15:06,  2.80it/s]
                                                   
{'loss': '0.7282', 'grad_norm': '44.25', 'learning_rate': '4.503e-06', 'num_tokens': '2.118e+06', 'mean_token_accuracy': '0.85', 'epoch': '0.5498'}
+{'loss': '1.631', 'grad_norm': '20.38', 'learning_rate': '4.502e-06', 'num_tokens': '2.118e+06', 'mean_token_accuracy': '0.6182', 'epoch': '0.55'}
+{'loss': '1.24', 'grad_norm': '14.69', 'learning_rate': '4.5e-06', 'num_tokens': '2.119e+06', 'mean_token_accuracy': '0.7261', 'epoch': '0.5502'}
+{'loss': '1.872', 'grad_norm': '22.62', 'learning_rate': '4.498e-06', 'num_tokens': '2.119e+06', 'mean_token_accuracy': '0.6213', 'epoch': '0.5504'}
+{'loss': '1.393', 'grad_norm': '15.88', 'learning_rate': '4.496e-06', 'num_tokens': '2.12e+06', 'mean_token_accuracy': '0.6738', 'epoch': '0.5505'}
+{'loss': '0.9795', 'grad_norm': '38.25', 'learning_rate': '4.495e-06', 'num_tokens': '2.12e+06', 'mean_token_accuracy': '0.8429', 'epoch': '0.5507'}
+{'loss': '1.069', 'grad_norm': '11.75', 'learning_rate': '4.493e-06', 'num_tokens': '2.121e+06', 'mean_token_accuracy': '0.7098', 'epoch': '0.5509'}
+{'loss': '2.027', 'grad_norm': '12.69', 'learning_rate': '4.491e-06', 'num_tokens': '2.121e+06', 'mean_token_accuracy': '0.5729', 'epoch': '0.5511'}
+{'loss': '1.254', 'grad_norm': '13.31', 'learning_rate': '4.489e-06', 'num_tokens': '2.122e+06', 'mean_token_accuracy': '0.6845', 'epoch': '0.5512'}
+{'loss': '1.138', 'grad_norm': '11.5', 'learning_rate': '4.488e-06', 'num_tokens': '2.123e+06', 'mean_token_accuracy': '0.7095', 'epoch': '0.5514'}
+{'loss': '1.3', 'grad_norm': '12.75', 'learning_rate': '4.486e-06', 'num_tokens': '2.123e+06', 'mean_token_accuracy': '0.6455', 'epoch': '0.5516'}
+{'loss': '1.524', 'grad_norm': '11.38', 'learning_rate': '4.484e-06', 'num_tokens': '2.124e+06', 'mean_token_accuracy': '0.6638', 'epoch': '0.5518'}
+{'loss': '1.165', 'grad_norm': '13.06', 'learning_rate': '4.482e-06', 'num_tokens': '2.125e+06', 'mean_token_accuracy': '0.682', 'epoch': '0.5519'}
+{'loss': '1.753', 'grad_norm': '11', 'learning_rate': '4.481e-06', 'num_tokens': '2.126e+06', 'mean_token_accuracy': '0.6361', 'epoch': '0.5521'}
+{'loss': '1.272', 'grad_norm': '13.06', 'learning_rate': '4.479e-06', 'num_tokens': '2.127e+06', 'mean_token_accuracy': '0.6759', 'epoch': '0.5523'}
+{'loss': '1.216', 'grad_norm': '11.12', 'learning_rate': '4.477e-06', 'num_tokens': '2.127e+06', 'mean_token_accuracy': '0.6563', 'epoch': '0.5524'}
+{'loss': '1.557', 'grad_norm': '18.25', 'learning_rate': '4.476e-06', 'num_tokens': '2.128e+06', 'mean_token_accuracy': '0.6489', 'epoch': '0.5526'}
+{'loss': '1.538', 'grad_norm': '11', 'learning_rate': '4.474e-06', 'num_tokens': '2.129e+06', 'mean_token_accuracy': '0.6711', 'epoch': '0.5528'}
+{'loss': '0.7361', 'grad_norm': '26.5', 'learning_rate': '4.472e-06', 'num_tokens': '2.129e+06', 'mean_token_accuracy': '0.8351', 'epoch': '0.553'}
+{'loss': '0.9045', 'grad_norm': '16.25', 'learning_rate': '4.47e-06', 'num_tokens': '2.129e+06', 'mean_token_accuracy': '0.7556', 'epoch': '0.5531'}
+{'loss': '1.438', 'grad_norm': '12.44', 'learning_rate': '4.469e-06', 'num_tokens': '2.13e+06', 'mean_token_accuracy': '0.6381', 'epoch': '0.5533'}
+{'loss': '1.215', 'grad_norm': '7.688', 'learning_rate': '4.467e-06', 'num_tokens': '2.132e+06', 'mean_token_accuracy': '0.7388', 'epoch': '0.5535'}
+{'loss': '1.258', 'grad_norm': '9.188', 'learning_rate': '4.465e-06', 'num_tokens': '2.133e+06', 'mean_token_accuracy': '0.7253', 'epoch': '0.5537'}
+{'loss': '2.457', 'grad_norm': '19.5', 'learning_rate': '4.463e-06', 'num_tokens': '2.133e+06', 'mean_token_accuracy': '0.5487', 'epoch': '0.5538'}
+{'loss': '1.023', 'grad_norm': '16', 'learning_rate': '4.462e-06', 'num_tokens': '2.134e+06', 'mean_token_accuracy': '0.7291', 'epoch': '0.554'}
+{'loss': '1.141', 'grad_norm': '12.94', 'learning_rate': '4.46e-06', 'num_tokens': '2.134e+06', 'mean_token_accuracy': '0.6809', 'epoch': '0.5542'}
+{'loss': '0.2824', 'grad_norm': '15.69', 'learning_rate': '4.458e-06', 'num_tokens': '2.134e+06', 'mean_token_accuracy': '0.9193', 'epoch': '0.5544'}
+{'loss': '0.4666', 'grad_norm': '6.844', 'learning_rate': '4.456e-06', 'num_tokens': '2.136e+06', 'mean_token_accuracy': '0.858', 'epoch': '0.5545'}
+{'loss': '1.105', 'grad_norm': '13.56', 'learning_rate': '4.455e-06', 'num_tokens': '2.137e+06', 'mean_token_accuracy': '0.7147', 'epoch': '0.5547'}
+{'loss': '1.774', 'grad_norm': '21.38', 'learning_rate': '4.453e-06', 'num_tokens': '2.137e+06', 'mean_token_accuracy': '0.6727', 'epoch': '0.5549'}
+{'loss': '0.8532', 'grad_norm': '21.5', 'learning_rate': '4.451e-06', 'num_tokens': '2.137e+06', 'mean_token_accuracy': '0.8161', 'epoch': '0.5551'}
+{'loss': '0.9134', 'grad_norm': '13.06', 'learning_rate': '4.449e-06', 'num_tokens': '2.138e+06', 'mean_token_accuracy': '0.7227', 'epoch': '0.5552'}
+{'loss': '1.051', 'grad_norm': '12.06', 'learning_rate': '4.448e-06', 'num_tokens': '2.138e+06', 'mean_token_accuracy': '0.7178', 'epoch': '0.5554'}
+{'loss': '2.387', 'grad_norm': '15.12', 'learning_rate': '4.446e-06', 'num_tokens': '2.139e+06', 'mean_token_accuracy': '0.5054', 'epoch': '0.5556'}
+{'loss': '1.74', 'grad_norm': '7.875', 'learning_rate': '4.444e-06', 'num_tokens': '2.142e+06', 'mean_token_accuracy': '0.6166', 'epoch': '0.5557'}
+{'loss': '1.9', 'grad_norm': '56', 'learning_rate': '4.443e-06', 'num_tokens': '2.142e+06', 'mean_token_accuracy': '0.5897', 'epoch': '0.5559'}
+{'loss': '0.5714', 'grad_norm': '7.562', 'learning_rate': '4.441e-06', 'num_tokens': '2.143e+06', 'mean_token_accuracy': '0.8391', 'epoch': '0.5561'}
+{'loss': '0.5934', 'grad_norm': '39.25', 'learning_rate': '4.439e-06', 'num_tokens': '2.143e+06', 'mean_token_accuracy': '0.8286', 'epoch': '0.5563'}
+{'loss': '0.6973', 'grad_norm': '9.188', 'learning_rate': '4.437e-06', 'num_tokens': '2.144e+06', 'mean_token_accuracy': '0.7814', 'epoch': '0.5564'}
+{'loss': '1.291', 'grad_norm': '13.06', 'learning_rate': '4.436e-06', 'num_tokens': '2.145e+06', 'mean_token_accuracy': '0.7113', 'epoch': '0.5566'}
+{'loss': '0.9872', 'grad_norm': '13.31', 'learning_rate': '4.434e-06', 'num_tokens': '2.146e+06', 'mean_token_accuracy': '0.728', 'epoch': '0.5568'}
+{'loss': '1.237', 'grad_norm': '13.81', 'learning_rate': '4.432e-06', 'num_tokens': '2.146e+06', 'mean_token_accuracy': '0.653', 'epoch': '0.557'}
+{'loss': '0.9929', 'grad_norm': '31.5', 'learning_rate': '4.43e-06', 'num_tokens': '2.147e+06', 'mean_token_accuracy': '0.839', 'epoch': '0.5571'}
+{'loss': '2.057', 'grad_norm': '12.62', 'learning_rate': '4.429e-06', 'num_tokens': '2.147e+06', 'mean_token_accuracy': '0.5593', 'epoch': '0.5573'}
+{'loss': '0.9', 'grad_norm': '11.19', 'learning_rate': '4.427e-06', 'num_tokens': '2.148e+06', 'mean_token_accuracy': '0.7236', 'epoch': '0.5575'}
+{'loss': '1.153', 'grad_norm': '12.62', 'learning_rate': '4.425e-06', 'num_tokens': '2.149e+06', 'mean_token_accuracy': '0.6492', 'epoch': '0.5577'}
+{'loss': '2.016', 'grad_norm': '19.38', 'learning_rate': '4.423e-06', 'num_tokens': '2.149e+06', 'mean_token_accuracy': '0.5809', 'epoch': '0.5578'}
+{'loss': '0.9321', 'grad_norm': '12.56', 'learning_rate': '4.422e-06', 'num_tokens': '2.15e+06', 'mean_token_accuracy': '0.7373', 'epoch': '0.558'}
+{'loss': '1.144', 'grad_norm': '17.75', 'learning_rate': '4.42e-06', 'num_tokens': '2.15e+06', 'mean_token_accuracy': '0.6809', 'epoch': '0.5582'}
+{'loss': '1.639', 'grad_norm': '20.88', 'learning_rate': '4.418e-06', 'num_tokens': '2.151e+06', 'mean_token_accuracy': '0.6508', 'epoch': '0.5584'}
+{'loss': '1.557', 'grad_norm': '20', 'learning_rate': '4.416e-06', 'num_tokens': '2.151e+06', 'mean_token_accuracy': '0.6639', 'epoch': '0.5585'}
+{'loss': '1.287', 'grad_norm': '19.5', 'learning_rate': '4.415e-06', 'num_tokens': '2.151e+06', 'mean_token_accuracy': '0.7747', 'epoch': '0.5587'}
+{'loss': '1.755', 'grad_norm': '22.88', 'learning_rate': '4.413e-06', 'num_tokens': '2.151e+06', 'mean_token_accuracy': '0.6295', 'epoch': '0.5589'}
+{'loss': '0.496', 'grad_norm': '15.62', 'learning_rate': '4.411e-06', 'num_tokens': '2.152e+06', 'mean_token_accuracy': '0.8923', 'epoch': '0.559'}
+{'loss': '1.204', 'grad_norm': '8.875', 'learning_rate': '4.41e-06', 'num_tokens': '2.153e+06', 'mean_token_accuracy': '0.7256', 'epoch': '0.5592'}
+
 56%|█████▌    | 3221/5758 [18:18<15:06,  2.80it/s]
 56%|█████▌    | 3222/5758 [18:19<14:50,  2.85it/s]
                                                   

 56%|█████▌    | 3222/5758 [18:19<14:50,  2.85it/s]
 56%|█████▌    | 3223/5758 [18:19<14:44,  2.87it/s]
                                                   

 56%|█████▌    | 3223/5758 [18:19<14:44,  2.87it/s]
 56%|█████▌    | 3224/5758 [18:19<14:45,  2.86it/s]
                                                   

 56%|█████▌    | 3224/5758 [18:19<14:45,  2.86it/s]
 56%|█████▌    | 3225/5758 [18:20<14:20,  2.94it/s]
                                                   

 56%|█████▌    | 3225/5758 [18:20<14:20,  2.94it/s]
 56%|█████▌    | 3226/5758 [18:20<14:01,  3.01it/s]
                                                   

 56%|█████▌    | 3226/5758 [18:20<14:01,  3.01it/s]
 56%|█████▌    | 3227/5758 [18:20<14:34,  2.89it/s]
                                                   

 56%|█████▌    | 3227/5758 [18:20<14:34,  2.89it/s]
 56%|█████▌    | 3228/5758 [18:21<14:43,  2.86it/s]
                                                   

 56%|█████▌    | 3228/5758 [18:21<14:43,  2.86it/s]
 56%|█████▌    | 3229/5758 [18:21<14:02,  3.00it/s]
                                                   

 56%|█████▌    | 3229/5758 [18:21<14:02,  3.00it/s]
 56%|█████▌    | 3230/5758 [18:21<14:34,  2.89it/s]
                                                   

 56%|█████▌    | 3230/5758 [18:21<14:34,  2.89it/s]
 56%|█████▌    | 3231/5758 [18:22<14:59,  2.81it/s]
                                                   

 56%|█████▌    | 3231/5758 [18:22<14:59,  2.81it/s]
 56%|█████▌    | 3232/5758 [18:22<15:27,  2.72it/s]
                                                   

 56%|█████▌    | 3232/5758 [18:22<15:27,  2.72it/s]
 56%|█████▌    | 3233/5758 [18:22<14:54,  2.82it/s]
                                                   

 56%|█████▌    | 3233/5758 [18:22<14:54,  2.82it/s]
 56%|█████▌    | 3234/5758 [18:23<14:15,  2.95it/s]
                                                   

 56%|█████▌    | 3234/5758 [18:23<14:15,  2.95it/s]
 56%|█████▌    | 3235/5758 [18:23<14:25,  2.91it/s]
                                                   

 56%|█████▌    | 3235/5758 [18:23<14:25,  2.91it/s]
 56%|█████▌    | 3236/5758 [18:23<13:54,  3.02it/s]
                                                   

 56%|█████▌    | 3236/5758 [18:23<13:54,  3.02it/s]
 56%|█████▌    | 3237/5758 [18:24<15:18,  2.74it/s]
                                                   

 56%|█████▌    | 3237/5758 [18:24<15:18,  2.74it/s]
 56%|█████▌    | 3238/5758 [18:24<14:46,  2.84it/s]
                                                   

 56%|█████▌    | 3238/5758 [18:24<14:46,  2.84it/s]
 56%|█████▋    | 3239/5758 [18:24<14:21,  2.92it/s]
                                                   

 56%|█████▋    | 3239/5758 [18:24<14:21,  2.92it/s]
 56%|█████▋    | 3240/5758 [18:25<14:46,  2.84it/s]
                                                   

 56%|█████▋    | 3240/5758 [18:25<14:46,  2.84it/s]
 56%|█████▋    | 3241/5758 [18:25<14:28,  2.90it/s]
                                                   

 56%|█████▋    | 3241/5758 [18:25<14:28,  2.90it/s]
 56%|█████▋    | 3242/5758 [18:25<13:54,  3.02it/s]
                                                   

 56%|█████▋    | 3242/5758 [18:25<13:54,  3.02it/s]
 56%|█████▋    | 3243/5758 [18:26<13:56,  3.01it/s]
                                                   

 56%|█████▋    | 3243/5758 [18:26<13:56,  3.01it/s]
 56%|█████▋    | 3244/5758 [18:26<14:09,  2.96it/s]
                                                   

 56%|█████▋    | 3244/5758 [18:26<14:09,  2.96it/s]
 56%|█████▋    | 3245/5758 [18:26<13:43,  3.05it/s]
                                                   

 56%|█████▋    | 3245/5758 [18:26<13:43,  3.05it/s]
 56%|█████▋    | 3246/5758 [18:27<14:10,  2.96it/s]
                                                   

 56%|█████▋    | 3246/5758 [18:27<14:10,  2.96it/s]
 56%|█████▋    | 3247/5758 [18:27<14:43,  2.84it/s]
                                                   

 56%|█████▋    | 3247/5758 [18:27<14:43,  2.84it/s]
 56%|█████▋    | 3248/5758 [18:27<14:11,  2.95it/s]
                                                   

 56%|█████▋    | 3248/5758 [18:27<14:11,  2.95it/s]
 56%|█████▋    | 3249/5758 [18:28<14:42,  2.84it/s]
                                                   

 56%|█████▋    | 3249/5758 [18:28<14:42,  2.84it/s]
 56%|█████▋    | 3250/5758 [18:28<14:52,  2.81it/s]
                                                   

 56%|█████▋    | 3250/5758 [18:28<14:52,  2.81it/s]
 56%|█████▋    | 3251/5758 [18:28<14:06,  2.96it/s]
                                                   

 56%|█████▋    | 3251/5758 [18:28<14:06,  2.96it/s]
 56%|█████▋    | 3252/5758 [18:29<14:51,  2.81it/s]
                                                   

 56%|█████▋    | 3252/5758 [18:29<14:51,  2.81it/s]
 56%|█████▋    | 3253/5758 [18:29<15:08,  2.76it/s]
                                                   

 56%|█████▋    | 3253/5758 [18:29<15:08,  2.76it/s]
 57%|█████▋    | 3254/5758 [18:30<14:20,  2.91it/s]
                                                   

 57%|█████▋    | 3254/5758 [18:30<14:20,  2.91it/s]
 57%|█████▋    | 3255/5758 [18:30<13:50,  3.01it/s]
                                                   

 57%|█████▋    | 3255/5758 [18:30<13:50,  3.01it/s]
 57%|█████▋    | 3256/5758 [18:30<14:11,  2.94it/s]
                                                   

 57%|█████▋    | 3256/5758 [18:30<14:11,  2.94it/s]
 57%|█████▋    | 3257/5758 [18:31<13:48,  3.02it/s]
                                                   

 57%|█████▋    | 3257/5758 [18:31<13:48,  3.02it/s]
 57%|█████▋    | 3258/5758 [18:31<13:44,  3.03it/s]
                                                   

 57%|█████▋    | 3258/5758 [18:31<13:44,  3.03it/s]
 57%|█████▋    | 3259/5758 [18:31<13:21,  3.12it/s]
                                                   

 57%|█████▋    | 3259/5758 [18:31<13:21,  3.12it/s]
 57%|█████▋    | 3260/5758 [18:32<13:50,  3.01it/s]
                                                   

 57%|█████▋    | 3260/5758 [18:32<13:50,  3.01it/s]
 57%|█████▋    | 3261/5758 [18:32<13:31,  3.08it/s]
                                                   

 57%|█████▋    | 3261/5758 [18:32<13:31,  3.08it/s]
 57%|█████▋    | 3262/5758 [18:32<13:20,  3.12it/s]
                                                   

 57%|█████▋    | 3262/5758 [18:32<13:20,  3.12it/s]
 57%|█████▋    | 3263/5758 [18:33<14:37,  2.84it/s]
                                                   

 57%|█████▋    | 3263/5758 [18:33<14:37,  2.84it/s]
 57%|█████▋    | 3264/5758 [18:33<15:12,  2.73it/s]
                                                   

 57%|█████▋    | 3264/5758 [18:33<15:12,  2.73it/s]
 57%|█████▋    | 3265/5758 [18:34<17:31,  2.37it/s]
                                                   

 57%|█████▋    | 3265/5758 [18:34<17:31,  2.37it/s]
 57%|█████▋    | 3266/5758 [18:34<17:10,  2.42it/s]
                                                   

 57%|█████▋    | 3266/5758 [18:34<17:10,  2.42it/s]
 57%|█████▋    | 3267/5758 [18:34<16:44,  2.48it/s]
                                                   

 57%|█████▋    | 3267/5758 [18:34<16:44,  2.48it/s]
 57%|█████▋    | 3268/5758 [18:35<16:08,  2.57it/s]
                                                   

 57%|█████▋    | 3268/5758 [18:35<16:08,  2.57it/s]
 57%|█████▋    | 3269/5758 [18:35<15:30,  2.67it/s]
                                                   

 57%|█████▋    | 3269/5758 [18:35<15:30,  2.67it/s]
 57%|█████▋    | 3270/5758 [18:35<15:28,  2.68it/s]
                                                   

 57%|█████▋    | 3270/5758 [18:35<15:28,  2.68it/s]
 57%|█████▋    | 3271/5758 [18:36<14:28,  2.86it/s]
                                                   

 57%|█████▋    | 3271/5758 [18:36<14:28,  2.86it/s]
 57%|█████▋    | 3272/5758 [18:36<14:47,  2.80it/s]
                                                   

 57%|█████▋    | 3272/5758 [18:36<14:47,  2.80it/s]
 57%|█████▋    | 3273/5758 [18:36<15:04,  2.75it/s]
                                                   

 57%|█████▋    | 3273/5758 [18:36<15:04,  2.75it/s]
 57%|█████▋    | 3274/5758 [18:37<14:36,  2.83it/s]
                                                   

 57%|█████▋    | 3274/5758 [18:37<14:36,  2.83it/s]
 57%|█████▋    | 3275/5758 [18:37<14:34,  2.84it/s]
                                                   

 57%|█████▋    | 3275/5758 [18:37<14:34,  2.84it/s]
 57%|█████▋    | 3276/5758 [18:37<14:51,  2.78it/s]
                                                   
{'loss': '1.057', 'grad_norm': '20.25', 'learning_rate': '4.408e-06', 'num_tokens': '2.153e+06', 'mean_token_accuracy': '0.6936', 'epoch': '0.5594'}
+{'loss': '1.165', 'grad_norm': '12.94', 'learning_rate': '4.406e-06', 'num_tokens': '2.154e+06', 'mean_token_accuracy': '0.7165', 'epoch': '0.5596'}
+{'loss': '1.076', 'grad_norm': '10.44', 'learning_rate': '4.404e-06', 'num_tokens': '2.155e+06', 'mean_token_accuracy': '0.7443', 'epoch': '0.5597'}
+{'loss': '1.752', 'grad_norm': '22.25', 'learning_rate': '4.403e-06', 'num_tokens': '2.155e+06', 'mean_token_accuracy': '0.6432', 'epoch': '0.5599'}
+{'loss': '1.349', 'grad_norm': '12.19', 'learning_rate': '4.401e-06', 'num_tokens': '2.156e+06', 'mean_token_accuracy': '0.6571', 'epoch': '0.5601'}
+{'loss': '1.223', 'grad_norm': '11.31', 'learning_rate': '4.399e-06', 'num_tokens': '2.157e+06', 'mean_token_accuracy': '0.6778', 'epoch': '0.5603'}
+{'loss': '1.74', 'grad_norm': '16.5', 'learning_rate': '4.397e-06', 'num_tokens': '2.157e+06', 'mean_token_accuracy': '0.6221', 'epoch': '0.5604'}
+{'loss': '0.718', 'grad_norm': '26.5', 'learning_rate': '4.396e-06', 'num_tokens': '2.157e+06', 'mean_token_accuracy': '0.9', 'epoch': '0.5606'}
+{'loss': '1.182', 'grad_norm': '13.19', 'learning_rate': '4.394e-06', 'num_tokens': '2.158e+06', 'mean_token_accuracy': '0.6748', 'epoch': '0.5608'}
+{'loss': '2.2', 'grad_norm': '19.12', 'learning_rate': '4.392e-06', 'num_tokens': '2.159e+06', 'mean_token_accuracy': '0.5582', 'epoch': '0.561'}
+{'loss': '0.7206', 'grad_norm': '26.5', 'learning_rate': '4.39e-06', 'num_tokens': '2.159e+06', 'mean_token_accuracy': '0.8095', 'epoch': '0.5611'}
+{'loss': '1.881', 'grad_norm': '17.88', 'learning_rate': '4.389e-06', 'num_tokens': '2.159e+06', 'mean_token_accuracy': '0.6284', 'epoch': '0.5613'}
+{'loss': '1.451', 'grad_norm': '12.69', 'learning_rate': '4.387e-06', 'num_tokens': '2.16e+06', 'mean_token_accuracy': '0.6118', 'epoch': '0.5615'}
+{'loss': '1.823', 'grad_norm': '11.75', 'learning_rate': '4.385e-06', 'num_tokens': '2.161e+06', 'mean_token_accuracy': '0.6006', 'epoch': '0.5617'}
+{'loss': '1.57', 'grad_norm': '15', 'learning_rate': '4.383e-06', 'num_tokens': '2.161e+06', 'mean_token_accuracy': '0.6886', 'epoch': '0.5618'}
+{'loss': '1.131', 'grad_norm': '9.062', 'learning_rate': '4.382e-06', 'num_tokens': '2.163e+06', 'mean_token_accuracy': '0.7253', 'epoch': '0.562'}
+{'loss': '1.068', 'grad_norm': '28.12', 'learning_rate': '4.38e-06', 'num_tokens': '2.163e+06', 'mean_token_accuracy': '0.7487', 'epoch': '0.5622'}
+{'loss': '1.541', 'grad_norm': '11.44', 'learning_rate': '4.378e-06', 'num_tokens': '2.164e+06', 'mean_token_accuracy': '0.6596', 'epoch': '0.5623'}
+{'loss': '0.5453', 'grad_norm': '7', 'learning_rate': '4.377e-06', 'num_tokens': '2.165e+06', 'mean_token_accuracy': '0.8321', 'epoch': '0.5625'}
+{'loss': '1.929', 'grad_norm': '11.19', 'learning_rate': '4.375e-06', 'num_tokens': '2.166e+06', 'mean_token_accuracy': '0.6', 'epoch': '0.5627'}
+{'loss': '0.9831', 'grad_norm': '29.75', 'learning_rate': '4.373e-06', 'num_tokens': '2.166e+06', 'mean_token_accuracy': '0.7617', 'epoch': '0.5629'}
+{'loss': '1.201', 'grad_norm': '13.88', 'learning_rate': '4.371e-06', 'num_tokens': '2.167e+06', 'mean_token_accuracy': '0.6494', 'epoch': '0.563'}
+{'loss': '1.156', 'grad_norm': '11.94', 'learning_rate': '4.37e-06', 'num_tokens': '2.168e+06', 'mean_token_accuracy': '0.6802', 'epoch': '0.5632'}
+{'loss': '1.999', 'grad_norm': '22.88', 'learning_rate': '4.368e-06', 'num_tokens': '2.168e+06', 'mean_token_accuracy': '0.5865', 'epoch': '0.5634'}
+{'loss': '1.419', 'grad_norm': '30.25', 'learning_rate': '4.366e-06', 'num_tokens': '2.168e+06', 'mean_token_accuracy': '0.7151', 'epoch': '0.5636'}
+{'loss': '1.276', 'grad_norm': '6.812', 'learning_rate': '4.364e-06', 'num_tokens': '2.17e+06', 'mean_token_accuracy': '0.7242', 'epoch': '0.5637'}
+{'loss': '1.557', 'grad_norm': '39', 'learning_rate': '4.363e-06', 'num_tokens': '2.171e+06', 'mean_token_accuracy': '0.688', 'epoch': '0.5639'}
+{'loss': '0.6389', 'grad_norm': '32.5', 'learning_rate': '4.361e-06', 'num_tokens': '2.171e+06', 'mean_token_accuracy': '0.825', 'epoch': '0.5641'}
+{'loss': '1.175', 'grad_norm': '10.25', 'learning_rate': '4.359e-06', 'num_tokens': '2.172e+06', 'mean_token_accuracy': '0.6633', 'epoch': '0.5643'}
+{'loss': '1.125', 'grad_norm': '19.25', 'learning_rate': '4.357e-06', 'num_tokens': '2.172e+06', 'mean_token_accuracy': '0.7395', 'epoch': '0.5644'}
+{'loss': '2.184', 'grad_norm': '20', 'learning_rate': '4.356e-06', 'num_tokens': '2.172e+06', 'mean_token_accuracy': '0.5266', 'epoch': '0.5646'}
+{'loss': '0.5745', 'grad_norm': '8.125', 'learning_rate': '4.354e-06', 'num_tokens': '2.173e+06', 'mean_token_accuracy': '0.8042', 'epoch': '0.5648'}
+{'loss': '1.94', 'grad_norm': '17.75', 'learning_rate': '4.352e-06', 'num_tokens': '2.174e+06', 'mean_token_accuracy': '0.5829', 'epoch': '0.565'}
+{'loss': '1.078', 'grad_norm': '13.31', 'learning_rate': '4.35e-06', 'num_tokens': '2.175e+06', 'mean_token_accuracy': '0.6987', 'epoch': '0.5651'}
+{'loss': '2.082', 'grad_norm': '27.5', 'learning_rate': '4.349e-06', 'num_tokens': '2.175e+06', 'mean_token_accuracy': '0.5536', 'epoch': '0.5653'}
+{'loss': '0.6462', 'grad_norm': '6.719', 'learning_rate': '4.347e-06', 'num_tokens': '2.177e+06', 'mean_token_accuracy': '0.8088', 'epoch': '0.5655'}
+{'loss': '1.706', 'grad_norm': '30', 'learning_rate': '4.345e-06', 'num_tokens': '2.177e+06', 'mean_token_accuracy': '0.6724', 'epoch': '0.5656'}
+{'loss': '2.08', 'grad_norm': '17.12', 'learning_rate': '4.344e-06', 'num_tokens': '2.177e+06', 'mean_token_accuracy': '0.58', 'epoch': '0.5658'}
+{'loss': '1.026', 'grad_norm': '12.88', 'learning_rate': '4.342e-06', 'num_tokens': '2.178e+06', 'mean_token_accuracy': '0.7107', 'epoch': '0.566'}
+{'loss': '1.427', 'grad_norm': '14.75', 'learning_rate': '4.34e-06', 'num_tokens': '2.178e+06', 'mean_token_accuracy': '0.6963', 'epoch': '0.5662'}
+{'loss': '0.7412', 'grad_norm': '20.12', 'learning_rate': '4.338e-06', 'num_tokens': '2.179e+06', 'mean_token_accuracy': '0.7989', 'epoch': '0.5663'}
+{'loss': '1.474', 'grad_norm': '8.688', 'learning_rate': '4.337e-06', 'num_tokens': '2.18e+06', 'mean_token_accuracy': '0.6722', 'epoch': '0.5665'}
+{'loss': '0.7841', 'grad_norm': '26.5', 'learning_rate': '4.335e-06', 'num_tokens': '2.18e+06', 'mean_token_accuracy': '0.8444', 'epoch': '0.5667'}
+{'loss': '1.834', 'grad_norm': '13.19', 'learning_rate': '4.333e-06', 'num_tokens': '2.181e+06', 'mean_token_accuracy': '0.6106', 'epoch': '0.5669'}
+{'loss': '1.351', 'grad_norm': '18.62', 'learning_rate': '4.331e-06', 'num_tokens': '2.181e+06', 'mean_token_accuracy': '0.7075', 'epoch': '0.567'}
+{'loss': '1.6', 'grad_norm': '13.44', 'learning_rate': '4.33e-06', 'num_tokens': '2.181e+06', 'mean_token_accuracy': '0.6522', 'epoch': '0.5672'}
+{'loss': '0.6139', 'grad_norm': '8.625', 'learning_rate': '4.328e-06', 'num_tokens': '2.183e+06', 'mean_token_accuracy': '0.8208', 'epoch': '0.5674'}
+{'loss': '0.926', 'grad_norm': '18.25', 'learning_rate': '4.326e-06', 'num_tokens': '2.183e+06', 'mean_token_accuracy': '0.7958', 'epoch': '0.5676'}
+{'loss': '1.564', 'grad_norm': '13.44', 'learning_rate': '4.324e-06', 'num_tokens': '2.184e+06', 'mean_token_accuracy': '0.5771', 'epoch': '0.5677'}
+{'loss': '0.868', 'grad_norm': '8.75', 'learning_rate': '4.323e-06', 'num_tokens': '2.185e+06', 'mean_token_accuracy': '0.7989', 'epoch': '0.5679'}
+{'loss': '1.095', 'grad_norm': '15.25', 'learning_rate': '4.321e-06', 'num_tokens': '2.185e+06', 'mean_token_accuracy': '0.6957', 'epoch': '0.5681'}
+{'loss': '0.6258', 'grad_norm': '8.562', 'learning_rate': '4.319e-06', 'num_tokens': '2.186e+06', 'mean_token_accuracy': '0.8119', 'epoch': '0.5683'}
+{'loss': '0.942', 'grad_norm': '17.38', 'learning_rate': '4.317e-06', 'num_tokens': '2.187e+06', 'mean_token_accuracy': '0.7593', 'epoch': '0.5684'}
+{'loss': '1.304', 'grad_norm': '12.94', 'learning_rate': '4.316e-06', 'num_tokens': '2.188e+06', 'mean_token_accuracy': '0.6493', 'epoch': '0.5686'}
+{'loss': '1.498', 'grad_norm': '12.69', 'learning_rate': '4.314e-06', 'num_tokens': '2.189e+06', 'mean_token_accuracy': '0.6758', 'epoch': '0.5688'}
+
 57%|█████▋    | 3276/5758 [18:37<14:51,  2.78it/s]
 57%|█████▋    | 3277/5758 [18:38<15:17,  2.70it/s]
                                                   

 57%|█████▋    | 3277/5758 [18:38<15:17,  2.70it/s]
 57%|█████▋    | 3278/5758 [18:38<14:41,  2.81it/s]
                                                   

 57%|█████▋    | 3278/5758 [18:38<14:41,  2.81it/s]
 57%|█████▋    | 3279/5758 [18:39<15:04,  2.74it/s]
                                                   

 57%|█████▋    | 3279/5758 [18:39<15:04,  2.74it/s]
 57%|█████▋    | 3280/5758 [18:39<14:52,  2.78it/s]
                                                   

 57%|█████▋    | 3280/5758 [18:39<14:52,  2.78it/s]
 57%|█████▋    | 3281/5758 [18:39<14:25,  2.86it/s]
                                                   

 57%|█████▋    | 3281/5758 [18:39<14:25,  2.86it/s]
 57%|█████▋    | 3282/5758 [18:40<14:06,  2.93it/s]
                                                   

 57%|█████▋    | 3282/5758 [18:40<14:06,  2.93it/s]
 57%|█████▋    | 3283/5758 [18:40<14:56,  2.76it/s]
                                                   

 57%|█████▋    | 3283/5758 [18:40<14:56,  2.76it/s]
 57%|█████▋    | 3284/5758 [18:40<14:55,  2.76it/s]
                                                   

 57%|█████▋    | 3284/5758 [18:40<14:55,  2.76it/s]
 57%|█████▋    | 3285/5758 [18:41<14:14,  2.89it/s]
                                                   

 57%|█████▋    | 3285/5758 [18:41<14:14,  2.89it/s]
 57%|█████▋    | 3286/5758 [18:41<13:34,  3.04it/s]
                                                   

 57%|█████▋    | 3286/5758 [18:41<13:34,  3.04it/s]
 57%|█████▋    | 3287/5758 [18:41<13:06,  3.14it/s]
                                                   

 57%|█████▋    | 3287/5758 [18:41<13:06,  3.14it/s]
 57%|█████▋    | 3288/5758 [18:41<12:44,  3.23it/s]
                                                   

 57%|█████▋    | 3288/5758 [18:42<12:44,  3.23it/s]
 57%|█████▋    | 3289/5758 [18:42<12:52,  3.20it/s]
                                                   

 57%|█████▋    | 3289/5758 [18:42<12:52,  3.20it/s]
 57%|█████▋    | 3290/5758 [18:42<14:04,  2.92it/s]
                                                   

 57%|█████▋    | 3290/5758 [18:42<14:04,  2.92it/s]
 57%|█████▋    | 3291/5758 [18:43<13:41,  3.00it/s]
                                                   

 57%|█████▋    | 3291/5758 [18:43<13:41,  3.00it/s]
 57%|█████▋    | 3292/5758 [18:43<14:34,  2.82it/s]
                                                   

 57%|█████▋    | 3292/5758 [18:43<14:34,  2.82it/s]
 57%|█████▋    | 3293/5758 [18:43<13:59,  2.94it/s]
                                                   

 57%|█████▋    | 3293/5758 [18:43<13:59,  2.94it/s]
 57%|█████▋    | 3294/5758 [18:44<14:05,  2.92it/s]
                                                   

 57%|█████▋    | 3294/5758 [18:44<14:05,  2.92it/s]
 57%|█████▋    | 3295/5758 [18:44<13:39,  3.01it/s]
                                                   

 57%|█████▋    | 3295/5758 [18:44<13:39,  3.01it/s]
 57%|█████▋    | 3296/5758 [18:44<13:43,  2.99it/s]
                                                   

 57%|█████▋    | 3296/5758 [18:44<13:43,  2.99it/s]
 57%|█████▋    | 3297/5758 [18:45<13:23,  3.06it/s]
                                                   

 57%|█████▋    | 3297/5758 [18:45<13:23,  3.06it/s]
 57%|█████▋    | 3298/5758 [18:45<13:09,  3.12it/s]
                                                   

 57%|█████▋    | 3298/5758 [18:45<13:09,  3.12it/s]
 57%|█████▋    | 3299/5758 [18:45<13:26,  3.05it/s]
                                                   

 57%|█████▋    | 3299/5758 [18:45<13:26,  3.05it/s]
 57%|█████▋    | 3300/5758 [18:46<13:21,  3.07it/s]
                                                   

 57%|█████▋    | 3300/5758 [18:46<13:21,  3.07it/s]
 57%|█████▋    | 3301/5758 [18:46<13:31,  3.03it/s]
                                                   

 57%|█████▋    | 3301/5758 [18:46<13:31,  3.03it/s]
 57%|█████▋    | 3302/5758 [18:46<13:06,  3.12it/s]
                                                   

 57%|█████▋    | 3302/5758 [18:46<13:06,  3.12it/s]
 57%|█████▋    | 3303/5758 [18:47<13:18,  3.07it/s]
                                                   

 57%|█████▋    | 3303/5758 [18:47<13:18,  3.07it/s]
 57%|█████▋    | 3304/5758 [18:47<13:02,  3.13it/s]
                                                   

 57%|█████▋    | 3304/5758 [18:47<13:02,  3.13it/s]
 57%|█████▋    | 3305/5758 [18:47<13:55,  2.93it/s]
                                                   

 57%|█████▋    | 3305/5758 [18:47<13:55,  2.93it/s]
 57%|█████▋    | 3306/5758 [18:48<13:26,  3.04it/s]
                                                   

 57%|█████▋    | 3306/5758 [18:48<13:26,  3.04it/s]
 57%|█████▋    | 3307/5758 [18:48<13:33,  3.01it/s]
                                                   

 57%|█████▋    | 3307/5758 [18:48<13:33,  3.01it/s]
 57%|█████▋    | 3308/5758 [18:48<13:41,  2.98it/s]
                                                   

 57%|█████▋    | 3308/5758 [18:48<13:41,  2.98it/s]
 57%|█████▋    | 3309/5758 [18:48<13:15,  3.08it/s]
                                                   

 57%|█████▋    | 3309/5758 [18:48<13:15,  3.08it/s]
 57%|█████▋    | 3310/5758 [18:49<13:29,  3.03it/s]
                                                   

 57%|█████▋    | 3310/5758 [18:49<13:29,  3.03it/s]
 58%|█████▊    | 3311/5758 [18:49<13:06,  3.11it/s]
                                                   

 58%|█████▊    | 3311/5758 [18:49<13:06,  3.11it/s]
 58%|█████▊    | 3312/5758 [18:49<13:00,  3.14it/s]
                                                   

 58%|█████▊    | 3312/5758 [18:49<13:00,  3.14it/s]
 58%|█████▊    | 3313/5758 [18:50<13:28,  3.02it/s]
                                                   

 58%|█████▊    | 3313/5758 [18:50<13:28,  3.02it/s]
 58%|█████▊    | 3314/5758 [18:50<13:43,  2.97it/s]
                                                   

 58%|█████▊    | 3314/5758 [18:50<13:43,  2.97it/s]
 58%|█████▊    | 3315/5758 [18:50<13:16,  3.07it/s]
                                                   

 58%|█████▊    | 3315/5758 [18:50<13:16,  3.07it/s]
 58%|█████▊    | 3316/5758 [18:51<15:03,  2.70it/s]
                                                   

 58%|█████▊    | 3316/5758 [18:51<15:03,  2.70it/s]
 58%|█████▊    | 3317/5758 [18:51<15:09,  2.68it/s]
                                                   

 58%|█████▊    | 3317/5758 [18:51<15:09,  2.68it/s]
 58%|█████▊    | 3318/5758 [18:52<14:35,  2.79it/s]
                                                   

 58%|█████▊    | 3318/5758 [18:52<14:35,  2.79it/s]
 58%|█████▊    | 3319/5758 [18:52<14:12,  2.86it/s]
                                                   

 58%|█████▊    | 3319/5758 [18:52<14:12,  2.86it/s]
 58%|█████▊    | 3320/5758 [18:52<14:15,  2.85it/s]
                                                   

 58%|█████▊    | 3320/5758 [18:52<14:15,  2.85it/s]
 58%|█████▊    | 3321/5758 [18:53<13:46,  2.95it/s]
                                                   

 58%|█████▊    | 3321/5758 [18:53<13:46,  2.95it/s]
 58%|█████▊    | 3322/5758 [18:53<13:18,  3.05it/s]
                                                   

 58%|█████▊    | 3322/5758 [18:53<13:18,  3.05it/s]
 58%|█████▊    | 3323/5758 [18:53<13:00,  3.12it/s]
                                                   

 58%|█████▊    | 3323/5758 [18:53<13:00,  3.12it/s]
 58%|█████▊    | 3324/5758 [18:54<12:29,  3.25it/s]
                                                   

 58%|█████▊    | 3324/5758 [18:54<12:29,  3.25it/s]
 58%|█████▊    | 3325/5758 [18:54<12:40,  3.20it/s]
                                                   

 58%|█████▊    | 3325/5758 [18:54<12:40,  3.20it/s]
 58%|█████▊    | 3326/5758 [18:54<12:36,  3.21it/s]
                                                   

 58%|█████▊    | 3326/5758 [18:54<12:36,  3.21it/s]
 58%|█████▊    | 3327/5758 [18:55<13:19,  3.04it/s]
                                                   

 58%|█████▊    | 3327/5758 [18:55<13:19,  3.04it/s]
 58%|█████▊    | 3328/5758 [18:55<13:31,  2.99it/s]
                                                   

 58%|█████▊    | 3328/5758 [18:55<13:31,  2.99it/s]
 58%|█████▊    | 3329/5758 [18:55<13:15,  3.06it/s]
                                                   

 58%|█████▊    | 3329/5758 [18:55<13:15,  3.06it/s]
 58%|█████▊    | 3330/5758 [18:55<12:52,  3.14it/s]
                                                   

 58%|█████▊    | 3330/5758 [18:55<12:52,  3.14it/s]
 58%|█████▊    | 3331/5758 [18:56<12:50,  3.15it/s]
                                                   
{'loss': '1.294', 'grad_norm': '14.56', 'learning_rate': '4.312e-06', 'num_tokens': '2.189e+06', 'mean_token_accuracy': '0.6613', 'epoch': '0.5689'}
+{'loss': '1.142', 'grad_norm': '9.562', 'learning_rate': '4.311e-06', 'num_tokens': '2.19e+06', 'mean_token_accuracy': '0.7321', 'epoch': '0.5691'}
+{'loss': '1.061', 'grad_norm': '11.94', 'learning_rate': '4.309e-06', 'num_tokens': '2.191e+06', 'mean_token_accuracy': '0.7236', 'epoch': '0.5693'}
+{'loss': '1.111', 'grad_norm': '14.06', 'learning_rate': '4.307e-06', 'num_tokens': '2.192e+06', 'mean_token_accuracy': '0.6822', 'epoch': '0.5695'}
+{'loss': '0.8803', 'grad_norm': '22.12', 'learning_rate': '4.305e-06', 'num_tokens': '2.192e+06', 'mean_token_accuracy': '0.7866', 'epoch': '0.5696'}
+{'loss': '1.431', 'grad_norm': '12.25', 'learning_rate': '4.304e-06', 'num_tokens': '2.193e+06', 'mean_token_accuracy': '0.6278', 'epoch': '0.5698'}
+{'loss': '1.589', 'grad_norm': '41.25', 'learning_rate': '4.302e-06', 'num_tokens': '2.193e+06', 'mean_token_accuracy': '0.6705', 'epoch': '0.57'}
+{'loss': '1.046', 'grad_norm': '29.5', 'learning_rate': '4.3e-06', 'num_tokens': '2.193e+06', 'mean_token_accuracy': '0.7862', 'epoch': '0.5702'}
+{'loss': '0.6873', 'grad_norm': '9.5', 'learning_rate': '4.298e-06', 'num_tokens': '2.194e+06', 'mean_token_accuracy': '0.7867', 'epoch': '0.5703'}
+{'loss': '1.054', 'grad_norm': '11.69', 'learning_rate': '4.297e-06', 'num_tokens': '2.195e+06', 'mean_token_accuracy': '0.7151', 'epoch': '0.5705'}
+{'loss': '1.89', 'grad_norm': '17.88', 'learning_rate': '4.295e-06', 'num_tokens': '2.195e+06', 'mean_token_accuracy': '0.5856', 'epoch': '0.5707'}
+{'loss': '2.012', 'grad_norm': '13.81', 'learning_rate': '4.293e-06', 'num_tokens': '2.196e+06', 'mean_token_accuracy': '0.5724', 'epoch': '0.5709'}
+{'loss': '1.082', 'grad_norm': '13.62', 'learning_rate': '4.291e-06', 'num_tokens': '2.197e+06', 'mean_token_accuracy': '0.6699', 'epoch': '0.571'}
+{'loss': '0.6527', 'grad_norm': '32.75', 'learning_rate': '4.29e-06', 'num_tokens': '2.197e+06', 'mean_token_accuracy': '0.8523', 'epoch': '0.5712'}
+{'loss': '1.394', 'grad_norm': '13.44', 'learning_rate': '4.288e-06', 'num_tokens': '2.197e+06', 'mean_token_accuracy': '0.6487', 'epoch': '0.5714'}
+{'loss': '1.143', 'grad_norm': '15.88', 'learning_rate': '4.286e-06', 'num_tokens': '2.198e+06', 'mean_token_accuracy': '0.7126', 'epoch': '0.5716'}
+{'loss': '1.112', 'grad_norm': '29.88', 'learning_rate': '4.284e-06', 'num_tokens': '2.198e+06', 'mean_token_accuracy': '0.7021', 'epoch': '0.5717'}
+{'loss': '1.745', 'grad_norm': '16.75', 'learning_rate': '4.283e-06', 'num_tokens': '2.198e+06', 'mean_token_accuracy': '0.639', 'epoch': '0.5719'}
+{'loss': '2.255', 'grad_norm': '14.81', 'learning_rate': '4.281e-06', 'num_tokens': '2.199e+06', 'mean_token_accuracy': '0.5195', 'epoch': '0.5721'}
+{'loss': '0.752', 'grad_norm': '10.81', 'learning_rate': '4.279e-06', 'num_tokens': '2.2e+06', 'mean_token_accuracy': '0.7954', 'epoch': '0.5722'}
+{'loss': '2.012', 'grad_norm': '17.75', 'learning_rate': '4.278e-06', 'num_tokens': '2.2e+06', 'mean_token_accuracy': '0.5639', 'epoch': '0.5724'}
+{'loss': '0.8566', 'grad_norm': '25.88', 'learning_rate': '4.276e-06', 'num_tokens': '2.2e+06', 'mean_token_accuracy': '0.8011', 'epoch': '0.5726'}
+{'loss': '0.9202', 'grad_norm': '26.12', 'learning_rate': '4.274e-06', 'num_tokens': '2.2e+06', 'mean_token_accuracy': '0.7556', 'epoch': '0.5728'}
+{'loss': '1.093', 'grad_norm': '18', 'learning_rate': '4.272e-06', 'num_tokens': '2.201e+06', 'mean_token_accuracy': '0.7792', 'epoch': '0.5729'}
+{'loss': '0.6381', 'grad_norm': '7.469', 'learning_rate': '4.271e-06', 'num_tokens': '2.202e+06', 'mean_token_accuracy': '0.808', 'epoch': '0.5731'}
+{'loss': '1.717', 'grad_norm': '8', 'learning_rate': '4.269e-06', 'num_tokens': '2.204e+06', 'mean_token_accuracy': '0.6241', 'epoch': '0.5733'}
+{'loss': '1.542', 'grad_norm': '43', 'learning_rate': '4.267e-06', 'num_tokens': '2.204e+06', 'mean_token_accuracy': '0.6765', 'epoch': '0.5735'}
+{'loss': '1.811', 'grad_norm': '23.62', 'learning_rate': '4.265e-06', 'num_tokens': '2.204e+06', 'mean_token_accuracy': '0.5952', 'epoch': '0.5736'}
+{'loss': '0.9653', 'grad_norm': '21.12', 'learning_rate': '4.264e-06', 'num_tokens': '2.205e+06', 'mean_token_accuracy': '0.7829', 'epoch': '0.5738'}
+{'loss': '0.6191', 'grad_norm': '26.88', 'learning_rate': '4.262e-06', 'num_tokens': '2.205e+06', 'mean_token_accuracy': '0.8855', 'epoch': '0.574'}
+{'loss': '1.448', 'grad_norm': '23.88', 'learning_rate': '4.26e-06', 'num_tokens': '2.205e+06', 'mean_token_accuracy': '0.6938', 'epoch': '0.5742'}
+{'loss': '1.222', 'grad_norm': '12.12', 'learning_rate': '4.258e-06', 'num_tokens': '2.206e+06', 'mean_token_accuracy': '0.6426', 'epoch': '0.5743'}
+{'loss': '0.6285', 'grad_norm': '6.75', 'learning_rate': '4.257e-06', 'num_tokens': '2.208e+06', 'mean_token_accuracy': '0.814', 'epoch': '0.5745'}
+{'loss': '0.8672', 'grad_norm': '13.56', 'learning_rate': '4.255e-06', 'num_tokens': '2.208e+06', 'mean_token_accuracy': '0.7749', 'epoch': '0.5747'}
+{'loss': '0.6799', 'grad_norm': '7.438', 'learning_rate': '4.253e-06', 'num_tokens': '2.21e+06', 'mean_token_accuracy': '0.7939', 'epoch': '0.5749'}
+{'loss': '1.278', 'grad_norm': '16.5', 'learning_rate': '4.251e-06', 'num_tokens': '2.21e+06', 'mean_token_accuracy': '0.7547', 'epoch': '0.575'}
+{'loss': '1.51', 'grad_norm': '34', 'learning_rate': '4.25e-06', 'num_tokens': '2.211e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.5752'}
+{'loss': '1.349', 'grad_norm': '18.88', 'learning_rate': '4.248e-06', 'num_tokens': '2.211e+06', 'mean_token_accuracy': '0.6942', 'epoch': '0.5754'}
+{'loss': '2.285', 'grad_norm': '11.81', 'learning_rate': '4.246e-06', 'num_tokens': '2.212e+06', 'mean_token_accuracy': '0.5503', 'epoch': '0.5755'}
+{'loss': '0.598', 'grad_norm': '7.906', 'learning_rate': '4.245e-06', 'num_tokens': '2.213e+06', 'mean_token_accuracy': '0.8217', 'epoch': '0.5757'}
+{'loss': '1.672', 'grad_norm': '21', 'learning_rate': '4.243e-06', 'num_tokens': '2.213e+06', 'mean_token_accuracy': '0.672', 'epoch': '0.5759'}
+{'loss': '1.854', 'grad_norm': '13.44', 'learning_rate': '4.241e-06', 'num_tokens': '2.214e+06', 'mean_token_accuracy': '0.6281', 'epoch': '0.5761'}
+{'loss': '0.6728', 'grad_norm': '12.5', 'learning_rate': '4.239e-06', 'num_tokens': '2.214e+06', 'mean_token_accuracy': '0.8333', 'epoch': '0.5762'}
+{'loss': '0.9125', 'grad_norm': '11.31', 'learning_rate': '4.238e-06', 'num_tokens': '2.215e+06', 'mean_token_accuracy': '0.7422', 'epoch': '0.5764'}
+{'loss': '1.51', 'grad_norm': '13.19', 'learning_rate': '4.236e-06', 'num_tokens': '2.216e+06', 'mean_token_accuracy': '0.6864', 'epoch': '0.5766'}
+{'loss': '1.264', 'grad_norm': '13.56', 'learning_rate': '4.234e-06', 'num_tokens': '2.216e+06', 'mean_token_accuracy': '0.7466', 'epoch': '0.5768'}
+{'loss': '1.169', 'grad_norm': '13.25', 'learning_rate': '4.232e-06', 'num_tokens': '2.217e+06', 'mean_token_accuracy': '0.6687', 'epoch': '0.5769'}
+{'loss': '0.7592', 'grad_norm': '12.38', 'learning_rate': '4.231e-06', 'num_tokens': '2.217e+06', 'mean_token_accuracy': '0.8095', 'epoch': '0.5771'}
+{'loss': '0.774', 'grad_norm': '55', 'learning_rate': '4.229e-06', 'num_tokens': '2.217e+06', 'mean_token_accuracy': '0.7667', 'epoch': '0.5773'}
+{'loss': '1.778', 'grad_norm': '19.75', 'learning_rate': '4.227e-06', 'num_tokens': '2.218e+06', 'mean_token_accuracy': '0.6304', 'epoch': '0.5775'}
+{'loss': '1.099', 'grad_norm': '12.56', 'learning_rate': '4.225e-06', 'num_tokens': '2.218e+06', 'mean_token_accuracy': '0.6877', 'epoch': '0.5776'}
+{'loss': '1.384', 'grad_norm': '15.75', 'learning_rate': '4.224e-06', 'num_tokens': '2.219e+06', 'mean_token_accuracy': '0.6585', 'epoch': '0.5778'}
+{'loss': '1.269', 'grad_norm': '39.75', 'learning_rate': '4.222e-06', 'num_tokens': '2.219e+06', 'mean_token_accuracy': '0.699', 'epoch': '0.578'}
+{'loss': '0.566', 'grad_norm': '42.5', 'learning_rate': '4.22e-06', 'num_tokens': '2.219e+06', 'mean_token_accuracy': '0.8395', 'epoch': '0.5782'}
+{'loss': '1.412', 'grad_norm': '19.38', 'learning_rate': '4.218e-06', 'num_tokens': '2.22e+06', 'mean_token_accuracy': '0.7', 'epoch': '0.5783'}
+
 58%|█████▊    | 3331/5758 [18:56<12:50,  3.15it/s]
 58%|█████▊    | 3332/5758 [18:56<13:03,  3.10it/s]
                                                   

 58%|█████▊    | 3332/5758 [18:56<13:03,  3.10it/s]
 58%|█████▊    | 3333/5758 [18:56<13:12,  3.06it/s]
                                                   

 58%|█████▊    | 3333/5758 [18:56<13:12,  3.06it/s]
 58%|█████▊    | 3334/5758 [18:57<12:47,  3.16it/s]
                                                   

 58%|█████▊    | 3334/5758 [18:57<12:47,  3.16it/s]
 58%|█████▊    | 3335/5758 [18:57<12:41,  3.18it/s]
                                                   

 58%|█████▊    | 3335/5758 [18:57<12:41,  3.18it/s]
 58%|█████▊    | 3336/5758 [18:57<13:03,  3.09it/s]
                                                   

 58%|█████▊    | 3336/5758 [18:57<13:03,  3.09it/s]
 58%|█████▊    | 3337/5758 [18:58<12:48,  3.15it/s]
                                                   

 58%|█████▊    | 3337/5758 [18:58<12:48,  3.15it/s]
 58%|█████▊    | 3338/5758 [18:58<13:01,  3.10it/s]
                                                   

 58%|█████▊    | 3338/5758 [18:58<13:01,  3.10it/s]
 58%|█████▊    | 3339/5758 [18:58<12:43,  3.17it/s]
                                                   

 58%|█████▊    | 3339/5758 [18:58<12:43,  3.17it/s]
 58%|█████▊    | 3340/5758 [18:59<13:37,  2.96it/s]
                                                   

 58%|█████▊    | 3340/5758 [18:59<13:37,  2.96it/s]
 58%|█████▊    | 3341/5758 [18:59<13:21,  3.02it/s]
                                                   

 58%|█████▊    | 3341/5758 [18:59<13:21,  3.02it/s]
 58%|█████▊    | 3342/5758 [19:00<15:27,  2.60it/s]
                                                   

 58%|█████▊    | 3342/5758 [19:00<15:27,  2.60it/s]
 58%|█████▊    | 3343/5758 [19:00<15:27,  2.60it/s]
                                                   

 58%|█████▊    | 3343/5758 [19:00<15:27,  2.60it/s]
 58%|█████▊    | 3344/5758 [19:00<14:40,  2.74it/s]
                                                   

 58%|█████▊    | 3344/5758 [19:00<14:40,  2.74it/s]
 58%|█████▊    | 3345/5758 [19:01<14:09,  2.84it/s]
                                                   

 58%|█████▊    | 3345/5758 [19:01<14:09,  2.84it/s]
 58%|█████▊    | 3346/5758 [19:01<13:36,  2.95it/s]
                                                   

 58%|█████▊    | 3346/5758 [19:01<13:36,  2.95it/s]
 58%|█████▊    | 3347/5758 [19:01<13:16,  3.03it/s]
                                                   

 58%|█████▊    | 3347/5758 [19:01<13:16,  3.03it/s]
 58%|█████▊    | 3348/5758 [19:02<13:51,  2.90it/s]
                                                   

 58%|█████▊    | 3348/5758 [19:02<13:51,  2.90it/s]
 58%|█████▊    | 3349/5758 [19:02<14:20,  2.80it/s]
                                                   

 58%|█████▊    | 3349/5758 [19:02<14:20,  2.80it/s]
 58%|█████▊    | 3350/5758 [19:02<14:40,  2.74it/s]
                                                   

 58%|█████▊    | 3350/5758 [19:02<14:40,  2.74it/s]
 58%|█████▊    | 3351/5758 [19:03<14:46,  2.71it/s]
                                                   

 58%|█████▊    | 3351/5758 [19:03<14:46,  2.71it/s]
 58%|█████▊    | 3352/5758 [19:03<14:15,  2.81it/s]
                                                   

 58%|█████▊    | 3352/5758 [19:03<14:15,  2.81it/s]
 58%|█████▊    | 3353/5758 [19:03<13:47,  2.91it/s]
                                                   

 58%|█████▊    | 3353/5758 [19:03<13:47,  2.91it/s]
 58%|█████▊    | 3354/5758 [19:04<13:35,  2.95it/s]
                                                   

 58%|█████▊    | 3354/5758 [19:04<13:35,  2.95it/s]
 58%|█████▊    | 3355/5758 [19:04<13:23,  2.99it/s]
                                                   

 58%|█████▊    | 3355/5758 [19:04<13:23,  2.99it/s]
 58%|█████▊    | 3356/5758 [19:04<13:11,  3.04it/s]
                                                   

 58%|█████▊    | 3356/5758 [19:04<13:11,  3.04it/s]
 58%|█████▊    | 3357/5758 [19:05<13:58,  2.86it/s]
                                                   

 58%|█████▊    | 3357/5758 [19:05<13:58,  2.86it/s]
 58%|█████▊    | 3358/5758 [19:05<13:26,  2.98it/s]
                                                   

 58%|█████▊    | 3358/5758 [19:05<13:26,  2.98it/s]
 58%|█████▊    | 3359/5758 [19:05<14:00,  2.85it/s]
                                                   

 58%|█████▊    | 3359/5758 [19:05<14:00,  2.85it/s]
 58%|█████▊    | 3360/5758 [19:06<13:53,  2.88it/s]
                                                   

 58%|█████▊    | 3360/5758 [19:06<13:53,  2.88it/s]
 58%|█████▊    | 3361/5758 [19:06<13:59,  2.85it/s]
                                                   

 58%|█████▊    | 3361/5758 [19:06<13:59,  2.85it/s]
 58%|█████▊    | 3362/5758 [19:06<14:14,  2.80it/s]
                                                   

 58%|█████▊    | 3362/5758 [19:06<14:14,  2.80it/s]
 58%|█████▊    | 3363/5758 [19:07<14:15,  2.80it/s]
                                                   

 58%|█████▊    | 3363/5758 [19:07<14:15,  2.80it/s]
 58%|█████▊    | 3364/5758 [19:07<13:30,  2.95it/s]
                                                   

 58%|█████▊    | 3364/5758 [19:07<13:30,  2.95it/s]
 58%|█████▊    | 3365/5758 [19:07<12:58,  3.08it/s]
                                                   

 58%|█████▊    | 3365/5758 [19:07<12:58,  3.08it/s]
 58%|█████▊    | 3366/5758 [19:08<13:32,  2.95it/s]
                                                   

 58%|█████▊    | 3366/5758 [19:08<13:32,  2.95it/s]
 58%|█████▊    | 3367/5758 [19:08<12:59,  3.07it/s]
                                                   

 58%|█████▊    | 3367/5758 [19:08<12:59,  3.07it/s]
 58%|█████▊    | 3368/5758 [19:09<14:27,  2.75it/s]
                                                   

 58%|█████▊    | 3368/5758 [19:09<14:27,  2.75it/s]
 59%|█████▊    | 3369/5758 [19:09<13:29,  2.95it/s]
                                                   

 59%|█████▊    | 3369/5758 [19:09<13:29,  2.95it/s]
 59%|█████▊    | 3370/5758 [19:09<13:21,  2.98it/s]
                                                   

 59%|█████▊    | 3370/5758 [19:09<13:21,  2.98it/s]
 59%|█████▊    | 3371/5758 [19:09<12:59,  3.06it/s]
                                                   

 59%|█████▊    | 3371/5758 [19:09<12:59,  3.06it/s]
 59%|█████▊    | 3372/5758 [19:10<12:47,  3.11it/s]
                                                   

 59%|█████▊    | 3372/5758 [19:10<12:47,  3.11it/s]
 59%|█████▊    | 3373/5758 [19:10<12:56,  3.07it/s]
                                                   

 59%|█████▊    | 3373/5758 [19:10<12:56,  3.07it/s]
 59%|█████▊    | 3374/5758 [19:10<12:53,  3.08it/s]
                                                   

 59%|█████▊    | 3374/5758 [19:10<12:53,  3.08it/s]
 59%|█████▊    | 3375/5758 [19:11<12:33,  3.16it/s]
                                                   

 59%|█████▊    | 3375/5758 [19:11<12:33,  3.16it/s]
 59%|█████▊    | 3376/5758 [19:11<13:20,  2.98it/s]
                                                   

 59%|█████▊    | 3376/5758 [19:11<13:20,  2.98it/s]
 59%|█████▊    | 3377/5758 [19:11<13:10,  3.01it/s]
                                                   

 59%|█████▊    | 3377/5758 [19:11<13:10,  3.01it/s]
 59%|█████▊    | 3378/5758 [19:12<12:52,  3.08it/s]
                                                   

 59%|█████▊    | 3378/5758 [19:12<12:52,  3.08it/s]
 59%|█████▊    | 3379/5758 [19:12<13:28,  2.94it/s]
                                                   

 59%|█████▊    | 3379/5758 [19:12<13:28,  2.94it/s]
 59%|█████▊    | 3380/5758 [19:12<12:39,  3.13it/s]
                                                   

 59%|█████▊    | 3380/5758 [19:12<12:39,  3.13it/s]
 59%|█████▊    | 3381/5758 [19:13<12:52,  3.08it/s]
                                                   

 59%|█████▊    | 3381/5758 [19:13<12:52,  3.08it/s]
 59%|█████▊    | 3382/5758 [19:13<12:47,  3.10it/s]
                                                   

 59%|█████▊    | 3382/5758 [19:13<12:47,  3.10it/s]
 59%|█████▉    | 3383/5758 [19:13<13:26,  2.94it/s]
                                                   

 59%|█████▉    | 3383/5758 [19:13<13:26,  2.94it/s]
 59%|█████▉    | 3384/5758 [19:14<12:59,  3.05it/s]
                                                   

 59%|█████▉    | 3384/5758 [19:14<12:59,  3.05it/s]
 59%|█████▉    | 3385/5758 [19:14<12:36,  3.14it/s]
                                                   

 59%|█████▉    | 3385/5758 [19:14<12:36,  3.14it/s]
 59%|█████▉    | 3386/5758 [19:14<13:24,  2.95it/s]
                                                   
{'loss': '1.456', 'grad_norm': '12.94', 'learning_rate': '4.217e-06', 'num_tokens': '2.22e+06', 'mean_token_accuracy': '0.6645', 'epoch': '0.5785'}
+{'loss': '1.784', 'grad_norm': '16.62', 'learning_rate': '4.215e-06', 'num_tokens': '2.221e+06', 'mean_token_accuracy': '0.5812', 'epoch': '0.5787'}
+{'loss': '1.311', 'grad_norm': '6.438', 'learning_rate': '4.213e-06', 'num_tokens': '2.224e+06', 'mean_token_accuracy': '0.7089', 'epoch': '0.5788'}
+{'loss': '0.6048', 'grad_norm': '14.81', 'learning_rate': '4.212e-06', 'num_tokens': '2.224e+06', 'mean_token_accuracy': '0.8645', 'epoch': '0.579'}
+{'loss': '1.241', 'grad_norm': '37.25', 'learning_rate': '4.21e-06', 'num_tokens': '2.225e+06', 'mean_token_accuracy': '0.7125', 'epoch': '0.5792'}
+{'loss': '1.594', 'grad_norm': '17.75', 'learning_rate': '4.208e-06', 'num_tokens': '2.225e+06', 'mean_token_accuracy': '0.6941', 'epoch': '0.5794'}
+{'loss': '1.013', 'grad_norm': '16.25', 'learning_rate': '4.206e-06', 'num_tokens': '2.225e+06', 'mean_token_accuracy': '0.7585', 'epoch': '0.5795'}
+{'loss': '1.331', 'grad_norm': '34.25', 'learning_rate': '4.205e-06', 'num_tokens': '2.225e+06', 'mean_token_accuracy': '0.7347', 'epoch': '0.5797'}
+{'loss': '1.184', 'grad_norm': '15.62', 'learning_rate': '4.203e-06', 'num_tokens': '2.226e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.5799'}
+{'loss': '1.119', 'grad_norm': '12.75', 'learning_rate': '4.201e-06', 'num_tokens': '2.227e+06', 'mean_token_accuracy': '0.7143', 'epoch': '0.5801'}
+{'loss': '0.6106', 'grad_norm': '7.125', 'learning_rate': '4.199e-06', 'num_tokens': '2.228e+06', 'mean_token_accuracy': '0.809', 'epoch': '0.5802'}
+{'loss': '2.359', 'grad_norm': '10.25', 'learning_rate': '4.198e-06', 'num_tokens': '2.229e+06', 'mean_token_accuracy': '0.5712', 'epoch': '0.5804'}
+{'loss': '1.333', 'grad_norm': '14.06', 'learning_rate': '4.196e-06', 'num_tokens': '2.23e+06', 'mean_token_accuracy': '0.6367', 'epoch': '0.5806'}
+{'loss': '1.458', 'grad_norm': '7.875', 'learning_rate': '4.194e-06', 'num_tokens': '2.232e+06', 'mean_token_accuracy': '0.6863', 'epoch': '0.5808'}
+{'loss': '1.387', 'grad_norm': '12.12', 'learning_rate': '4.192e-06', 'num_tokens': '2.232e+06', 'mean_token_accuracy': '0.6308', 'epoch': '0.5809'}
+{'loss': '0.8561', 'grad_norm': '21.5', 'learning_rate': '4.191e-06', 'num_tokens': '2.232e+06', 'mean_token_accuracy': '0.7853', 'epoch': '0.5811'}
+{'loss': '1.131', 'grad_norm': '12.62', 'learning_rate': '4.189e-06', 'num_tokens': '2.233e+06', 'mean_token_accuracy': '0.6838', 'epoch': '0.5813'}
+{'loss': '2.303', 'grad_norm': '49.5', 'learning_rate': '4.187e-06', 'num_tokens': '2.233e+06', 'mean_token_accuracy': '0.625', 'epoch': '0.5815'}
+{'loss': '1.38', 'grad_norm': '36.25', 'learning_rate': '4.185e-06', 'num_tokens': '2.233e+06', 'mean_token_accuracy': '0.7405', 'epoch': '0.5816'}
+{'loss': '1.223', 'grad_norm': '13.31', 'learning_rate': '4.184e-06', 'num_tokens': '2.234e+06', 'mean_token_accuracy': '0.6886', 'epoch': '0.5818'}
+{'loss': '1.467', 'grad_norm': '30.12', 'learning_rate': '4.182e-06', 'num_tokens': '2.234e+06', 'mean_token_accuracy': '0.7333', 'epoch': '0.582'}
+{'loss': '1.168', 'grad_norm': '16.75', 'learning_rate': '4.18e-06', 'num_tokens': '2.235e+06', 'mean_token_accuracy': '0.7268', 'epoch': '0.5821'}
+{'loss': '1.058', 'grad_norm': '9.25', 'learning_rate': '4.179e-06', 'num_tokens': '2.236e+06', 'mean_token_accuracy': '0.7113', 'epoch': '0.5823'}
+{'loss': '0.5189', 'grad_norm': '30.5', 'learning_rate': '4.177e-06', 'num_tokens': '2.236e+06', 'mean_token_accuracy': '0.9059', 'epoch': '0.5825'}
+{'loss': '1.399', 'grad_norm': '12.88', 'learning_rate': '4.175e-06', 'num_tokens': '2.237e+06', 'mean_token_accuracy': '0.631', 'epoch': '0.5827'}
+{'loss': '1.334', 'grad_norm': '22', 'learning_rate': '4.173e-06', 'num_tokens': '2.237e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.5828'}
+{'loss': '0.9495', 'grad_norm': '13.5', 'learning_rate': '4.172e-06', 'num_tokens': '2.238e+06', 'mean_token_accuracy': '0.7195', 'epoch': '0.583'}
+{'loss': '1.463', 'grad_norm': '15.19', 'learning_rate': '4.17e-06', 'num_tokens': '2.238e+06', 'mean_token_accuracy': '0.6378', 'epoch': '0.5832'}
+{'loss': '1.695', 'grad_norm': '25.25', 'learning_rate': '4.168e-06', 'num_tokens': '2.239e+06', 'mean_token_accuracy': '0.5792', 'epoch': '0.5834'}
+{'loss': '0.9924', 'grad_norm': '14.19', 'learning_rate': '4.166e-06', 'num_tokens': '2.239e+06', 'mean_token_accuracy': '0.7827', 'epoch': '0.5835'}
+{'loss': '2.234', 'grad_norm': '11.06', 'learning_rate': '4.165e-06', 'num_tokens': '2.24e+06', 'mean_token_accuracy': '0.5323', 'epoch': '0.5837'}
+{'loss': '0.7914', 'grad_norm': '15.25', 'learning_rate': '4.163e-06', 'num_tokens': '2.24e+06', 'mean_token_accuracy': '0.8309', 'epoch': '0.5839'}
+{'loss': '1.873', 'grad_norm': '14.19', 'learning_rate': '4.161e-06', 'num_tokens': '2.241e+06', 'mean_token_accuracy': '0.6112', 'epoch': '0.5841'}
+{'loss': '1.182', 'grad_norm': '47.5', 'learning_rate': '4.159e-06', 'num_tokens': '2.241e+06', 'mean_token_accuracy': '0.726', 'epoch': '0.5842'}
+{'loss': '1.067', 'grad_norm': '13.06', 'learning_rate': '4.158e-06', 'num_tokens': '2.242e+06', 'mean_token_accuracy': '0.6854', 'epoch': '0.5844'}
+{'loss': '1.129', 'grad_norm': '13.75', 'learning_rate': '4.156e-06', 'num_tokens': '2.242e+06', 'mean_token_accuracy': '0.7027', 'epoch': '0.5846'}
+{'loss': '1.171', 'grad_norm': '13.25', 'learning_rate': '4.154e-06', 'num_tokens': '2.243e+06', 'mean_token_accuracy': '0.6891', 'epoch': '0.5848'}
+{'loss': '1.319', 'grad_norm': '7.906', 'learning_rate': '4.152e-06', 'num_tokens': '2.244e+06', 'mean_token_accuracy': '0.7058', 'epoch': '0.5849'}
+{'loss': '2.011', 'grad_norm': '15.44', 'learning_rate': '4.151e-06', 'num_tokens': '2.245e+06', 'mean_token_accuracy': '0.5963', 'epoch': '0.5851'}
+{'loss': '1.147', 'grad_norm': '10.88', 'learning_rate': '4.149e-06', 'num_tokens': '2.246e+06', 'mean_token_accuracy': '0.6802', 'epoch': '0.5853'}
+{'loss': '1.147', 'grad_norm': '12', 'learning_rate': '4.147e-06', 'num_tokens': '2.247e+06', 'mean_token_accuracy': '0.6594', 'epoch': '0.5854'}
+{'loss': '1.425', 'grad_norm': '37.5', 'learning_rate': '4.146e-06', 'num_tokens': '2.247e+06', 'mean_token_accuracy': '0.6627', 'epoch': '0.5856'}
+{'loss': '1.002', 'grad_norm': '20.75', 'learning_rate': '4.144e-06', 'num_tokens': '2.247e+06', 'mean_token_accuracy': '0.7937', 'epoch': '0.5858'}
+{'loss': '1.096', 'grad_norm': '10.25', 'learning_rate': '4.142e-06', 'num_tokens': '2.248e+06', 'mean_token_accuracy': '0.7478', 'epoch': '0.586'}
+{'loss': '1.135', 'grad_norm': '30.75', 'learning_rate': '4.14e-06', 'num_tokens': '2.248e+06', 'mean_token_accuracy': '0.7246', 'epoch': '0.5861'}
+{'loss': '1.349', 'grad_norm': '14.5', 'learning_rate': '4.139e-06', 'num_tokens': '2.249e+06', 'mean_token_accuracy': '0.6294', 'epoch': '0.5863'}
+{'loss': '2.251', 'grad_norm': '16.75', 'learning_rate': '4.137e-06', 'num_tokens': '2.249e+06', 'mean_token_accuracy': '0.5316', 'epoch': '0.5865'}
+{'loss': '1.128', 'grad_norm': '12.25', 'learning_rate': '4.135e-06', 'num_tokens': '2.25e+06', 'mean_token_accuracy': '0.7011', 'epoch': '0.5867'}
+{'loss': '1.472', 'grad_norm': '13.81', 'learning_rate': '4.133e-06', 'num_tokens': '2.25e+06', 'mean_token_accuracy': '0.6895', 'epoch': '0.5868'}
+{'loss': '0.6953', 'grad_norm': '39.5', 'learning_rate': '4.132e-06', 'num_tokens': '2.25e+06', 'mean_token_accuracy': '0.8793', 'epoch': '0.587'}
+{'loss': '1.757', 'grad_norm': '20.12', 'learning_rate': '4.13e-06', 'num_tokens': '2.251e+06', 'mean_token_accuracy': '0.6185', 'epoch': '0.5872'}
+{'loss': '0.6964', 'grad_norm': '13.38', 'learning_rate': '4.128e-06', 'num_tokens': '2.251e+06', 'mean_token_accuracy': '0.8444', 'epoch': '0.5874'}
+{'loss': '1.055', 'grad_norm': '18.25', 'learning_rate': '4.126e-06', 'num_tokens': '2.251e+06', 'mean_token_accuracy': '0.7411', 'epoch': '0.5875'}
+{'loss': '1.834', 'grad_norm': '26.88', 'learning_rate': '4.125e-06', 'num_tokens': '2.252e+06', 'mean_token_accuracy': '0.6235', 'epoch': '0.5877'}
+{'loss': '2.179', 'grad_norm': '12.5', 'learning_rate': '4.123e-06', 'num_tokens': '2.252e+06', 'mean_token_accuracy': '0.5373', 'epoch': '0.5879'}
+
 59%|█████▉    | 3386/5758 [19:14<13:24,  2.95it/s]
 59%|█████▉    | 3387/5758 [19:15<13:56,  2.83it/s]
                                                   

 59%|█████▉    | 3387/5758 [19:15<13:56,  2.83it/s]
 59%|█████▉    | 3388/5758 [19:15<13:17,  2.97it/s]
                                                   

 59%|█████▉    | 3388/5758 [19:15<13:17,  2.97it/s]
 59%|█████▉    | 3389/5758 [19:15<13:53,  2.84it/s]
                                                   

 59%|█████▉    | 3389/5758 [19:15<13:53,  2.84it/s]
 59%|█████▉    | 3390/5758 [19:16<13:32,  2.92it/s]
                                                   

 59%|█████▉    | 3390/5758 [19:16<13:32,  2.92it/s]
 59%|█████▉    | 3391/5758 [19:16<13:27,  2.93it/s]
                                                   

 59%|█████▉    | 3391/5758 [19:16<13:27,  2.93it/s]
 59%|█████▉    | 3392/5758 [19:16<12:58,  3.04it/s]
                                                   

 59%|█████▉    | 3392/5758 [19:16<12:58,  3.04it/s]
 59%|█████▉    | 3393/5758 [19:17<13:01,  3.03it/s]
                                                   

 59%|█████▉    | 3393/5758 [19:17<13:01,  3.03it/s]
 59%|█████▉    | 3394/5758 [19:17<14:17,  2.76it/s]
                                                   

 59%|█████▉    | 3394/5758 [19:17<14:17,  2.76it/s]
 59%|█████▉    | 3395/5758 [19:17<13:26,  2.93it/s]
                                                   

 59%|█████▉    | 3395/5758 [19:18<13:26,  2.93it/s]
 59%|█████▉    | 3396/5758 [19:18<13:55,  2.83it/s]
                                                   

 59%|█████▉    | 3396/5758 [19:18<13:55,  2.83it/s]
 59%|█████▉    | 3397/5758 [19:18<13:26,  2.93it/s]
                                                   

 59%|█████▉    | 3397/5758 [19:18<13:26,  2.93it/s]
 59%|█████▉    | 3398/5758 [19:18<12:55,  3.04it/s]
                                                   

 59%|█████▉    | 3398/5758 [19:18<12:55,  3.04it/s]
 59%|█████▉    | 3399/5758 [19:19<13:03,  3.01it/s]
                                                   

 59%|█████▉    | 3399/5758 [19:19<13:03,  3.01it/s]
 59%|█████▉    | 3400/5758 [19:19<12:35,  3.12it/s]
                                                   

 59%|█████▉    | 3400/5758 [19:19<12:35,  3.12it/s]
 59%|█████▉    | 3401/5758 [19:19<12:15,  3.20it/s]
                                                   

 59%|█████▉    | 3401/5758 [19:19<12:15,  3.20it/s]
 59%|█████▉    | 3402/5758 [19:20<12:44,  3.08it/s]
                                                   

 59%|█████▉    | 3402/5758 [19:20<12:44,  3.08it/s]
 59%|█████▉    | 3403/5758 [19:20<12:59,  3.02it/s]
                                                   

 59%|█████▉    | 3403/5758 [19:20<12:59,  3.02it/s]
 59%|█████▉    | 3404/5758 [19:21<13:35,  2.89it/s]
                                                   

 59%|█████▉    | 3404/5758 [19:21<13:35,  2.89it/s]
 59%|█████▉    | 3405/5758 [19:21<14:28,  2.71it/s]
                                                   

 59%|█████▉    | 3405/5758 [19:21<14:28,  2.71it/s]
 59%|█████▉    | 3406/5758 [19:21<14:18,  2.74it/s]
                                                   

 59%|█████▉    | 3406/5758 [19:21<14:18,  2.74it/s]
 59%|█████▉    | 3407/5758 [19:22<13:38,  2.87it/s]
                                                   

 59%|█████▉    | 3407/5758 [19:22<13:38,  2.87it/s]
 59%|█████▉    | 3408/5758 [19:22<14:17,  2.74it/s]
                                                   

 59%|█████▉    | 3408/5758 [19:22<14:17,  2.74it/s]
 59%|█████▉    | 3409/5758 [19:22<13:44,  2.85it/s]
                                                   

 59%|█████▉    | 3409/5758 [19:22<13:44,  2.85it/s]
 59%|█████▉    | 3410/5758 [19:23<13:15,  2.95it/s]
                                                   

 59%|█████▉    | 3410/5758 [19:23<13:15,  2.95it/s]
 59%|█████▉    | 3411/5758 [19:23<13:17,  2.94it/s]
                                                   

 59%|█████▉    | 3411/5758 [19:23<13:17,  2.94it/s]
 59%|█████▉    | 3412/5758 [19:23<12:57,  3.02it/s]
                                                   

 59%|█████▉    | 3412/5758 [19:23<12:57,  3.02it/s]
 59%|█████▉    | 3413/5758 [19:24<13:19,  2.93it/s]
                                                   

 59%|█████▉    | 3413/5758 [19:24<13:19,  2.93it/s]
 59%|█████▉    | 3414/5758 [19:24<12:52,  3.03it/s]
                                                   

 59%|█████▉    | 3414/5758 [19:24<12:52,  3.03it/s]
 59%|█████▉    | 3415/5758 [19:24<12:53,  3.03it/s]
                                                   

 59%|█████▉    | 3415/5758 [19:24<12:53,  3.03it/s]
 59%|█████▉    | 3416/5758 [19:25<13:06,  2.98it/s]
                                                   

 59%|█████▉    | 3416/5758 [19:25<13:06,  2.98it/s]
 59%|█████▉    | 3417/5758 [19:25<14:05,  2.77it/s]
                                                   

 59%|█████▉    | 3417/5758 [19:25<14:05,  2.77it/s]
 59%|█████▉    | 3418/5758 [19:25<15:07,  2.58it/s]
                                                   

 59%|█████▉    | 3418/5758 [19:25<15:07,  2.58it/s]
 59%|█████▉    | 3419/5758 [19:26<14:15,  2.73it/s]
                                                   

 59%|█████▉    | 3419/5758 [19:26<14:15,  2.73it/s]
 59%|█████▉    | 3420/5758 [19:26<13:23,  2.91it/s]
                                                   

 59%|█████▉    | 3420/5758 [19:26<13:23,  2.91it/s]
 59%|█████▉    | 3421/5758 [19:26<13:27,  2.90it/s]
                                                   

 59%|█████▉    | 3421/5758 [19:26<13:27,  2.90it/s]
 59%|█████▉    | 3422/5758 [19:27<13:14,  2.94it/s]
                                                   

 59%|█████▉    | 3422/5758 [19:27<13:14,  2.94it/s]
 59%|█████▉    | 3423/5758 [19:27<14:28,  2.69it/s]
                                                   

 59%|█████▉    | 3423/5758 [19:27<14:28,  2.69it/s]
 59%|█████▉    | 3424/5758 [19:28<13:32,  2.87it/s]
                                                   

 59%|█████▉    | 3424/5758 [19:28<13:32,  2.87it/s]
 59%|█████▉    | 3425/5758 [19:28<13:44,  2.83it/s]
                                                   

 59%|█████▉    | 3425/5758 [19:28<13:44,  2.83it/s]
 59%|█████▉    | 3426/5758 [19:28<13:12,  2.94it/s]
                                                   

 59%|█████▉    | 3426/5758 [19:28<13:12,  2.94it/s]
 60%|█████▉    | 3427/5758 [19:29<12:55,  3.01it/s]
                                                   

 60%|█████▉    | 3427/5758 [19:29<12:55,  3.01it/s]
 60%|█████▉    | 3428/5758 [19:29<12:27,  3.12it/s]
                                                   

 60%|█████▉    | 3428/5758 [19:29<12:27,  3.12it/s]
 60%|█████▉    | 3429/5758 [19:29<12:49,  3.03it/s]
                                                   

 60%|█████▉    | 3429/5758 [19:29<12:49,  3.03it/s]
 60%|█████▉    | 3430/5758 [19:30<13:31,  2.87it/s]
                                                   

 60%|█████▉    | 3430/5758 [19:30<13:31,  2.87it/s]
 60%|█████▉    | 3431/5758 [19:30<13:56,  2.78it/s]
                                                   

 60%|█████▉    | 3431/5758 [19:30<13:56,  2.78it/s]
 60%|█████▉    | 3432/5758 [19:30<13:50,  2.80it/s]
                                                   

 60%|█████▉    | 3432/5758 [19:30<13:50,  2.80it/s]
 60%|█████▉    | 3433/5758 [19:31<13:19,  2.91it/s]
                                                   

 60%|█████▉    | 3433/5758 [19:31<13:19,  2.91it/s]
 60%|█████▉    | 3434/5758 [19:31<12:49,  3.02it/s]
                                                   

 60%|█████▉    | 3434/5758 [19:31<12:49,  3.02it/s]
 60%|█████▉    | 3435/5758 [19:31<13:17,  2.91it/s]
                                                   

 60%|█████▉    | 3435/5758 [19:31<13:17,  2.91it/s]
 60%|█████▉    | 3436/5758 [19:32<12:46,  3.03it/s]
                                                   

 60%|█████▉    | 3436/5758 [19:32<12:46,  3.03it/s]
 60%|█████▉    | 3437/5758 [19:32<12:20,  3.14it/s]
                                                   

 60%|█████▉    | 3437/5758 [19:32<12:20,  3.14it/s]
 60%|█████▉    | 3438/5758 [19:32<12:04,  3.20it/s]
                                                   

 60%|█████▉    | 3438/5758 [19:32<12:04,  3.20it/s]
 60%|█████▉    | 3439/5758 [19:32<12:23,  3.12it/s]
                                                   

 60%|█████▉    | 3439/5758 [19:32<12:23,  3.12it/s]
 60%|█████▉    | 3440/5758 [19:33<12:04,  3.20it/s]
                                                   

 60%|█████▉    | 3440/5758 [19:33<12:04,  3.20it/s]
 60%|█████▉    | 3441/5758 [19:33<12:21,  3.13it/s]
                                                   
{'loss': '2.239', 'grad_norm': '32.5', 'learning_rate': '4.121e-06', 'num_tokens': '2.252e+06', 'mean_token_accuracy': '0.6', 'epoch': '0.5881'}
+{'loss': '1.372', 'grad_norm': '19.5', 'learning_rate': '4.119e-06', 'num_tokens': '2.253e+06', 'mean_token_accuracy': '0.6748', 'epoch': '0.5882'}
+{'loss': '0.6493', 'grad_norm': '8.688', 'learning_rate': '4.118e-06', 'num_tokens': '2.254e+06', 'mean_token_accuracy': '0.8112', 'epoch': '0.5884'}
+{'loss': '2.072', 'grad_norm': '23.62', 'learning_rate': '4.116e-06', 'num_tokens': '2.254e+06', 'mean_token_accuracy': '0.5769', 'epoch': '0.5886'}
+{'loss': '1.161', 'grad_norm': '12.5', 'learning_rate': '4.114e-06', 'num_tokens': '2.255e+06', 'mean_token_accuracy': '0.6734', 'epoch': '0.5887'}
+{'loss': '1.012', 'grad_norm': '19', 'learning_rate': '4.113e-06', 'num_tokens': '2.255e+06', 'mean_token_accuracy': '0.7788', 'epoch': '0.5889'}
+{'loss': '0.6242', 'grad_norm': '8.188', 'learning_rate': '4.111e-06', 'num_tokens': '2.256e+06', 'mean_token_accuracy': '0.8099', 'epoch': '0.5891'}
+{'loss': '1.228', 'grad_norm': '12.38', 'learning_rate': '4.109e-06', 'num_tokens': '2.257e+06', 'mean_token_accuracy': '0.6543', 'epoch': '0.5893'}
+{'loss': '1.224', 'grad_norm': '12.69', 'learning_rate': '4.107e-06', 'num_tokens': '2.258e+06', 'mean_token_accuracy': '0.6676', 'epoch': '0.5894'}
+{'loss': '1.088', 'grad_norm': '37.5', 'learning_rate': '4.106e-06', 'num_tokens': '2.258e+06', 'mean_token_accuracy': '0.7846', 'epoch': '0.5896'}
+{'loss': '1.094', 'grad_norm': '13.62', 'learning_rate': '4.104e-06', 'num_tokens': '2.259e+06', 'mean_token_accuracy': '0.6835', 'epoch': '0.5898'}
+{'loss': '1.297', 'grad_norm': '12.75', 'learning_rate': '4.102e-06', 'num_tokens': '2.259e+06', 'mean_token_accuracy': '0.6522', 'epoch': '0.59'}
+{'loss': '1.081', 'grad_norm': '12.81', 'learning_rate': '4.1e-06', 'num_tokens': '2.26e+06', 'mean_token_accuracy': '0.7052', 'epoch': '0.5901'}
+{'loss': '1.66', 'grad_norm': '21.5', 'learning_rate': '4.099e-06', 'num_tokens': '2.261e+06', 'mean_token_accuracy': '0.6512', 'epoch': '0.5903'}
+{'loss': '1.758', 'grad_norm': '12.56', 'learning_rate': '4.097e-06', 'num_tokens': '2.261e+06', 'mean_token_accuracy': '0.6414', 'epoch': '0.5905'}
+{'loss': '1.07', 'grad_norm': '25.88', 'learning_rate': '4.095e-06', 'num_tokens': '2.261e+06', 'mean_token_accuracy': '0.7532', 'epoch': '0.5907'}
+{'loss': '1.817', 'grad_norm': '14.56', 'learning_rate': '4.093e-06', 'num_tokens': '2.262e+06', 'mean_token_accuracy': '0.6426', 'epoch': '0.5908'}
+{'loss': '1.242', 'grad_norm': '11.19', 'learning_rate': '4.092e-06', 'num_tokens': '2.263e+06', 'mean_token_accuracy': '0.666', 'epoch': '0.591'}
+{'loss': '1.119', 'grad_norm': '6', 'learning_rate': '4.09e-06', 'num_tokens': '2.265e+06', 'mean_token_accuracy': '0.7335', 'epoch': '0.5912'}
+{'loss': '1.723', 'grad_norm': '9.938', 'learning_rate': '4.088e-06', 'num_tokens': '2.267e+06', 'mean_token_accuracy': '0.6078', 'epoch': '0.5914'}
+{'loss': '1.605', 'grad_norm': '14.5', 'learning_rate': '4.086e-06', 'num_tokens': '2.267e+06', 'mean_token_accuracy': '0.6483', 'epoch': '0.5915'}
+{'loss': '2.174', 'grad_norm': '11.31', 'learning_rate': '4.085e-06', 'num_tokens': '2.268e+06', 'mean_token_accuracy': '0.5577', 'epoch': '0.5917'}
+{'loss': '1.417', 'grad_norm': '29.75', 'learning_rate': '4.083e-06', 'num_tokens': '2.268e+06', 'mean_token_accuracy': '0.7737', 'epoch': '0.5919'}
+{'loss': '1.561', 'grad_norm': '18.25', 'learning_rate': '4.081e-06', 'num_tokens': '2.268e+06', 'mean_token_accuracy': '0.6919', 'epoch': '0.592'}
+{'loss': '0.9144', 'grad_norm': '14.75', 'learning_rate': '4.08e-06', 'num_tokens': '2.269e+06', 'mean_token_accuracy': '0.7826', 'epoch': '0.5922'}
+{'loss': '1.857', 'grad_norm': '7.406', 'learning_rate': '4.078e-06', 'num_tokens': '2.27e+06', 'mean_token_accuracy': '0.6224', 'epoch': '0.5924'}
+{'loss': '1.026', 'grad_norm': '13.19', 'learning_rate': '4.076e-06', 'num_tokens': '2.271e+06', 'mean_token_accuracy': '0.7123', 'epoch': '0.5926'}
+{'loss': '0.8921', 'grad_norm': '9.875', 'learning_rate': '4.074e-06', 'num_tokens': '2.272e+06', 'mean_token_accuracy': '0.7754', 'epoch': '0.5927'}
+{'loss': '1.088', 'grad_norm': '38', 'learning_rate': '4.073e-06', 'num_tokens': '2.272e+06', 'mean_token_accuracy': '0.7361', 'epoch': '0.5929'}
+{'loss': '1.334', 'grad_norm': '12.31', 'learning_rate': '4.071e-06', 'num_tokens': '2.273e+06', 'mean_token_accuracy': '0.6571', 'epoch': '0.5931'}
+{'loss': '1.769', 'grad_norm': '24.75', 'learning_rate': '4.069e-06', 'num_tokens': '2.273e+06', 'mean_token_accuracy': '0.6029', 'epoch': '0.5933'}
+{'loss': '0.6444', 'grad_norm': '6.812', 'learning_rate': '4.067e-06', 'num_tokens': '2.275e+06', 'mean_token_accuracy': '0.8152', 'epoch': '0.5934'}
+{'loss': '1.937', 'grad_norm': '16.75', 'learning_rate': '4.066e-06', 'num_tokens': '2.276e+06', 'mean_token_accuracy': '0.5993', 'epoch': '0.5936'}
+{'loss': '1.225', 'grad_norm': '11.56', 'learning_rate': '4.064e-06', 'num_tokens': '2.276e+06', 'mean_token_accuracy': '0.6756', 'epoch': '0.5938'}
+{'loss': '0.3456', 'grad_norm': '25.38', 'learning_rate': '4.062e-06', 'num_tokens': '2.276e+06', 'mean_token_accuracy': '0.9073', 'epoch': '0.594'}
+{'loss': '0.6607', 'grad_norm': '12.75', 'learning_rate': '4.06e-06', 'num_tokens': '2.277e+06', 'mean_token_accuracy': '0.8202', 'epoch': '0.5941'}
+{'loss': '1.24', 'grad_norm': '12.62', 'learning_rate': '4.059e-06', 'num_tokens': '2.278e+06', 'mean_token_accuracy': '0.6691', 'epoch': '0.5943'}
+{'loss': '0.9127', 'grad_norm': '11', 'learning_rate': '4.057e-06', 'num_tokens': '2.279e+06', 'mean_token_accuracy': '0.7449', 'epoch': '0.5945'}
+{'loss': '1.201', 'grad_norm': '13.19', 'learning_rate': '4.055e-06', 'num_tokens': '2.279e+06', 'mean_token_accuracy': '0.6798', 'epoch': '0.5947'}
+{'loss': '1.129', 'grad_norm': '6.969', 'learning_rate': '4.053e-06', 'num_tokens': '2.281e+06', 'mean_token_accuracy': '0.7267', 'epoch': '0.5948'}
+{'loss': '1.209', 'grad_norm': '13', 'learning_rate': '4.052e-06', 'num_tokens': '2.282e+06', 'mean_token_accuracy': '0.6514', 'epoch': '0.595'}
+{'loss': '0.9389', 'grad_norm': '12.25', 'learning_rate': '4.05e-06', 'num_tokens': '2.283e+06', 'mean_token_accuracy': '0.7324', 'epoch': '0.5952'}
+{'loss': '1.366', 'grad_norm': '14', 'learning_rate': '4.048e-06', 'num_tokens': '2.283e+06', 'mean_token_accuracy': '0.6437', 'epoch': '0.5953'}
+{'loss': '0.8595', 'grad_norm': '14.19', 'learning_rate': '4.047e-06', 'num_tokens': '2.284e+06', 'mean_token_accuracy': '0.7924', 'epoch': '0.5955'}
+{'loss': '0.9611', 'grad_norm': '12.69', 'learning_rate': '4.045e-06', 'num_tokens': '2.284e+06', 'mean_token_accuracy': '0.7429', 'epoch': '0.5957'}
+{'loss': '2.25', 'grad_norm': '18.62', 'learning_rate': '4.043e-06', 'num_tokens': '2.285e+06', 'mean_token_accuracy': '0.5533', 'epoch': '0.5959'}
+{'loss': '0.6647', 'grad_norm': '12.56', 'learning_rate': '4.041e-06', 'num_tokens': '2.285e+06', 'mean_token_accuracy': '0.8272', 'epoch': '0.596'}
+{'loss': '0.4988', 'grad_norm': '20.62', 'learning_rate': '4.04e-06', 'num_tokens': '2.286e+06', 'mean_token_accuracy': '0.8901', 'epoch': '0.5962'}
+{'loss': '1.954', 'grad_norm': '14', 'learning_rate': '4.038e-06', 'num_tokens': '2.286e+06', 'mean_token_accuracy': '0.5904', 'epoch': '0.5964'}
+{'loss': '1.511', 'grad_norm': '10.25', 'learning_rate': '4.036e-06', 'num_tokens': '2.287e+06', 'mean_token_accuracy': '0.6695', 'epoch': '0.5966'}
+{'loss': '1.553', 'grad_norm': '21.12', 'learning_rate': '4.034e-06', 'num_tokens': '2.288e+06', 'mean_token_accuracy': '0.6933', 'epoch': '0.5967'}
+{'loss': '2.19', 'grad_norm': '13.69', 'learning_rate': '4.033e-06', 'num_tokens': '2.288e+06', 'mean_token_accuracy': '0.5128', 'epoch': '0.5969'}
+{'loss': '1.275', 'grad_norm': '13.38', 'learning_rate': '4.031e-06', 'num_tokens': '2.289e+06', 'mean_token_accuracy': '0.6511', 'epoch': '0.5971'}
+{'loss': '1.839', 'grad_norm': '11.69', 'learning_rate': '4.029e-06', 'num_tokens': '2.29e+06', 'mean_token_accuracy': '0.6009', 'epoch': '0.5973'}
+{'loss': '0.6732', 'grad_norm': '15.31', 'learning_rate': '4.027e-06', 'num_tokens': '2.29e+06', 'mean_token_accuracy': '0.8447', 'epoch': '0.5974'}
+
 60%|█████▉    | 3441/5758 [19:33<12:21,  3.13it/s]
 60%|█████▉    | 3442/5758 [19:33<12:21,  3.12it/s]
                                                   

 60%|█████▉    | 3442/5758 [19:33<12:21,  3.12it/s]
 60%|█████▉    | 3443/5758 [19:34<12:06,  3.18it/s]
                                                   

 60%|█████▉    | 3443/5758 [19:34<12:06,  3.18it/s]
 60%|█████▉    | 3444/5758 [19:34<11:53,  3.24it/s]
                                                   

 60%|█████▉    | 3444/5758 [19:34<11:53,  3.24it/s]
 60%|█████▉    | 3445/5758 [19:34<12:18,  3.13it/s]
                                                   

 60%|█████▉    | 3445/5758 [19:34<12:18,  3.13it/s]
 60%|█████▉    | 3446/5758 [19:35<11:55,  3.23it/s]
                                                   

 60%|█████▉    | 3446/5758 [19:35<11:55,  3.23it/s]
 60%|█████▉    | 3447/5758 [19:35<11:55,  3.23it/s]
                                                   

 60%|█████▉    | 3447/5758 [19:35<11:55,  3.23it/s]
 60%|█████▉    | 3448/5758 [19:35<12:12,  3.15it/s]
                                                   

 60%|█████▉    | 3448/5758 [19:35<12:12,  3.15it/s]
 60%|█████▉    | 3449/5758 [19:36<13:25,  2.87it/s]
                                                   

 60%|█████▉    | 3449/5758 [19:36<13:25,  2.87it/s]
 60%|█████▉    | 3450/5758 [19:36<13:50,  2.78it/s]
                                                   

 60%|█████▉    | 3450/5758 [19:36<13:50,  2.78it/s]
 60%|█████▉    | 3451/5758 [19:37<14:25,  2.67it/s]
                                                   

 60%|█████▉    | 3451/5758 [19:37<14:25,  2.67it/s]
 60%|█████▉    | 3452/5758 [19:37<13:27,  2.85it/s]
                                                   

 60%|█████▉    | 3452/5758 [19:37<13:27,  2.85it/s]
 60%|█████▉    | 3453/5758 [19:37<13:20,  2.88it/s]
                                                   

 60%|█████▉    | 3453/5758 [19:37<13:20,  2.88it/s]
 60%|█████▉    | 3454/5758 [19:37<12:53,  2.98it/s]
                                                   

 60%|█████▉    | 3454/5758 [19:37<12:53,  2.98it/s]
 60%|██████    | 3455/5758 [19:38<12:30,  3.07it/s]
                                                   

 60%|██████    | 3455/5758 [19:38<12:30,  3.07it/s]
 60%|██████    | 3456/5758 [19:38<12:46,  3.00it/s]
                                                   

 60%|██████    | 3456/5758 [19:38<12:46,  3.00it/s]
 60%|██████    | 3457/5758 [19:38<13:06,  2.92it/s]
                                                   

 60%|██████    | 3457/5758 [19:38<13:06,  2.92it/s]
 60%|██████    | 3458/5758 [19:39<12:35,  3.04it/s]
                                                   

 60%|██████    | 3458/5758 [19:39<12:35,  3.04it/s]
 60%|██████    | 3459/5758 [19:39<12:54,  2.97it/s]
                                                   

 60%|██████    | 3459/5758 [19:39<12:54,  2.97it/s]
 60%|██████    | 3460/5758 [19:39<12:24,  3.09it/s]
                                                   

 60%|██████    | 3460/5758 [19:39<12:24,  3.09it/s]
 60%|██████    | 3461/5758 [19:40<12:45,  3.00it/s]
                                                   

 60%|██████    | 3461/5758 [19:40<12:45,  3.00it/s]
 60%|██████    | 3462/5758 [19:40<13:23,  2.86it/s]
                                                   

 60%|██████    | 3462/5758 [19:40<13:23,  2.86it/s]
 60%|██████    | 3463/5758 [19:40<12:44,  3.00it/s]
                                                   

 60%|██████    | 3463/5758 [19:40<12:44,  3.00it/s]
 60%|██████    | 3464/5758 [19:41<12:56,  2.96it/s]
                                                   

 60%|██████    | 3464/5758 [19:41<12:56,  2.96it/s]
 60%|██████    | 3465/5758 [19:41<13:03,  2.93it/s]
                                                   

 60%|██████    | 3465/5758 [19:41<13:03,  2.93it/s]
 60%|██████    | 3466/5758 [19:42<13:06,  2.91it/s]
                                                   

 60%|██████    | 3466/5758 [19:42<13:06,  2.91it/s]
 60%|██████    | 3467/5758 [19:42<13:02,  2.93it/s]
                                                   

 60%|██████    | 3467/5758 [19:42<13:02,  2.93it/s]
 60%|██████    | 3468/5758 [19:42<12:47,  2.98it/s]
                                                   

 60%|██████    | 3468/5758 [19:42<12:47,  2.98it/s]
 60%|���█████    | 3469/5758 [19:43<13:19,  2.86it/s]
                                                   

 60%|██████    | 3469/5758 [19:43<13:19,  2.86it/s]
 60%|██████    | 3470/5758 [19:43<13:53,  2.74it/s]
                                                   

 60%|██████    | 3470/5758 [19:43<13:53,  2.74it/s]
 60%|██████    | 3471/5758 [19:43<13:07,  2.90it/s]
                                                   

 60%|██████    | 3471/5758 [19:43<13:07,  2.90it/s]
 60%|██████    | 3472/5758 [19:44<13:16,  2.87it/s]
                                                   

 60%|██████    | 3472/5758 [19:44<13:16,  2.87it/s]
 60%|██████    | 3473/5758 [19:44<12:56,  2.94it/s]
                                                   

 60%|██████    | 3473/5758 [19:44<12:56,  2.94it/s]
 60%|██████    | 3474/5758 [19:44<12:48,  2.97it/s]
                                                   

 60%|██████    | 3474/5758 [19:44<12:48,  2.97it/s]
 60%|██████    | 3475/5758 [19:45<13:47,  2.76it/s]
                                                   

 60%|██████    | 3475/5758 [19:45<13:47,  2.76it/s]
 60%|██████    | 3476/5758 [19:45<13:35,  2.80it/s]
                                                   

 60%|██████    | 3476/5758 [19:45<13:35,  2.80it/s]
 60%|██████    | 3477/5758 [19:45<13:09,  2.89it/s]
                                                   

 60%|██████    | 3477/5758 [19:45<13:09,  2.89it/s]
 60%|██████    | 3478/5758 [19:46<14:16,  2.66it/s]
                                                   

 60%|██████    | 3478/5758 [19:46<14:16,  2.66it/s]
 60%|██████    | 3479/5758 [19:46<14:48,  2.57it/s]
                                                   

 60%|██████    | 3479/5758 [19:46<14:48,  2.57it/s]
 60%|██████    | 3480/5758 [19:47<13:55,  2.73it/s]
                                                   

 60%|██████    | 3480/5758 [19:47<13:55,  2.73it/s]
 60%|██████    | 3481/5758 [19:47<13:15,  2.86it/s]
                                                   

 60%|██████    | 3481/5758 [19:47<13:15,  2.86it/s]
 60%|██████    | 3482/5758 [19:47<13:13,  2.87it/s]
                                                   

 60%|██████    | 3482/5758 [19:47<13:13,  2.87it/s]
 60%|██████    | 3483/5758 [19:48<12:57,  2.92it/s]
                                                   

 60%|██████    | 3483/5758 [19:48<12:57,  2.92it/s]
 61%|██████    | 3484/5758 [19:48<12:30,  3.03it/s]
                                                   

 61%|██████    | 3484/5758 [19:48<12:30,  3.03it/s]
 61%|██████    | 3485/5758 [19:48<12:02,  3.15it/s]
                                                   

 61%|██████    | 3485/5758 [19:48<12:02,  3.15it/s]
 61%|██████    | 3486/5758 [19:48<12:48,  2.96it/s]
                                                   

 61%|██████    | 3486/5758 [19:48<12:48,  2.96it/s]
 61%|██████    | 3487/5758 [19:49<12:16,  3.08it/s]
                                                   

 61%|██████    | 3487/5758 [19:49<12:16,  3.08it/s]
 61%|██████    | 3488/5758 [19:49<12:51,  2.94it/s]
                                                   

 61%|██████    | 3488/5758 [19:49<12:51,  2.94it/s]
 61%|██████    | 3489/5758 [19:49<12:13,  3.09it/s]
                                                   

 61%|██████    | 3489/5758 [19:49<12:13,  3.09it/s]
 61%|██████    | 3490/5758 [19:50<12:58,  2.91it/s]
                                                   

 61%|██████    | 3490/5758 [19:50<12:58,  2.91it/s]
 61%|██████    | 3491/5758 [19:50<12:24,  3.04it/s]
                                                   

 61%|██████    | 3491/5758 [19:50<12:24,  3.04it/s]
 61%|██████    | 3492/5758 [19:50<12:49,  2.95it/s]
                                                   

 61%|██████    | 3492/5758 [19:51<12:49,  2.95it/s]
 61%|██████    | 3493/5758 [19:51<12:23,  3.05it/s]
                                                   

 61%|██████    | 3493/5758 [19:51<12:23,  3.05it/s]
 61%|██████    | 3494/5758 [19:51<12:05,  3.12it/s]
                                                   

 61%|██████    | 3494/5758 [19:51<12:05,  3.12it/s]
 61%|██████    | 3495/5758 [19:51<12:47,  2.95it/s]
                                                   

 61%|██████    | 3495/5758 [19:51<12:47,  2.95it/s]
 61%|██████    | 3496/5758 [19:52<12:25,  3.04it/s]
                                                   
{'loss': '0.5123', 'grad_norm': '5.906', 'learning_rate': '4.026e-06', 'num_tokens': '2.292e+06', 'mean_token_accuracy': '0.8416', 'epoch': '0.5976'}
+{'loss': '1.422', 'grad_norm': '15.12', 'learning_rate': '4.024e-06', 'num_tokens': '2.292e+06', 'mean_token_accuracy': '0.6992', 'epoch': '0.5978'}
+{'loss': '1.056', 'grad_norm': '9.25', 'learning_rate': '4.022e-06', 'num_tokens': '2.294e+06', 'mean_token_accuracy': '0.7311', 'epoch': '0.598'}
+{'loss': '0.6313', 'grad_norm': '8.875', 'learning_rate': '4.02e-06', 'num_tokens': '2.295e+06', 'mean_token_accuracy': '0.7939', 'epoch': '0.5981'}
+{'loss': '1.094', 'grad_norm': '18.38', 'learning_rate': '4.019e-06', 'num_tokens': '2.295e+06', 'mean_token_accuracy': '0.7617', 'epoch': '0.5983'}
+{'loss': '1.517', 'grad_norm': '14.12', 'learning_rate': '4.017e-06', 'num_tokens': '2.296e+06', 'mean_token_accuracy': '0.6107', 'epoch': '0.5985'}
+{'loss': '1.247', 'grad_norm': '12.44', 'learning_rate': '4.015e-06', 'num_tokens': '2.297e+06', 'mean_token_accuracy': '0.671', 'epoch': '0.5986'}
+{'loss': '1.613', 'grad_norm': '43', 'learning_rate': '4.014e-06', 'num_tokens': '2.297e+06', 'mean_token_accuracy': '0.7368', 'epoch': '0.5988'}
+{'loss': '1.074', 'grad_norm': '7.688', 'learning_rate': '4.012e-06', 'num_tokens': '2.298e+06', 'mean_token_accuracy': '0.7441', 'epoch': '0.599'}
+{'loss': '1.369', 'grad_norm': '16.5', 'learning_rate': '4.01e-06', 'num_tokens': '2.299e+06', 'mean_token_accuracy': '0.6364', 'epoch': '0.5992'}
+{'loss': '0.9053', 'grad_norm': '24.38', 'learning_rate': '4.008e-06', 'num_tokens': '2.299e+06', 'mean_token_accuracy': '0.7778', 'epoch': '0.5993'}
+{'loss': '1.078', 'grad_norm': '14.69', 'learning_rate': '4.007e-06', 'num_tokens': '2.299e+06', 'mean_token_accuracy': '0.7079', 'epoch': '0.5995'}
+{'loss': '0.5389', 'grad_norm': '5.781', 'learning_rate': '4.005e-06', 'num_tokens': '2.301e+06', 'mean_token_accuracy': '0.8499', 'epoch': '0.5997'}
+{'loss': '0.708', 'grad_norm': '38.75', 'learning_rate': '4.003e-06', 'num_tokens': '2.301e+06', 'mean_token_accuracy': '0.8889', 'epoch': '0.5999'}
+{'loss': '1.056', 'grad_norm': '18.12', 'learning_rate': '4.001e-06', 'num_tokens': '2.302e+06', 'mean_token_accuracy': '0.756', 'epoch': '0.6'}
+{'loss': '0.5452', 'grad_norm': '25.25', 'learning_rate': '4e-06', 'num_tokens': '2.302e+06', 'mean_token_accuracy': '0.8611', 'epoch': '0.6002'}
+{'loss': '1.486', 'grad_norm': '9.375', 'learning_rate': '3.998e-06', 'num_tokens': '2.303e+06', 'mean_token_accuracy': '0.6898', 'epoch': '0.6004'}
+{'loss': '1.712', 'grad_norm': '15.31', 'learning_rate': '3.996e-06', 'num_tokens': '2.303e+06', 'mean_token_accuracy': '0.6511', 'epoch': '0.6006'}
+{'loss': '0.7044', 'grad_norm': '6.844', 'learning_rate': '3.994e-06', 'num_tokens': '2.305e+06', 'mean_token_accuracy': '0.7855', 'epoch': '0.6007'}
+{'loss': '1.515', 'grad_norm': '36.75', 'learning_rate': '3.993e-06', 'num_tokens': '2.305e+06', 'mean_token_accuracy': '0.7429', 'epoch': '0.6009'}
+{'loss': '1.034', 'grad_norm': '9.688', 'learning_rate': '3.991e-06', 'num_tokens': '2.306e+06', 'mean_token_accuracy': '0.7651', 'epoch': '0.6011'}
+{'loss': '1.096', 'grad_norm': '6.062', 'learning_rate': '3.989e-06', 'num_tokens': '2.309e+06', 'mean_token_accuracy': '0.7282', 'epoch': '0.6013'}
+{'loss': '1.73', 'grad_norm': '8.688', 'learning_rate': '3.987e-06', 'num_tokens': '2.31e+06', 'mean_token_accuracy': '0.6361', 'epoch': '0.6014'}
+{'loss': '1.155', 'grad_norm': '20', 'learning_rate': '3.986e-06', 'num_tokens': '2.31e+06', 'mean_token_accuracy': '0.7203', 'epoch': '0.6016'}
+{'loss': '0.8272', 'grad_norm': '18.88', 'learning_rate': '3.984e-06', 'num_tokens': '2.31e+06', 'mean_token_accuracy': '0.8077', 'epoch': '0.6018'}
+{'loss': '1.17', 'grad_norm': '21.5', 'learning_rate': '3.982e-06', 'num_tokens': '2.311e+06', 'mean_token_accuracy': '0.7566', 'epoch': '0.6019'}
+{'loss': '0.5874', 'grad_norm': '6.906', 'learning_rate': '3.981e-06', 'num_tokens': '2.312e+06', 'mean_token_accuracy': '0.825', 'epoch': '0.6021'}
+{'loss': '1.186', 'grad_norm': '12.31', 'learning_rate': '3.979e-06', 'num_tokens': '2.313e+06', 'mean_token_accuracy': '0.6916', 'epoch': '0.6023'}
+{'loss': '1.384', 'grad_norm': '14.62', 'learning_rate': '3.977e-06', 'num_tokens': '2.313e+06', 'mean_token_accuracy': '0.6369', 'epoch': '0.6025'}
+{'loss': '1.757', 'grad_norm': '19.88', 'learning_rate': '3.975e-06', 'num_tokens': '2.314e+06', 'mean_token_accuracy': '0.6622', 'epoch': '0.6026'}
+{'loss': '1.021', 'grad_norm': '12.44', 'learning_rate': '3.974e-06', 'num_tokens': '2.314e+06', 'mean_token_accuracy': '0.7372', 'epoch': '0.6028'}
+{'loss': '1.594', 'grad_norm': '17.5', 'learning_rate': '3.972e-06', 'num_tokens': '2.315e+06', 'mean_token_accuracy': '0.639', 'epoch': '0.603'}
+{'loss': '1.214', 'grad_norm': '12.62', 'learning_rate': '3.97e-06', 'num_tokens': '2.316e+06', 'mean_token_accuracy': '0.6812', 'epoch': '0.6032'}
+{'loss': '1.47', 'grad_norm': '19.25', 'learning_rate': '3.968e-06', 'num_tokens': '2.316e+06', 'mean_token_accuracy': '0.6548', 'epoch': '0.6033'}
+{'loss': '1.663', 'grad_norm': '13.12', 'learning_rate': '3.967e-06', 'num_tokens': '2.317e+06', 'mean_token_accuracy': '0.653', 'epoch': '0.6035'}
+{'loss': '1.043', 'grad_norm': '7.031', 'learning_rate': '3.965e-06', 'num_tokens': '2.318e+06', 'mean_token_accuracy': '0.7629', 'epoch': '0.6037'}
+{'loss': '0.9441', 'grad_norm': '8.062', 'learning_rate': '3.963e-06', 'num_tokens': '2.32e+06', 'mean_token_accuracy': '0.7629', 'epoch': '0.6039'}
+{'loss': '1.55', 'grad_norm': '9.312', 'learning_rate': '3.961e-06', 'num_tokens': '2.321e+06', 'mean_token_accuracy': '0.6789', 'epoch': '0.604'}
+{'loss': '2.158', 'grad_norm': '11.06', 'learning_rate': '3.96e-06', 'num_tokens': '2.322e+06', 'mean_token_accuracy': '0.5521', 'epoch': '0.6042'}
+{'loss': '0.5354', 'grad_norm': '8.188', 'learning_rate': '3.958e-06', 'num_tokens': '2.323e+06', 'mean_token_accuracy': '0.8444', 'epoch': '0.6044'}
+{'loss': '1.5', 'grad_norm': '23', 'learning_rate': '3.956e-06', 'num_tokens': '2.323e+06', 'mean_token_accuracy': '0.6209', 'epoch': '0.6046'}
+{'loss': '1.994', 'grad_norm': '19.12', 'learning_rate': '3.954e-06', 'num_tokens': '2.324e+06', 'mean_token_accuracy': '0.5853', 'epoch': '0.6047'}
+{'loss': '1.126', 'grad_norm': '33.25', 'learning_rate': '3.953e-06', 'num_tokens': '2.324e+06', 'mean_token_accuracy': '0.8105', 'epoch': '0.6049'}
+{'loss': '1.06', 'grad_norm': '11.75', 'learning_rate': '3.951e-06', 'num_tokens': '2.325e+06', 'mean_token_accuracy': '0.7001', 'epoch': '0.6051'}
+{'loss': '0.916', 'grad_norm': '14.5', 'learning_rate': '3.949e-06', 'num_tokens': '2.325e+06', 'mean_token_accuracy': '0.7333', 'epoch': '0.6052'}
+{'loss': '0.849', 'grad_norm': '10.75', 'learning_rate': '3.948e-06', 'num_tokens': '2.327e+06', 'mean_token_accuracy': '0.7842', 'epoch': '0.6054'}
+{'loss': '1.416', 'grad_norm': '14.5', 'learning_rate': '3.946e-06', 'num_tokens': '2.327e+06', 'mean_token_accuracy': '0.6355', 'epoch': '0.6056'}
+{'loss': '2.371', 'grad_norm': '18.88', 'learning_rate': '3.944e-06', 'num_tokens': '2.328e+06', 'mean_token_accuracy': '0.5241', 'epoch': '0.6058'}
+{'loss': '1.022', 'grad_norm': '44.75', 'learning_rate': '3.942e-06', 'num_tokens': '2.328e+06', 'mean_token_accuracy': '0.8088', 'epoch': '0.6059'}
+{'loss': '1.029', 'grad_norm': '28.62', 'learning_rate': '3.941e-06', 'num_tokens': '2.328e+06', 'mean_token_accuracy': '0.7846', 'epoch': '0.6061'}
+{'loss': '1.012', 'grad_norm': '13.81', 'learning_rate': '3.939e-06', 'num_tokens': '2.329e+06', 'mean_token_accuracy': '0.7087', 'epoch': '0.6063'}
+{'loss': '1.161', 'grad_norm': '16.75', 'learning_rate': '3.937e-06', 'num_tokens': '2.329e+06', 'mean_token_accuracy': '0.7519', 'epoch': '0.6065'}
+{'loss': '1.153', 'grad_norm': '14.25', 'learning_rate': '3.935e-06', 'num_tokens': '2.329e+06', 'mean_token_accuracy': '0.6927', 'epoch': '0.6066'}
+{'loss': '1.858', 'grad_norm': '36.25', 'learning_rate': '3.934e-06', 'num_tokens': '2.33e+06', 'mean_token_accuracy': '0.614', 'epoch': '0.6068'}
+{'loss': '2.58', 'grad_norm': '9.688', 'learning_rate': '3.932e-06', 'num_tokens': '2.331e+06', 'mean_token_accuracy': '0.48', 'epoch': '0.607'}
+
 61%|██████    | 3496/5758 [19:52<12:25,  3.04it/s]
 61%|██████    | 3497/5758 [19:52<12:06,  3.11it/s]
                                                   

 61%|██████    | 3497/5758 [19:52<12:06,  3.11it/s]
 61%|██████    | 3498/5758 [19:52<12:37,  2.99it/s]
                                                   

 61%|██████    | 3498/5758 [19:52<12:37,  2.99it/s]
 61%|██████    | 3499/5758 [19:53<12:45,  2.95it/s]
                                                   

 61%|██████    | 3499/5758 [19:53<12:45,  2.95it/s]
 61%|██████    | 3500/5758 [19:53<12:26,  3.03it/s]
                                                   

 61%|██████    | 3500/5758 [19:53<12:26,  3.03it/s]{'loss': '1.126', 'grad_norm': '11.88', 'learning_rate': '3.93e-06', 'num_tokens': '2.331e+06', 'mean_token_accuracy': '0.6867', 'epoch': '0.6072'}
+{'loss': '2.085', 'grad_norm': '16.88', 'learning_rate': '3.928e-06', 'num_tokens': '2.332e+06', 'mean_token_accuracy': '0.5967', 'epoch': '0.6073'}
+{'loss': '2.036', 'grad_norm': '13.69', 'learning_rate': '3.927e-06', 'num_tokens': '2.333e+06', 'mean_token_accuracy': '0.5785', 'epoch': '0.6075'}
+{'loss': '0.581', 'grad_norm': '5.906', 'learning_rate': '3.925e-06', 'num_tokens': '2.334e+06', 'mean_token_accuracy': '0.8176', 'epoch': '0.6077'}
+{'loss': '1.18', 'grad_norm': '17.25', 'learning_rate': '3.923e-06', 'num_tokens': '2.335e+06', 'mean_token_accuracy': '0.7267', 'epoch': '0.6078'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.62it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.61it/s]
+
 61%|██████    | 3501/5758 [19:55<33:31,  1.12it/s]
                                                   

 61%|██████    | 3501/5758 [19:55<33:31,  1.12it/s]
 61%|██████    | 3502/5758 [19:56<27:11,  1.38it/s]
                                                   

 61%|██████    | 3502/5758 [19:56<27:11,  1.38it/s]
 61%|██████    | 3503/5758 [19:56<22:54,  1.64it/s]
                                                   

 61%|██████    | 3503/5758 [19:56<22:54,  1.64it/s]
 61%|██████    | 3504/5758 [19:56<20:17,  1.85it/s]
                                                   

 61%|██████    | 3504/5758 [19:56<20:17,  1.85it/s]
 61%|██████    | 3505/5758 [19:57<17:29,  2.15it/s]
                                                   

 61%|██████    | 3505/5758 [19:57<17:29,  2.15it/s]
 61%|██████    | 3506/5758 [19:57<15:54,  2.36it/s]
                                                   

 61%|██████    | 3506/5758 [19:57<15:54,  2.36it/s]
 61%|██████    | 3507/5758 [19:57<14:23,  2.61it/s]
                                                   

 61%|██████    | 3507/5758 [19:57<14:23,  2.61it/s]
 61%|██████    | 3508/5758 [19:58<13:25,  2.79it/s]
                                                   

 61%|██████    | 3508/5758 [19:58<13:25,  2.79it/s]
 61%|██████    | 3509/5758 [19:58<13:43,  2.73it/s]
                                                   

 61%|██████    | 3509/5758 [19:58<13:43,  2.73it/s]
 61%|██████    | 3510/5758 [19:58<13:08,  2.85it/s]
                                                   

 61%|██████    | 3510/5758 [19:58<13:08,  2.85it/s]
 61%|██████    | 3511/5758 [19:59<13:25,  2.79it/s]
                                                   

 61%|██████    | 3511/5758 [19:59<13:25,  2.79it/s]
 61%|██████    | 3512/5758 [19:59<12:33,  2.98it/s]
                                                   

 61%|██████    | 3512/5758 [19:59<12:33,  2.98it/s]
 61%|██████    | 3513/5758 [19:59<12:03,  3.10it/s]
                                                   

 61%|██████    | 3513/5758 [19:59<12:03,  3.10it/s]
 61%|██████    | 3514/5758 [20:00<11:44,  3.19it/s]
                                                   

 61%|██████    | 3514/5758 [20:00<11:44,  3.19it/s]
 61%|██████    | 3515/5758 [20:00<12:19,  3.03it/s]
                                                   

 61%|██████    | 3515/5758 [20:00<12:19,  3.03it/s]
 61%|██████    | 3516/5758 [20:00<12:54,  2.90it/s]
                                                   

 61%|██████    | 3516/5758 [20:00<12:54,  2.90it/s]
 61%|██████    | 3517/5758 [20:01<13:16,  2.81it/s]
                                                   

 61%|██████    | 3517/5758 [20:01<13:16,  2.81it/s]
 61%|██████    | 3518/5758 [20:01<12:50,  2.91it/s]
                                                   

 61%|██████    | 3518/5758 [20:01<12:50,  2.91it/s]
 61%|██████    | 3519/5758 [20:01<12:25,  3.00it/s]
                                                   

 61%|██████    | 3519/5758 [20:01<12:25,  3.00it/s]
 61%|██████    | 3520/5758 [20:02<12:51,  2.90it/s]
                                                   

 61%|██████    | 3520/5758 [20:02<12:51,  2.90it/s]
 61%|██████    | 3521/5758 [20:02<12:20,  3.02it/s]
                                                   

 61%|██████    | 3521/5758 [20:02<12:20,  3.02it/s]
 61%|██████    | 3522/5758 [20:02<12:46,  2.92it/s]
                                                   

 61%|██████    | 3522/5758 [20:02<12:46,  2.92it/s]
 61%|██████    | 3523/5758 [20:03<13:09,  2.83it/s]
                                                   

 61%|██████    | 3523/5758 [20:03<13:09,  2.83it/s]
 61%|██████    | 3524/5758 [20:03<12:31,  2.97it/s]
                                                   

 61%|██████    | 3524/5758 [20:03<12:31,  2.97it/s]
 61%|██████    | 3525/5758 [20:03<12:28,  2.98it/s]
                                                   

 61%|██████    | 3525/5758 [20:03<12:28,  2.98it/s]
 61%|██████    | 3526/5758 [20:04<12:15,  3.03it/s]
                                                   

 61%|██████    | 3526/5758 [20:04<12:15,  3.03it/s]
 61%|██████▏   | 3527/5758 [20:04<12:11,  3.05it/s]
                                                   

 61%|██████▏   | 3527/5758 [20:04<12:11,  3.05it/s]
 61%|██████▏   | 3528/5758 [20:04<11:50,  3.14it/s]
                                                   

 61%|██████▏   | 3528/5758 [20:04<11:50,  3.14it/s]
 61%|██████▏   | 3529/5758 [20:05<13:16,  2.80it/s]
                                                   

 61%|██████▏   | 3529/5758 [20:05<13:16,  2.80it/s]
 61%|██████▏   | 3530/5758 [20:05<12:50,  2.89it/s]
                                                   

 61%|██████▏   | 3530/5758 [20:05<12:50,  2.89it/s]
 61%|██████▏   | 3531/5758 [20:05<12:55,  2.87it/s]
                                                   

 61%|██████▏   | 3531/5758 [20:05<12:55,  2.87it/s]
 61%|██████▏   | 3532/5758 [20:06<13:39,  2.72it/s]
                                                   

 61%|██████▏   | 3532/5758 [20:06<13:39,  2.72it/s]
 61%|██████▏   | 3533/5758 [20:06<13:19,  2.78it/s]
                                                   

 61%|██████▏   | 3533/5758 [20:06<13:19,  2.78it/s]
 61%|██████▏   | 3534/5758 [20:06<12:45,  2.91it/s]
                                                   

 61%|██████▏   | 3534/5758 [20:06<12:45,  2.91it/s]
 61%|██████▏   | 3535/5758 [20:07<12:37,  2.93it/s]
                                                   

 61%|██████▏   | 3535/5758 [20:07<12:37,  2.93it/s]
 61%|██████▏   | 3536/5758 [20:07<12:52,  2.88it/s]
                                                   

 61%|██████▏   | 3536/5758 [20:07<12:52,  2.88it/s]
 61%|██████▏   | 3537/5758 [20:07<12:26,  2.98it/s]
                                                   

 61%|██████▏   | 3537/5758 [20:07<12:26,  2.98it/s]
 61%|██████▏   | 3538/5758 [20:08<12:00,  3.08it/s]
                                                   

 61%|██████▏   | 3538/5758 [20:08<12:00,  3.08it/s]
 61%|██████▏   | 3539/5758 [20:08<11:58,  3.09it/s]
                                                   

 61%|██████▏   | 3539/5758 [20:08<11:58,  3.09it/s]
 61%|██████▏   | 3540/5758 [20:08<12:34,  2.94it/s]
                                                   

 61%|██████▏   | 3540/5758 [20:08<12:34,  2.94it/s]
 61%|██████▏   | 3541/5758 [20:09<13:00,  2.84it/s]
                                                   

 61%|██████▏   | 3541/5758 [20:09<13:00,  2.84it/s]
 62%|██████▏   | 3542/5758 [20:09<12:45,  2.89it/s]
                                                   

 62%|██████▏   | 3542/5758 [20:09<12:45,  2.89it/s]
 62%|██████▏   | 3543/5758 [20:09<12:10,  3.03it/s]
                                                   

 62%|██████▏   | 3543/5758 [20:09<12:10,  3.03it/s]
 62%|██████▏   | 3544/5758 [20:10<11:35,  3.19it/s]
                                                   

 62%|██████▏   | 3544/5758 [20:10<11:35,  3.19it/s]
 62%|██████▏   | 3545/5758 [20:10<11:32,  3.20it/s]
                                                   

 62%|██████▏   | 3545/5758 [20:10<11:32,  3.20it/s]
 62%|██████▏   | 3546/5758 [20:10<11:19,  3.26it/s]
                                                   

 62%|██████▏   | 3546/5758 [20:10<11:19,  3.26it/s]
 62%|██████▏   | 3547/5758 [20:11<11:55,  3.09it/s]
                                                   

 62%|██████▏   | 3547/5758 [20:11<11:55,  3.09it/s]
 62%|██████▏   | 3548/5758 [20:11<11:42,  3.15it/s]
                                                   

 62%|██████▏   | 3548/5758 [20:11<11:42,  3.15it/s]
 62%|██████▏   | 3549/5758 [20:11<11:55,  3.09it/s]
                                                   

 62%|██████▏   | 3549/5758 [20:11<11:55,  3.09it/s]
 62%|██████▏   | 3550/5758 [20:12<12:29,  2.94it/s]
                                                   

 62%|██████▏   | 3550/5758 [20:12<12:29,  2.94it/s]
 62%|██████▏   | 3551/5758 [20:12<12:38,  2.91it/s]
                                                   

 62%|██████▏   | 3551/5758 [20:12<12:38,  2.91it/s]
 62%|██████▏   | 3552/5758 [20:12<13:09,  2.79it/s]
                                                   

 62%|██████▏   | 3552/5758 [20:12<13:09,  2.79it/s]
 62%|██████▏   | 3553/5758 [20:13<12:46,  2.87it/s]
                                                   

 62%|██████▏   | 3553/5758 [20:13<12:46,  2.87it/s]
 62%|██████▏   | 3554/5758 [20:13<13:53,  2.64it/s]
                                                   

 62%|██████▏   | 3554/5758 [20:13<13:53,  2.64it/s]
 62%|██████▏   | 3555/5758 [20:14<12:57,  2.83it/s]
                                                   

 62%|██████▏   | 3555/5758 [20:14<12:57,  2.83it/s]
 62%|██████▏   | 3556/5758 [20:14<12:42,  2.89it/s]
                                                   
{'loss': '1.164', 'grad_norm': '11.94', 'learning_rate': '3.922e-06', 'num_tokens': '2.336e+06', 'mean_token_accuracy': '0.6626', 'epoch': '0.608'}
+{'loss': '1.703', 'grad_norm': '15', 'learning_rate': '3.92e-06', 'num_tokens': '2.336e+06', 'mean_token_accuracy': '0.6336', 'epoch': '0.6082'}
+{'loss': '1.726', 'grad_norm': '16.38', 'learning_rate': '3.918e-06', 'num_tokens': '2.337e+06', 'mean_token_accuracy': '0.6308', 'epoch': '0.6084'}
+{'loss': '1.113', 'grad_norm': '13.94', 'learning_rate': '3.916e-06', 'num_tokens': '2.337e+06', 'mean_token_accuracy': '0.696', 'epoch': '0.6085'}
+{'loss': '2.372', 'grad_norm': '37.75', 'learning_rate': '3.915e-06', 'num_tokens': '2.337e+06', 'mean_token_accuracy': '0.5125', 'epoch': '0.6087'}
+{'loss': '0.9638', 'grad_norm': '21', 'learning_rate': '3.913e-06', 'num_tokens': '2.337e+06', 'mean_token_accuracy': '0.7885', 'epoch': '0.6089'}
+{'loss': '0.8596', 'grad_norm': '10.75', 'learning_rate': '3.911e-06', 'num_tokens': '2.339e+06', 'mean_token_accuracy': '0.7306', 'epoch': '0.6091'}
+{'loss': '1.128', 'grad_norm': '13.31', 'learning_rate': '3.909e-06', 'num_tokens': '2.339e+06', 'mean_token_accuracy': '0.7075', 'epoch': '0.6092'}
+{'loss': '0.7289', 'grad_norm': '8.938', 'learning_rate': '3.908e-06', 'num_tokens': '2.34e+06', 'mean_token_accuracy': '0.7726', 'epoch': '0.6094'}
+{'loss': '1.306', 'grad_norm': '26.88', 'learning_rate': '3.906e-06', 'num_tokens': '2.341e+06', 'mean_token_accuracy': '0.7143', 'epoch': '0.6096'}
+{'loss': '0.856', 'grad_norm': '8.062', 'learning_rate': '3.904e-06', 'num_tokens': '2.342e+06', 'mean_token_accuracy': '0.7717', 'epoch': '0.6098'}
+{'loss': '1.08', 'grad_norm': '13.62', 'learning_rate': '3.902e-06', 'num_tokens': '2.342e+06', 'mean_token_accuracy': '0.7178', 'epoch': '0.6099'}
+{'loss': '1.771', 'grad_norm': '37.75', 'learning_rate': '3.901e-06', 'num_tokens': '2.343e+06', 'mean_token_accuracy': '0.6242', 'epoch': '0.6101'}
+{'loss': '0.7311', 'grad_norm': '10.88', 'learning_rate': '3.899e-06', 'num_tokens': '2.344e+06', 'mean_token_accuracy': '0.7759', 'epoch': '0.6103'}
+{'loss': '1.674', 'grad_norm': '11.88', 'learning_rate': '3.897e-06', 'num_tokens': '2.345e+06', 'mean_token_accuracy': '0.6525', 'epoch': '0.6105'}
+{'loss': '0.3814', 'grad_norm': '15.81', 'learning_rate': '3.895e-06', 'num_tokens': '2.345e+06', 'mean_token_accuracy': '0.9153', 'epoch': '0.6106'}
+{'loss': '0.6374', 'grad_norm': '9.25', 'learning_rate': '3.894e-06', 'num_tokens': '2.346e+06', 'mean_token_accuracy': '0.8079', 'epoch': '0.6108'}
+{'loss': '1.939', 'grad_norm': '12.12', 'learning_rate': '3.892e-06', 'num_tokens': '2.347e+06', 'mean_token_accuracy': '0.5988', 'epoch': '0.611'}
+{'loss': '0.9289', 'grad_norm': '19.38', 'learning_rate': '3.89e-06', 'num_tokens': '2.347e+06', 'mean_token_accuracy': '0.7407', 'epoch': '0.6111'}
+{'loss': '0.7269', 'grad_norm': '28.5', 'learning_rate': '3.889e-06', 'num_tokens': '2.347e+06', 'mean_token_accuracy': '0.8261', 'epoch': '0.6113'}
+{'loss': '1.885', 'grad_norm': '18.75', 'learning_rate': '3.887e-06', 'num_tokens': '2.347e+06', 'mean_token_accuracy': '0.6577', 'epoch': '0.6115'}
+{'loss': '2.115', 'grad_norm': '11.44', 'learning_rate': '3.885e-06', 'num_tokens': '2.348e+06', 'mean_token_accuracy': '0.5956', 'epoch': '0.6117'}
+{'loss': '1.187', 'grad_norm': '15.19', 'learning_rate': '3.883e-06', 'num_tokens': '2.349e+06', 'mean_token_accuracy': '0.7026', 'epoch': '0.6118'}
+{'loss': '0.6873', 'grad_norm': '8.125', 'learning_rate': '3.882e-06', 'num_tokens': '2.35e+06', 'mean_token_accuracy': '0.7803', 'epoch': '0.612'}
+{'loss': '1.071', 'grad_norm': '11.5', 'learning_rate': '3.88e-06', 'num_tokens': '2.351e+06', 'mean_token_accuracy': '0.6763', 'epoch': '0.6122'}
+{'loss': '1.008', 'grad_norm': '11.38', 'learning_rate': '3.878e-06', 'num_tokens': '2.352e+06', 'mean_token_accuracy': '0.7204', 'epoch': '0.6124'}
+{'loss': '1.395', 'grad_norm': '12.94', 'learning_rate': '3.876e-06', 'num_tokens': '2.353e+06', 'mean_token_accuracy': '0.6378', 'epoch': '0.6125'}
+{'loss': '2.061', 'grad_norm': '26.88', 'learning_rate': '3.875e-06', 'num_tokens': '2.353e+06', 'mean_token_accuracy': '0.6556', 'epoch': '0.6127'}
+{'loss': '1.92', 'grad_norm': '15.56', 'learning_rate': '3.873e-06', 'num_tokens': '2.354e+06', 'mean_token_accuracy': '0.61', 'epoch': '0.6129'}
+{'loss': '0.7297', 'grad_norm': '15.38', 'learning_rate': '3.871e-06', 'num_tokens': '2.354e+06', 'mean_token_accuracy': '0.8125', 'epoch': '0.6131'}
+{'loss': '1.077', 'grad_norm': '33', 'learning_rate': '3.869e-06', 'num_tokens': '2.354e+06', 'mean_token_accuracy': '0.7778', 'epoch': '0.6132'}
+{'loss': '1.038', 'grad_norm': '17.12', 'learning_rate': '3.868e-06', 'num_tokens': '2.354e+06', 'mean_token_accuracy': '0.7461', 'epoch': '0.6134'}
+{'loss': '0.7654', 'grad_norm': '19.38', 'learning_rate': '3.866e-06', 'num_tokens': '2.355e+06', 'mean_token_accuracy': '0.8558', 'epoch': '0.6136'}
+{'loss': '1.688', 'grad_norm': '11.12', 'learning_rate': '3.864e-06', 'num_tokens': '2.355e+06', 'mean_token_accuracy': '0.6488', 'epoch': '0.6138'}
+{'loss': '1.037', 'grad_norm': '11.31', 'learning_rate': '3.862e-06', 'num_tokens': '2.356e+06', 'mean_token_accuracy': '0.7037', 'epoch': '0.6139'}
+{'loss': '1.25', 'grad_norm': '10.75', 'learning_rate': '3.861e-06', 'num_tokens': '2.357e+06', 'mean_token_accuracy': '0.6816', 'epoch': '0.6141'}
+{'loss': '1.532', 'grad_norm': '10.44', 'learning_rate': '3.859e-06', 'num_tokens': '2.359e+06', 'mean_token_accuracy': '0.6643', 'epoch': '0.6143'}
+{'loss': '1.502', 'grad_norm': '13.94', 'learning_rate': '3.857e-06', 'num_tokens': '2.359e+06', 'mean_token_accuracy': '0.6287', 'epoch': '0.6144'}
+{'loss': '0.6584', 'grad_norm': '13.88', 'learning_rate': '3.856e-06', 'num_tokens': '2.36e+06', 'mean_token_accuracy': '0.8161', 'epoch': '0.6146'}
+{'loss': '1.486', 'grad_norm': '10.12', 'learning_rate': '3.854e-06', 'num_tokens': '2.361e+06', 'mean_token_accuracy': '0.6521', 'epoch': '0.6148'}
+{'loss': '1.102', 'grad_norm': '21.88', 'learning_rate': '3.852e-06', 'num_tokens': '2.361e+06', 'mean_token_accuracy': '0.7634', 'epoch': '0.615'}
+{'loss': '1.19', 'grad_norm': '11.75', 'learning_rate': '3.85e-06', 'num_tokens': '2.362e+06', 'mean_token_accuracy': '0.6719', 'epoch': '0.6151'}
+{'loss': '1.43', 'grad_norm': '9.812', 'learning_rate': '3.849e-06', 'num_tokens': '2.363e+06', 'mean_token_accuracy': '0.6551', 'epoch': '0.6153'}
+{'loss': '1.004', 'grad_norm': '36.5', 'learning_rate': '3.847e-06', 'num_tokens': '2.363e+06', 'mean_token_accuracy': '0.7581', 'epoch': '0.6155'}
+{'loss': '1.062', 'grad_norm': '19.5', 'learning_rate': '3.845e-06', 'num_tokens': '2.363e+06', 'mean_token_accuracy': '0.7814', 'epoch': '0.6157'}
+{'loss': '0.4588', 'grad_norm': '7.844', 'learning_rate': '3.843e-06', 'num_tokens': '2.365e+06', 'mean_token_accuracy': '0.8571', 'epoch': '0.6158'}
+{'loss': '1.272', 'grad_norm': '11.38', 'learning_rate': '3.842e-06', 'num_tokens': '2.366e+06', 'mean_token_accuracy': '0.6405', 'epoch': '0.616'}
+{'loss': '1.745', 'grad_norm': '16.88', 'learning_rate': '3.84e-06', 'num_tokens': '2.366e+06', 'mean_token_accuracy': '0.6541', 'epoch': '0.6162'}
+{'loss': '1.558', 'grad_norm': '15.94', 'learning_rate': '3.838e-06', 'num_tokens': '2.366e+06', 'mean_token_accuracy': '0.6659', 'epoch': '0.6164'}
+{'loss': '1.137', 'grad_norm': '13.81', 'learning_rate': '3.836e-06', 'num_tokens': '2.367e+06', 'mean_token_accuracy': '0.6922', 'epoch': '0.6165'}
+{'loss': '1.612', 'grad_norm': '10.06', 'learning_rate': '3.835e-06', 'num_tokens': '2.368e+06', 'mean_token_accuracy': '0.6629', 'epoch': '0.6167'}
+{'loss': '1.693', 'grad_norm': '14.25', 'learning_rate': '3.833e-06', 'num_tokens': '2.369e+06', 'mean_token_accuracy': '0.6603', 'epoch': '0.6169'}
+{'loss': '2.232', 'grad_norm': '17', 'learning_rate': '3.831e-06', 'num_tokens': '2.369e+06', 'mean_token_accuracy': '0.5471', 'epoch': '0.6171'}
+{'loss': '0.8318', 'grad_norm': '13.75', 'learning_rate': '3.829e-06', 'num_tokens': '2.369e+06', 'mean_token_accuracy': '0.7911', 'epoch': '0.6172'}
+{'loss': '0.9953', 'grad_norm': '46.5', 'learning_rate': '3.828e-06', 'num_tokens': '2.369e+06', 'mean_token_accuracy': '0.7647', 'epoch': '0.6174'}
+
 62%|██████▏   | 3556/5758 [20:14<12:42,  2.89it/s]
 62%|██████▏   | 3557/5758 [20:14<12:33,  2.92it/s]
                                                   

 62%|██████▏   | 3557/5758 [20:14<12:33,  2.92it/s]
 62%|██████▏   | 3558/5758 [20:14<12:04,  3.04it/s]
                                                   

 62%|██████▏   | 3558/5758 [20:14<12:04,  3.04it/s]
 62%|██████▏   | 3559/5758 [20:15<11:51,  3.09it/s]
                                                   

 62%|██████▏   | 3559/5758 [20:15<11:51,  3.09it/s]
 62%|██████▏   | 3560/5758 [20:15<12:08,  3.02it/s]
                                                   

 62%|██████▏   | 3560/5758 [20:15<12:08,  3.02it/s]
 62%|██████▏   | 3561/5758 [20:15<11:53,  3.08it/s]
                                                   

 62%|██████▏   | 3561/5758 [20:15<11:53,  3.08it/s]
 62%|██████▏   | 3562/5758 [20:16<11:55,  3.07it/s]
                                                   

 62%|██████▏   | 3562/5758 [20:16<11:55,  3.07it/s]
 62%|██████▏   | 3563/5758 [20:16<11:52,  3.08it/s]
                                                   

 62%|██████▏   | 3563/5758 [20:16<11:52,  3.08it/s]
 62%|██████▏   | 3564/5758 [20:16<11:55,  3.07it/s]
                                                   

 62%|██████▏   | 3564/5758 [20:16<11:55,  3.07it/s]
 62%|██████▏   | 3565/5758 [20:17<12:36,  2.90it/s]
                                                   

 62%|██████▏   | 3565/5758 [20:17<12:36,  2.90it/s]
 62%|██████▏   | 3566/5758 [20:17<12:06,  3.02it/s]
                                                   

 62%|██████▏   | 3566/5758 [20:17<12:06,  3.02it/s]
 62%|██████▏   | 3567/5758 [20:17<11:52,  3.07it/s]
                                                   

 62%|██████▏   | 3567/5758 [20:17<11:52,  3.07it/s]
 62%|██████▏   | 3568/5758 [20:18<11:45,  3.11it/s]
                                                   

 62%|██████▏   | 3568/5758 [20:18<11:45,  3.11it/s]
 62%|██████▏   | 3569/5758 [20:18<11:51,  3.08it/s]
                                                   

 62%|██████▏   | 3569/5758 [20:18<11:51,  3.08it/s]
 62%|██████▏   | 3570/5758 [20:18<11:36,  3.14it/s]
                                                   

 62%|██████▏   | 3570/5758 [20:18<11:36,  3.14it/s]
 62%|██████▏   | 3571/5758 [20:19<11:57,  3.05it/s]
                                                   

 62%|██████▏   | 3571/5758 [20:19<11:57,  3.05it/s]
 62%|██████▏   | 3572/5758 [20:19<11:34,  3.15it/s]
                                                   

 62%|██████▏   | 3572/5758 [20:19<11:34,  3.15it/s]
 62%|██████▏   | 3573/5758 [20:19<11:22,  3.20it/s]
                                                   

 62%|██████▏   | 3573/5758 [20:19<11:22,  3.20it/s]
 62%|██████▏   | 3574/5758 [20:20<11:51,  3.07it/s]
                                                   

 62%|██████▏   | 3574/5758 [20:20<11:51,  3.07it/s]
 62%|██████▏   | 3575/5758 [20:20<12:05,  3.01it/s]
                                                   

 62%|██████▏   | 3575/5758 [20:20<12:05,  3.01it/s]
 62%|██████▏   | 3576/5758 [20:20<12:44,  2.85it/s]
                                                   

 62%|██████▏   | 3576/5758 [20:20<12:44,  2.85it/s]
 62%|██████▏   | 3577/5758 [20:21<12:04,  3.01it/s]
                                                   

 62%|██████▏   | 3577/5758 [20:21<12:04,  3.01it/s]
 62%|██████▏   | 3578/5758 [20:21<11:43,  3.10it/s]
                                                   

 62%|██████▏   | 3578/5758 [20:21<11:43,  3.10it/s]
 62%|██████▏   | 3579/5758 [20:21<11:37,  3.12it/s]
                                                   

 62%|██████▏   | 3579/5758 [20:21<11:37,  3.12it/s]
 62%|██████▏   | 3580/5758 [20:22<13:11,  2.75it/s]
                                                   

 62%|██████▏   | 3580/5758 [20:22<13:11,  2.75it/s]
 62%|██████▏   | 3581/5758 [20:22<12:28,  2.91it/s]
                                                   

 62%|██████▏   | 3581/5758 [20:22<12:28,  2.91it/s]
 62%|██████▏   | 3582/5758 [20:22<12:44,  2.84it/s]
                                                   

 62%|██████▏   | 3582/5758 [20:22<12:44,  2.84it/s]
 62%|██████▏   | 3583/5758 [20:23<12:09,  2.98it/s]
                                                   

 62%|██████▏   | 3583/5758 [20:23<12:09,  2.98it/s]
 62%|██████▏   | 3584/5758 [20:23<11:48,  3.07it/s]
                                                   

 62%|██████▏   | 3584/5758 [20:23<11:48,  3.07it/s]
 62%|██████▏   | 3585/5758 [20:23<12:06,  2.99it/s]
                                                   

 62%|██████▏   | 3585/5758 [20:23<12:06,  2.99it/s]
 62%|██████▏   | 3586/5758 [20:24<11:43,  3.09it/s]
                                                   

 62%|██████▏   | 3586/5758 [20:24<11:43,  3.09it/s]
 62%|██████▏   | 3587/5758 [20:24<11:32,  3.14it/s]
                                                   

 62%|██████▏   | 3587/5758 [20:24<11:32,  3.14it/s]
 62%|██████▏   | 3588/5758 [20:24<11:25,  3.17it/s]
                                                   

 62%|██████▏   | 3588/5758 [20:24<11:25,  3.17it/s]
 62%|██████▏   | 3589/5758 [20:25<11:18,  3.20it/s]
                                                   

 62%|██████▏   | 3589/5758 [20:25<11:18,  3.20it/s]
 62%|██████▏   | 3590/5758 [20:25<11:10,  3.23it/s]
                                                   

 62%|██████▏   | 3590/5758 [20:25<11:10,  3.23it/s]
 62%|██████▏   | 3591/5758 [20:25<11:10,  3.23it/s]
                                                   

 62%|██████▏   | 3591/5758 [20:25<11:10,  3.23it/s]
 62%|██████▏   | 3592/5758 [20:26<11:02,  3.27it/s]
                                                   

 62%|██████▏   | 3592/5758 [20:26<11:02,  3.27it/s]
 62%|██████▏   | 3593/5758 [20:26<11:12,  3.22it/s]
                                                   

 62%|██████▏   | 3593/5758 [20:26<11:12,  3.22it/s]
 62%|██████▏   | 3594/5758 [20:26<11:59,  3.01it/s]
                                                   

 62%|██████▏   | 3594/5758 [20:26<11:59,  3.01it/s]
 62%|██████▏   | 3595/5758 [20:27<12:30,  2.88it/s]
                                                   

 62%|██████▏   | 3595/5758 [20:27<12:30,  2.88it/s]
 62%|██████▏   | 3596/5758 [20:27<12:19,  2.92it/s]
                                                   

 62%|██████▏   | 3596/5758 [20:27<12:19,  2.92it/s]
 62%|██████▏   | 3597/5758 [20:27<11:52,  3.03it/s]
                                                   

 62%|██████▏   | 3597/5758 [20:27<11:52,  3.03it/s]
 62%|██████▏   | 3598/5758 [20:28<11:42,  3.07it/s]
                                                   

 62%|██████▏   | 3598/5758 [20:28<11:42,  3.07it/s]
 63%|██████▎   | 3599/5758 [20:28<12:13,  2.95it/s]
                                                   

 63%|██████▎   | 3599/5758 [20:28<12:13,  2.95it/s]
 63%|██████▎   | 3600/5758 [20:28<12:07,  2.97it/s]
                                                   

 63%|██████▎   | 3600/5758 [20:28<12:07,  2.97it/s]
 63%|██████▎   | 3601/5758 [20:29<12:17,  2.93it/s]
                                                   

 63%|██████▎   | 3601/5758 [20:29<12:17,  2.93it/s]
 63%|██████▎   | 3602/5758 [20:29<12:12,  2.94it/s]
                                                   

 63%|██████▎   | 3602/5758 [20:29<12:12,  2.94it/s]
 63%|██████▎   | 3603/5758 [20:29<11:56,  3.01it/s]
                                                   

 63%|██████▎   | 3603/5758 [20:29<11:56,  3.01it/s]
 63%|██████▎   | 3604/5758 [20:30<12:34,  2.86it/s]
                                                   

 63%|██████▎   | 3604/5758 [20:30<12:34,  2.86it/s]
 63%|██████▎   | 3605/5758 [20:30<12:38,  2.84it/s]
                                                   

 63%|██████▎   | 3605/5758 [20:30<12:38,  2.84it/s]
 63%|██████▎   | 3606/5758 [20:31<14:05,  2.55it/s]
                                                   

 63%|██████▎   | 3606/5758 [20:31<14:05,  2.55it/s]
 63%|██████▎   | 3607/5758 [20:31<13:13,  2.71it/s]
                                                   

 63%|██████▎   | 3607/5758 [20:31<13:13,  2.71it/s]
 63%|██████▎   | 3608/5758 [20:31<12:34,  2.85it/s]
                                                   

 63%|██████▎   | 3608/5758 [20:31<12:34,  2.85it/s]
 63%|██████▎   | 3609/5758 [20:31<12:09,  2.94it/s]
                                                   

 63%|██████▎   | 3609/5758 [20:31<12:09,  2.94it/s]
 63%|██████▎   | 3610/5758 [20:32<12:04,  2.97it/s]
                                                   

 63%|██████▎   | 3610/5758 [20:32<12:04,  2.97it/s]
 63%|██████▎   | 3611/5758 [20:32<11:34,  3.09it/s]
                                                   
{'loss': '1.049', 'grad_norm': '11.5', 'learning_rate': '3.826e-06', 'num_tokens': '2.37e+06', 'mean_token_accuracy': '0.7206', 'epoch': '0.6176'}
+{'loss': '1.654', 'grad_norm': '16', 'learning_rate': '3.824e-06', 'num_tokens': '2.371e+06', 'mean_token_accuracy': '0.6355', 'epoch': '0.6177'}
+{'loss': '1.377', 'grad_norm': '14.25', 'learning_rate': '3.823e-06', 'num_tokens': '2.371e+06', 'mean_token_accuracy': '0.6609', 'epoch': '0.6179'}
+{'loss': '0.8632', 'grad_norm': '19.25', 'learning_rate': '3.821e-06', 'num_tokens': '2.372e+06', 'mean_token_accuracy': '0.8017', 'epoch': '0.6181'}
+{'loss': '1.221', 'grad_norm': '7.594', 'learning_rate': '3.819e-06', 'num_tokens': '2.373e+06', 'mean_token_accuracy': '0.6923', 'epoch': '0.6183'}
+{'loss': '2.402', 'grad_norm': '9.062', 'learning_rate': '3.817e-06', 'num_tokens': '2.375e+06', 'mean_token_accuracy': '0.5287', 'epoch': '0.6184'}
+{'loss': '0.6993', 'grad_norm': '12.75', 'learning_rate': '3.816e-06', 'num_tokens': '2.375e+06', 'mean_token_accuracy': '0.8325', 'epoch': '0.6186'}
+{'loss': '2.278', 'grad_norm': '29.62', 'learning_rate': '3.814e-06', 'num_tokens': '2.375e+06', 'mean_token_accuracy': '0.5602', 'epoch': '0.6188'}
+{'loss': '0.6308', 'grad_norm': '7.375', 'learning_rate': '3.812e-06', 'num_tokens': '2.377e+06', 'mean_token_accuracy': '0.8076', 'epoch': '0.619'}
+{'loss': '1.049', 'grad_norm': '14.94', 'learning_rate': '3.81e-06', 'num_tokens': '2.378e+06', 'mean_token_accuracy': '0.7487', 'epoch': '0.6191'}
+{'loss': '0.6327', 'grad_norm': '16.75', 'learning_rate': '3.809e-06', 'num_tokens': '2.378e+06', 'mean_token_accuracy': '0.8544', 'epoch': '0.6193'}
+{'loss': '1.143', 'grad_norm': '16', 'learning_rate': '3.807e-06', 'num_tokens': '2.378e+06', 'mean_token_accuracy': '0.7686', 'epoch': '0.6195'}
+{'loss': '0.9902', 'grad_norm': '12.31', 'learning_rate': '3.805e-06', 'num_tokens': '2.379e+06', 'mean_token_accuracy': '0.725', 'epoch': '0.6197'}
+{'loss': '2.185', 'grad_norm': '44.75', 'learning_rate': '3.803e-06', 'num_tokens': '2.379e+06', 'mean_token_accuracy': '0.6452', 'epoch': '0.6198'}
+{'loss': '0.9953', 'grad_norm': '11.62', 'learning_rate': '3.802e-06', 'num_tokens': '2.38e+06', 'mean_token_accuracy': '0.7075', 'epoch': '0.62'}
+{'loss': '2.089', 'grad_norm': '15.06', 'learning_rate': '3.8e-06', 'num_tokens': '2.38e+06', 'mean_token_accuracy': '0.597', 'epoch': '0.6202'}
+{'loss': '1.209', 'grad_norm': '27.88', 'learning_rate': '3.798e-06', 'num_tokens': '2.38e+06', 'mean_token_accuracy': '0.7368', 'epoch': '0.6204'}
+{'loss': '0.6071', 'grad_norm': '8.25', 'learning_rate': '3.796e-06', 'num_tokens': '2.382e+06', 'mean_token_accuracy': '0.8193', 'epoch': '0.6205'}
+{'loss': '0.891', 'grad_norm': '17.5', 'learning_rate': '3.795e-06', 'num_tokens': '2.382e+06', 'mean_token_accuracy': '0.8171', 'epoch': '0.6207'}
+{'loss': '1.507', 'grad_norm': '16.62', 'learning_rate': '3.793e-06', 'num_tokens': '2.382e+06', 'mean_token_accuracy': '0.6847', 'epoch': '0.6209'}
+{'loss': '1.923', 'grad_norm': '16.88', 'learning_rate': '3.791e-06', 'num_tokens': '2.383e+06', 'mean_token_accuracy': '0.6021', 'epoch': '0.621'}
+{'loss': '1.03', 'grad_norm': '11.88', 'learning_rate': '3.79e-06', 'num_tokens': '2.384e+06', 'mean_token_accuracy': '0.7836', 'epoch': '0.6212'}
+{'loss': '1.247', 'grad_norm': '13.38', 'learning_rate': '3.788e-06', 'num_tokens': '2.384e+06', 'mean_token_accuracy': '0.6604', 'epoch': '0.6214'}
+{'loss': '1.168', 'grad_norm': '27.38', 'learning_rate': '3.786e-06', 'num_tokens': '2.384e+06', 'mean_token_accuracy': '0.7514', 'epoch': '0.6216'}
+{'loss': '1.731', 'grad_norm': '10.31', 'learning_rate': '3.784e-06', 'num_tokens': '2.385e+06', 'mean_token_accuracy': '0.626', 'epoch': '0.6217'}
+{'loss': '1.36', 'grad_norm': '11.62', 'learning_rate': '3.783e-06', 'num_tokens': '2.386e+06', 'mean_token_accuracy': '0.7159', 'epoch': '0.6219'}
+{'loss': '1.634', 'grad_norm': '7.75', 'learning_rate': '3.781e-06', 'num_tokens': '2.388e+06', 'mean_token_accuracy': '0.6398', 'epoch': '0.6221'}
+{'loss': '0.8763', 'grad_norm': '12.75', 'learning_rate': '3.779e-06', 'num_tokens': '2.389e+06', 'mean_token_accuracy': '0.7043', 'epoch': '0.6223'}
+{'loss': '0.6753', 'grad_norm': '7.688', 'learning_rate': '3.777e-06', 'num_tokens': '2.39e+06', 'mean_token_accuracy': '0.8083', 'epoch': '0.6224'}
+{'loss': '1.337', 'grad_norm': '11', 'learning_rate': '3.776e-06', 'num_tokens': '2.391e+06', 'mean_token_accuracy': '0.6934', 'epoch': '0.6226'}
+{'loss': '0.6107', 'grad_norm': '14.94', 'learning_rate': '3.774e-06', 'num_tokens': '2.391e+06', 'mean_token_accuracy': '0.8424', 'epoch': '0.6228'}
+{'loss': '1.653', 'grad_norm': '16', 'learning_rate': '3.772e-06', 'num_tokens': '2.392e+06', 'mean_token_accuracy': '0.6301', 'epoch': '0.623'}
+{'loss': '1.02', 'grad_norm': '11.56', 'learning_rate': '3.77e-06', 'num_tokens': '2.392e+06', 'mean_token_accuracy': '0.726', 'epoch': '0.6231'}
+{'loss': '1.062', 'grad_norm': '12.75', 'learning_rate': '3.769e-06', 'num_tokens': '2.393e+06', 'mean_token_accuracy': '0.6756', 'epoch': '0.6233'}
+{'loss': '2.292', 'grad_norm': '9.188', 'learning_rate': '3.767e-06', 'num_tokens': '2.394e+06', 'mean_token_accuracy': '0.5578', 'epoch': '0.6235'}
+{'loss': '1.451', 'grad_norm': '19.25', 'learning_rate': '3.765e-06', 'num_tokens': '2.395e+06', 'mean_token_accuracy': '0.6762', 'epoch': '0.6237'}
+{'loss': '1.159', 'grad_norm': '13.44', 'learning_rate': '3.763e-06', 'num_tokens': '2.395e+06', 'mean_token_accuracy': '0.674', 'epoch': '0.6238'}
+{'loss': '1.115', 'grad_norm': '11.56', 'learning_rate': '3.762e-06', 'num_tokens': '2.396e+06', 'mean_token_accuracy': '0.7022', 'epoch': '0.624'}
+{'loss': '1.126', 'grad_norm': '14.19', 'learning_rate': '3.76e-06', 'num_tokens': '2.397e+06', 'mean_token_accuracy': '0.6962', 'epoch': '0.6242'}
+{'loss': '1.005', 'grad_norm': '9.75', 'learning_rate': '3.758e-06', 'num_tokens': '2.398e+06', 'mean_token_accuracy': '0.7365', 'epoch': '0.6243'}
+{'loss': '1.953', 'grad_norm': '11.69', 'learning_rate': '3.757e-06', 'num_tokens': '2.399e+06', 'mean_token_accuracy': '0.593', 'epoch': '0.6245'}
+{'loss': '0.8157', 'grad_norm': '16.5', 'learning_rate': '3.755e-06', 'num_tokens': '2.399e+06', 'mean_token_accuracy': '0.7994', 'epoch': '0.6247'}
+{'loss': '0.7429', 'grad_norm': '15.69', 'learning_rate': '3.753e-06', 'num_tokens': '2.399e+06', 'mean_token_accuracy': '0.8029', 'epoch': '0.6249'}
+{'loss': '1.464', 'grad_norm': '15.25', 'learning_rate': '3.751e-06', 'num_tokens': '2.4e+06', 'mean_token_accuracy': '0.6339', 'epoch': '0.625'}
+{'loss': '1.156', 'grad_norm': '10.94', 'learning_rate': '3.75e-06', 'num_tokens': '2.401e+06', 'mean_token_accuracy': '0.709', 'epoch': '0.6252'}
+{'loss': '1.009', 'grad_norm': '26', 'learning_rate': '3.748e-06', 'num_tokens': '2.401e+06', 'mean_token_accuracy': '0.7438', 'epoch': '0.6254'}
+{'loss': '0.9644', 'grad_norm': '18.62', 'learning_rate': '3.746e-06', 'num_tokens': '2.401e+06', 'mean_token_accuracy': '0.7926', 'epoch': '0.6256'}
+{'loss': '1.683', 'grad_norm': '12.75', 'learning_rate': '3.744e-06', 'num_tokens': '2.402e+06', 'mean_token_accuracy': '0.6269', 'epoch': '0.6257'}
+{'loss': '1.217', 'grad_norm': '12.94', 'learning_rate': '3.743e-06', 'num_tokens': '2.402e+06', 'mean_token_accuracy': '0.7257', 'epoch': '0.6259'}
+{'loss': '1.122', 'grad_norm': '14.19', 'learning_rate': '3.741e-06', 'num_tokens': '2.403e+06', 'mean_token_accuracy': '0.7319', 'epoch': '0.6261'}
+{'loss': '0.762', 'grad_norm': '24.88', 'learning_rate': '3.739e-06', 'num_tokens': '2.403e+06', 'mean_token_accuracy': '0.8017', 'epoch': '0.6263'}
+{'loss': '1.781', 'grad_norm': '12', 'learning_rate': '3.737e-06', 'num_tokens': '2.404e+06', 'mean_token_accuracy': '0.6', 'epoch': '0.6264'}
+{'loss': '1.262', 'grad_norm': '12.44', 'learning_rate': '3.736e-06', 'num_tokens': '2.405e+06', 'mean_token_accuracy': '0.6463', 'epoch': '0.6266'}
+{'loss': '0.6879', 'grad_norm': '7.531', 'learning_rate': '3.734e-06', 'num_tokens': '2.406e+06', 'mean_token_accuracy': '0.7996', 'epoch': '0.6268'}
+{'loss': '0.9046', 'grad_norm': '11.81', 'learning_rate': '3.732e-06', 'num_tokens': '2.407e+06', 'mean_token_accuracy': '0.7241', 'epoch': '0.627'}
+
 63%|██████▎   | 3611/5758 [20:32<11:34,  3.09it/s]
 63%|██████▎   | 3612/5758 [20:32<11:19,  3.16it/s]
                                                   

 63%|██████▎   | 3612/5758 [20:32<11:19,  3.16it/s]
 63%|██████▎   | 3613/5758 [20:33<11:18,  3.16it/s]
                                                   

 63%|██████▎   | 3613/5758 [20:33<11:18,  3.16it/s]
 63%|██████▎   | 3614/5758 [20:33<11:44,  3.04it/s]
                                                   

 63%|██████▎   | 3614/5758 [20:33<11:44,  3.04it/s]
 63%|██████▎   | 3615/5758 [20:33<11:40,  3.06it/s]
                                                   

 63%|██████▎   | 3615/5758 [20:33<11:40,  3.06it/s]
 63%|██████▎   | 3616/5758 [20:34<11:35,  3.08it/s]
                                                   

 63%|██████▎   | 3616/5758 [20:34<11:35,  3.08it/s]
 63%|██████▎   | 3617/5758 [20:34<12:02,  2.96it/s]
                                                   

 63%|██████▎   | 3617/5758 [20:34<12:02,  2.96it/s]
 63%|██████▎   | 3618/5758 [20:34<12:06,  2.95it/s]
                                                   

 63%|██████▎   | 3618/5758 [20:34<12:06,  2.95it/s]
 63%|██████▎   | 3619/5758 [20:35<12:30,  2.85it/s]
                                                   

 63%|██████▎   | 3619/5758 [20:35<12:30,  2.85it/s]
 63%|██████▎   | 3620/5758 [20:35<12:22,  2.88it/s]
                                                   

 63%|██████▎   | 3620/5758 [20:35<12:22,  2.88it/s]
 63%|██████▎   | 3621/5758 [20:36<12:43,  2.80it/s]
                                                   

 63%|██████▎   | 3621/5758 [20:36<12:43,  2.80it/s]
 63%|██████▎   | 3622/5758 [20:36<12:51,  2.77it/s]
                                                   

 63%|██████▎   | 3622/5758 [20:36<12:51,  2.77it/s]
 63%|██████▎   | 3623/5758 [20:36<13:10,  2.70it/s]
                                                   

 63%|██████▎   | 3623/5758 [20:36<13:10,  2.70it/s]
 63%|██████▎   | 3624/5758 [20:37<13:12,  2.69it/s]
                                                   

 63%|██████▎   | 3624/5758 [20:37<13:12,  2.69it/s]
 63%|██████▎   | 3625/5758 [20:37<12:54,  2.76it/s]
                                                   

 63%|██████▎   | 3625/5758 [20:37<12:54,  2.76it/s]
 63%|██████▎   | 3626/5758 [20:37<13:02,  2.72it/s]
                                                   

 63%|██████▎   | 3626/5758 [20:37<13:02,  2.72it/s]
 63%|██████▎   | 3627/5758 [20:38<14:12,  2.50it/s]
                                                   

 63%|██████▎   | 3627/5758 [20:38<14:12,  2.50it/s]
 63%|██████▎   | 3628/5758 [20:38<14:21,  2.47it/s]
                                                   

 63%|██████▎   | 3628/5758 [20:38<14:21,  2.47it/s]
 63%|██████▎   | 3629/5758 [20:39<14:11,  2.50it/s]
                                                   

 63%|██████▎   | 3629/5758 [20:39<14:11,  2.50it/s]
 63%|██████▎   | 3630/5758 [20:39<12:51,  2.76it/s]
                                                   

 63%|██████▎   | 3630/5758 [20:39<12:51,  2.76it/s]
 63%|██████▎   | 3631/5758 [20:39<12:17,  2.89it/s]
                                                   

 63%|██████▎   | 3631/5758 [20:39<12:17,  2.89it/s]
 63%|██████▎   | 3632/5758 [20:40<13:52,  2.55it/s]
                                                   

 63%|██████▎   | 3632/5758 [20:40<13:52,  2.55it/s]
 63%|██████▎   | 3633/5758 [20:40<12:56,  2.74it/s]
                                                   

 63%|██████▎   | 3633/5758 [20:40<12:56,  2.74it/s]
 63%|██████▎   | 3634/5758 [20:40<12:28,  2.84it/s]
                                                   

 63%|██████▎   | 3634/5758 [20:40<12:28,  2.84it/s]
 63%|██████▎   | 3635/5758 [20:41<11:59,  2.95it/s]
                                                   

 63%|██████▎   | 3635/5758 [20:41<11:59,  2.95it/s]
 63%|██████▎   | 3636/5758 [20:41<11:55,  2.97it/s]
                                                   

 63%|██████▎   | 3636/5758 [20:41<11:55,  2.97it/s]
 63%|██████▎   | 3637/5758 [20:41<12:07,  2.92it/s]
                                                   

 63%|██████▎   | 3637/5758 [20:41<12:07,  2.92it/s]
 63%|██████▎   | 3638/5758 [20:42<11:37,  3.04it/s]
                                                   

 63%|██████▎   | 3638/5758 [20:42<11:37,  3.04it/s]
 63%|██████▎   | 3639/5758 [20:42<12:12,  2.89it/s]
                                                   

 63%|██████▎   | 3639/5758 [20:42<12:12,  2.89it/s]
 63%|██████▎   | 3640/5758 [20:42<11:57,  2.95it/s]
                                                   

 63%|██████▎   | 3640/5758 [20:42<11:57,  2.95it/s]
 63%|██████▎   | 3641/5758 [20:43<12:00,  2.94it/s]
                                                   

 63%|██████▎   | 3641/5758 [20:43<12:00,  2.94it/s]
 63%|██████▎   | 3642/5758 [20:43<11:39,  3.02it/s]
                                                   

 63%|██████▎   | 3642/5758 [20:43<11:39,  3.02it/s]
 63%|██████▎   | 3643/5758 [20:43<11:56,  2.95it/s]
                                                   

 63%|██████▎   | 3643/5758 [20:43<11:56,  2.95it/s]
 63%|██████▎   | 3644/5758 [20:44<11:40,  3.02it/s]
                                                   

 63%|██████▎   | 3644/5758 [20:44<11:40,  3.02it/s]
 63%|██████▎   | 3645/5758 [20:44<12:02,  2.93it/s]
                                                   

 63%|██████▎   | 3645/5758 [20:44<12:02,  2.93it/s]
 63%|██████▎   | 3646/5758 [20:44<12:07,  2.90it/s]
                                                   

 63%|██████▎   | 3646/5758 [20:44<12:07,  2.90it/s]
 63%|██████▎   | 3647/5758 [20:45<11:36,  3.03it/s]
                                                   

 63%|██████▎   | 3647/5758 [20:45<11:36,  3.03it/s]
 63%|██████▎   | 3648/5758 [20:45<11:14,  3.13it/s]
                                                   

 63%|██████▎   | 3648/5758 [20:45<11:14,  3.13it/s]
 63%|██████▎   | 3649/5758 [20:45<11:40,  3.01it/s]
                                                   

 63%|██████▎   | 3649/5758 [20:45<11:40,  3.01it/s]
 63%|██████▎   | 3650/5758 [20:46<11:16,  3.12it/s]
                                                   

 63%|██████▎   | 3650/5758 [20:46<11:16,  3.12it/s]
 63%|██████▎   | 3651/5758 [20:46<11:03,  3.18it/s]
                                                   

 63%|██████▎   | 3651/5758 [20:46<11:03,  3.18it/s]
 63%|██████▎   | 3652/5758 [20:46<11:07,  3.16it/s]
                                                   

 63%|██████▎   | 3652/5758 [20:46<11:07,  3.16it/s]
 63%|██████▎   | 3653/5758 [20:47<10:55,  3.21it/s]
                                                   

 63%|██████▎   | 3653/5758 [20:47<10:55,  3.21it/s]
 63%|██████▎   | 3654/5758 [20:47<10:43,  3.27it/s]
                                                   

 63%|██████▎   | 3654/5758 [20:47<10:43,  3.27it/s]
 63%|██████▎   | 3655/5758 [20:47<10:37,  3.30it/s]
                                                   

 63%|██████▎   | 3655/5758 [20:47<10:37,  3.30it/s]
 63%|██████▎   | 3656/5758 [20:47<10:54,  3.21it/s]
                                                   

 63%|██████▎   | 3656/5758 [20:47<10:54,  3.21it/s]
 64%|██████▎   | 3657/5758 [20:48<10:44,  3.26it/s]
                                                   

 64%|██████▎   | 3657/5758 [20:48<10:44,  3.26it/s]
 64%|██████▎   | 3658/5758 [20:48<11:56,  2.93it/s]
                                                   

 64%|██████▎   | 3658/5758 [20:48<11:56,  2.93it/s]
 64%|██████▎   | 3659/5758 [20:49<11:59,  2.92it/s]
                                                   

 64%|██████▎   | 3659/5758 [20:49<11:59,  2.92it/s]
 64%|██████▎   | 3660/5758 [20:49<12:14,  2.86it/s]
                                                   

 64%|██████▎   | 3660/5758 [20:49<12:14,  2.86it/s]
 64%|██████▎   | 3661/5758 [20:49<11:58,  2.92it/s]
                                                   

 64%|██████▎   | 3661/5758 [20:49<11:58,  2.92it/s]
 64%|██████▎   | 3662/5758 [20:50<11:25,  3.06it/s]
                                                   

 64%|██████▎   | 3662/5758 [20:50<11:25,  3.06it/s]
 64%|██████▎   | 3663/5758 [20:50<11:24,  3.06it/s]
                                                   

 64%|██████▎   | 3663/5758 [20:50<11:24,  3.06it/s]
 64%|██████▎   | 3664/5758 [20:50<11:45,  2.97it/s]
                                                   

 64%|██████▎   | 3664/5758 [20:50<11:45,  2.97it/s]
 64%|██████▎   | 3665/5758 [20:51<12:07,  2.88it/s]
                                                   

 64%|██████▎   | 3665/5758 [20:51<12:07,  2.88it/s]
 64%|██████▎   | 3666/5758 [20:51<11:43,  2.98it/s]
                                                   
{'loss': '1.253', 'grad_norm': '15.38', 'learning_rate': '3.73e-06', 'num_tokens': '2.407e+06', 'mean_token_accuracy': '0.6682', 'epoch': '0.6271'}
+{'loss': '1.005', 'grad_norm': '13', 'learning_rate': '3.729e-06', 'num_tokens': '2.408e+06', 'mean_token_accuracy': '0.7117', 'epoch': '0.6273'}
+{'loss': '1.781', 'grad_norm': '16', 'learning_rate': '3.727e-06', 'num_tokens': '2.409e+06', 'mean_token_accuracy': '0.6436', 'epoch': '0.6275'}
+{'loss': '2.144', 'grad_norm': '22.38', 'learning_rate': '3.725e-06', 'num_tokens': '2.409e+06', 'mean_token_accuracy': '0.5777', 'epoch': '0.6276'}
+{'loss': '1.442', 'grad_norm': '15.81', 'learning_rate': '3.724e-06', 'num_tokens': '2.41e+06', 'mean_token_accuracy': '0.6923', 'epoch': '0.6278'}
+{'loss': '1.38', 'grad_norm': '13.75', 'learning_rate': '3.722e-06', 'num_tokens': '2.41e+06', 'mean_token_accuracy': '0.6095', 'epoch': '0.628'}
+{'loss': '2.366', 'grad_norm': '22.12', 'learning_rate': '3.72e-06', 'num_tokens': '2.411e+06', 'mean_token_accuracy': '0.5714', 'epoch': '0.6282'}
+{'loss': '1.13', 'grad_norm': '11.69', 'learning_rate': '3.718e-06', 'num_tokens': '2.411e+06', 'mean_token_accuracy': '0.7', 'epoch': '0.6283'}
+{'loss': '0.935', 'grad_norm': '13.81', 'learning_rate': '3.717e-06', 'num_tokens': '2.412e+06', 'mean_token_accuracy': '0.7306', 'epoch': '0.6285'}
+{'loss': '1.758', 'grad_norm': '10.31', 'learning_rate': '3.715e-06', 'num_tokens': '2.413e+06', 'mean_token_accuracy': '0.6284', 'epoch': '0.6287'}
+{'loss': '0.7018', 'grad_norm': '9.625', 'learning_rate': '3.713e-06', 'num_tokens': '2.414e+06', 'mean_token_accuracy': '0.7785', 'epoch': '0.6289'}
+{'loss': '1.479', 'grad_norm': '11.69', 'learning_rate': '3.711e-06', 'num_tokens': '2.415e+06', 'mean_token_accuracy': '0.6053', 'epoch': '0.629'}
+{'loss': '2.058', 'grad_norm': '21.25', 'learning_rate': '3.71e-06', 'num_tokens': '2.415e+06', 'mean_token_accuracy': '0.5669', 'epoch': '0.6292'}
+{'loss': '1.575', 'grad_norm': '37.25', 'learning_rate': '3.708e-06', 'num_tokens': '2.416e+06', 'mean_token_accuracy': '0.713', 'epoch': '0.6294'}
+{'loss': '1.343', 'grad_norm': '56.5', 'learning_rate': '3.706e-06', 'num_tokens': '2.416e+06', 'mean_token_accuracy': '0.7429', 'epoch': '0.6296'}
+{'loss': '1.193', 'grad_norm': '12.5', 'learning_rate': '3.704e-06', 'num_tokens': '2.416e+06', 'mean_token_accuracy': '0.6758', 'epoch': '0.6297'}
+{'loss': '1.692', 'grad_norm': '17', 'learning_rate': '3.703e-06', 'num_tokens': '2.417e+06', 'mean_token_accuracy': '0.6306', 'epoch': '0.6299'}
+{'loss': '1.713', 'grad_norm': '14.25', 'learning_rate': '3.701e-06', 'num_tokens': '2.418e+06', 'mean_token_accuracy': '0.6419', 'epoch': '0.6301'}
+{'loss': '0.9001', 'grad_norm': '12.56', 'learning_rate': '3.699e-06', 'num_tokens': '2.418e+06', 'mean_token_accuracy': '0.7847', 'epoch': '0.6303'}
+{'loss': '1.653', 'grad_norm': '69.5', 'learning_rate': '3.697e-06', 'num_tokens': '2.418e+06', 'mean_token_accuracy': '0.6935', 'epoch': '0.6304'}
+{'loss': '1.29', 'grad_norm': '14.06', 'learning_rate': '3.696e-06', 'num_tokens': '2.419e+06', 'mean_token_accuracy': '0.6402', 'epoch': '0.6306'}
+{'loss': '0.9441', 'grad_norm': '14.69', 'learning_rate': '3.694e-06', 'num_tokens': '2.419e+06', 'mean_token_accuracy': '0.7186', 'epoch': '0.6308'}
+{'loss': '0.8137', 'grad_norm': '12', 'learning_rate': '3.692e-06', 'num_tokens': '2.42e+06', 'mean_token_accuracy': '0.7452', 'epoch': '0.6309'}
+{'loss': '1.391', 'grad_norm': '11.94', 'learning_rate': '3.691e-06', 'num_tokens': '2.421e+06', 'mean_token_accuracy': '0.6299', 'epoch': '0.6311'}
+{'loss': '2.213', 'grad_norm': '18', 'learning_rate': '3.689e-06', 'num_tokens': '2.422e+06', 'mean_token_accuracy': '0.5683', 'epoch': '0.6313'}
+{'loss': '1.469', 'grad_norm': '21.75', 'learning_rate': '3.687e-06', 'num_tokens': '2.422e+06', 'mean_token_accuracy': '0.7156', 'epoch': '0.6315'}
+{'loss': '1.249', 'grad_norm': '40.75', 'learning_rate': '3.685e-06', 'num_tokens': '2.422e+06', 'mean_token_accuracy': '0.7909', 'epoch': '0.6316'}
+{'loss': '1.258', 'grad_norm': '14.12', 'learning_rate': '3.684e-06', 'num_tokens': '2.423e+06', 'mean_token_accuracy': '0.6539', 'epoch': '0.6318'}
+{'loss': '1.231', 'grad_norm': '14.56', 'learning_rate': '3.682e-06', 'num_tokens': '2.423e+06', 'mean_token_accuracy': '0.6632', 'epoch': '0.632'}
+{'loss': '2.225', 'grad_norm': '17', 'learning_rate': '3.68e-06', 'num_tokens': '2.424e+06', 'mean_token_accuracy': '0.5445', 'epoch': '0.6322'}
+{'loss': '2.027', 'grad_norm': '18', 'learning_rate': '3.678e-06', 'num_tokens': '2.424e+06', 'mean_token_accuracy': '0.6139', 'epoch': '0.6323'}
+{'loss': '1.176', 'grad_norm': '36.25', 'learning_rate': '3.677e-06', 'num_tokens': '2.424e+06', 'mean_token_accuracy': '0.7317', 'epoch': '0.6325'}
+{'loss': '1.022', 'grad_norm': '9.375', 'learning_rate': '3.675e-06', 'num_tokens': '2.425e+06', 'mean_token_accuracy': '0.7577', 'epoch': '0.6327'}
+{'loss': '1.194', 'grad_norm': '11.62', 'learning_rate': '3.673e-06', 'num_tokens': '2.426e+06', 'mean_token_accuracy': '0.7227', 'epoch': '0.6329'}
+{'loss': '1.316', 'grad_norm': '9.625', 'learning_rate': '3.671e-06', 'num_tokens': '2.427e+06', 'mean_token_accuracy': '0.7062', 'epoch': '0.633'}
+{'loss': '1.874', 'grad_norm': '10.19', 'learning_rate': '3.67e-06', 'num_tokens': '2.428e+06', 'mean_token_accuracy': '0.6074', 'epoch': '0.6332'}
+{'loss': '1.268', 'grad_norm': '16.5', 'learning_rate': '3.668e-06', 'num_tokens': '2.428e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.6334'}
+{'loss': '0.9736', 'grad_norm': '13.38', 'learning_rate': '3.666e-06', 'num_tokens': '2.429e+06', 'mean_token_accuracy': '0.7121', 'epoch': '0.6336'}
+{'loss': '1.593', 'grad_norm': '12.25', 'learning_rate': '3.664e-06', 'num_tokens': '2.43e+06', 'mean_token_accuracy': '0.5822', 'epoch': '0.6337'}
+{'loss': '1.262', 'grad_norm': '13.62', 'learning_rate': '3.663e-06', 'num_tokens': '2.43e+06', 'mean_token_accuracy': '0.6641', 'epoch': '0.6339'}
+{'loss': '1.526', 'grad_norm': '60.75', 'learning_rate': '3.661e-06', 'num_tokens': '2.431e+06', 'mean_token_accuracy': '0.6753', 'epoch': '0.6341'}
+{'loss': '1.339', 'grad_norm': '13.19', 'learning_rate': '3.659e-06', 'num_tokens': '2.431e+06', 'mean_token_accuracy': '0.6399', 'epoch': '0.6342'}
+{'loss': '1.536', 'grad_norm': '6.812', 'learning_rate': '3.658e-06', 'num_tokens': '2.434e+06', 'mean_token_accuracy': '0.6407', 'epoch': '0.6344'}
+{'loss': '1.055', 'grad_norm': '13', 'learning_rate': '3.656e-06', 'num_tokens': '2.435e+06', 'mean_token_accuracy': '0.707', 'epoch': '0.6346'}
+{'loss': '0.6389', 'grad_norm': '15.31', 'learning_rate': '3.654e-06', 'num_tokens': '2.435e+06', 'mean_token_accuracy': '0.8596', 'epoch': '0.6348'}
+{'loss': '1.27', 'grad_norm': '22.25', 'learning_rate': '3.652e-06', 'num_tokens': '2.435e+06', 'mean_token_accuracy': '0.7056', 'epoch': '0.6349'}
+{'loss': '1.69', 'grad_norm': '12.94', 'learning_rate': '3.651e-06', 'num_tokens': '2.436e+06', 'mean_token_accuracy': '0.6575', 'epoch': '0.6351'}
+{'loss': '1.346', 'grad_norm': '13.25', 'learning_rate': '3.649e-06', 'num_tokens': '2.437e+06', 'mean_token_accuracy': '0.6434', 'epoch': '0.6353'}
+{'loss': '1.946', 'grad_norm': '20.25', 'learning_rate': '3.647e-06', 'num_tokens': '2.437e+06', 'mean_token_accuracy': '0.6176', 'epoch': '0.6355'}
+{'loss': '1.788', 'grad_norm': '22.5', 'learning_rate': '3.645e-06', 'num_tokens': '2.437e+06', 'mean_token_accuracy': '0.6151', 'epoch': '0.6356'}
+{'loss': '1.149', 'grad_norm': '10.94', 'learning_rate': '3.644e-06', 'num_tokens': '2.438e+06', 'mean_token_accuracy': '0.7182', 'epoch': '0.6358'}
+{'loss': '0.8232', 'grad_norm': '12.81', 'learning_rate': '3.642e-06', 'num_tokens': '2.438e+06', 'mean_token_accuracy': '0.782', 'epoch': '0.636'}
+{'loss': '2.118', 'grad_norm': '13.44', 'learning_rate': '3.64e-06', 'num_tokens': '2.439e+06', 'mean_token_accuracy': '0.5624', 'epoch': '0.6362'}
+{'loss': '0.4642', 'grad_norm': '28', 'learning_rate': '3.638e-06', 'num_tokens': '2.439e+06', 'mean_token_accuracy': '0.8583', 'epoch': '0.6363'}
+{'loss': '0.8621', 'grad_norm': '19.62', 'learning_rate': '3.637e-06', 'num_tokens': '2.44e+06', 'mean_token_accuracy': '0.8085', 'epoch': '0.6365'}
+
 64%|██████▎   | 3666/5758 [20:51<11:43,  2.98it/s]
 64%|██████▎   | 3667/5758 [20:51<11:38,  2.99it/s]
                                                   

 64%|██████▎   | 3667/5758 [20:51<11:38,  2.99it/s]
 64%|██████▎   | 3668/5758 [20:52<11:12,  3.11it/s]
                                                   

 64%|██████▎   | 3668/5758 [20:52<11:12,  3.11it/s]
 64%|██████▎   | 3669/5758 [20:52<11:14,  3.10it/s]
                                                   

 64%|██████▎   | 3669/5758 [20:52<11:14,  3.10it/s]
 64%|██████▎   | 3670/5758 [20:52<10:53,  3.20it/s]
                                                   

 64%|██████▎   | 3670/5758 [20:52<10:53,  3.20it/s]
 64%|██████▍   | 3671/5758 [20:53<11:37,  2.99it/s]
                                                   

 64%|██████▍   | 3671/5758 [20:53<11:37,  2.99it/s]
 64%|██████▍   | 3672/5758 [20:53<11:11,  3.11it/s]
                                                   

 64%|██████▍   | 3672/5758 [20:53<11:11,  3.11it/s]
 64%|██████▍   | 3673/5758 [20:53<10:56,  3.17it/s]
                                                   

 64%|██████▍   | 3673/5758 [20:53<10:56,  3.17it/s]
 64%|██████▍   | 3674/5758 [20:53<10:43,  3.24it/s]
                                                   

 64%|██████▍   | 3674/5758 [20:53<10:43,  3.24it/s]
 64%|██████▍   | 3675/5758 [20:54<10:44,  3.23it/s]
                                                   

 64%|██████▍   | 3675/5758 [20:54<10:44,  3.23it/s]
 64%|██████▍   | 3676/5758 [20:54<10:47,  3.22it/s]
                                                   

 64%|██████▍   | 3676/5758 [20:54<10:47,  3.22it/s]
 64%|██████▍   | 3677/5758 [20:54<11:34,  3.00it/s]
                                                   

 64%|██████▍   | 3677/5758 [20:54<11:34,  3.00it/s]
 64%|██████▍   | 3678/5758 [20:55<11:28,  3.02it/s]
                                                   

 64%|██████▍   | 3678/5758 [20:55<11:28,  3.02it/s]
 64%|██████▍   | 3679/5758 [20:55<11:35,  2.99it/s]
                                                   

 64%|██████▍   | 3679/5758 [20:55<11:35,  2.99it/s]
 64%|██████▍   | 3680/5758 [20:55<12:21,  2.80it/s]
                                                   

 64%|██████▍   | 3680/5758 [20:55<12:21,  2.80it/s]
 64%|██████▍   | 3681/5758 [20:56<12:17,  2.82it/s]
                                                   

 64%|██████▍   | 3681/5758 [20:56<12:17,  2.82it/s]
 64%|██████▍   | 3682/5758 [20:56<12:18,  2.81it/s]
                                                   

 64%|██████▍   | 3682/5758 [20:56<12:18,  2.81it/s]
 64%|██████▍   | 3683/5758 [20:57<12:31,  2.76it/s]
                                                   

 64%|██████▍   | 3683/5758 [20:57<12:31,  2.76it/s]
 64%|██████▍   | 3684/5758 [20:57<12:30,  2.76it/s]
                                                   

 64%|██████▍   | 3684/5758 [20:57<12:30,  2.76it/s]
 64%|██████▍   | 3685/5758 [20:58<14:37,  2.36it/s]
                                                   

 64%|██████▍   | 3685/5758 [20:58<14:37,  2.36it/s]
 64%|██████▍   | 3686/5758 [20:58<13:57,  2.47it/s]
                                                   

 64%|██████▍   | 3686/5758 [20:58<13:57,  2.47it/s]
 64%|██████▍   | 3687/5758 [20:58<13:00,  2.65it/s]
                                                   

 64%|██████▍   | 3687/5758 [20:58<13:00,  2.65it/s]
 64%|██████▍   | 3688/5758 [20:59<12:37,  2.73it/s]
                                                   

 64%|██████▍   | 3688/5758 [20:59<12:37,  2.73it/s]
 64%|██████▍   | 3689/5758 [20:59<12:07,  2.84it/s]
                                                   

 64%|██████▍   | 3689/5758 [20:59<12:07,  2.84it/s]
 64%|██████▍   | 3690/5758 [20:59<11:45,  2.93it/s]
                                                   

 64%|██████▍   | 3690/5758 [20:59<11:45,  2.93it/s]
 64%|██████▍   | 3691/5758 [20:59<11:32,  2.99it/s]
                                                   

 64%|██████▍   | 3691/5758 [20:59<11:32,  2.99it/s]
 64%|██████▍   | 3692/5758 [21:00<11:37,  2.96it/s]
                                                   

 64%|██████▍   | 3692/5758 [21:00<11:37,  2.96it/s]
 64%|██████▍   | 3693/5758 [21:00<11:25,  3.01it/s]
                                                   

 64%|██████▍   | 3693/5758 [21:00<11:25,  3.01it/s]
 64%|██████▍   | 3694/5758 [21:00<11:23,  3.02it/s]
                                                   

 64%|██████▍   | 3694/5758 [21:00<11:23,  3.02it/s]
 64%|██████▍   | 3695/5758 [21:01<11:12,  3.07it/s]
                                                   

 64%|██████▍   | 3695/5758 [21:01<11:12,  3.07it/s]
 64%|██████▍   | 3696/5758 [21:01<11:11,  3.07it/s]
                                                   

 64%|██████▍   | 3696/5758 [21:01<11:11,  3.07it/s]
 64%|██████▍   | 3697/5758 [21:01<11:12,  3.06it/s]
                                                   

 64%|██████▍   | 3697/5758 [21:01<11:12,  3.06it/s]
 64%|██████▍   | 3698/5758 [21:02<11:16,  3.05it/s]
                                                   

 64%|██████▍   | 3698/5758 [21:02<11:16,  3.05it/s]
 64%|██████▍   | 3699/5758 [21:02<11:31,  2.98it/s]
                                                   

 64%|██████▍   | 3699/5758 [21:02<11:31,  2.98it/s]
 64%|██████▍   | 3700/5758 [21:02<11:11,  3.06it/s]
                                                   

 64%|██████▍   | 3700/5758 [21:02<11:11,  3.06it/s]
 64%|██████▍   | 3701/5758 [21:03<10:55,  3.14it/s]
                                                   

 64%|██████▍   | 3701/5758 [21:03<10:55,  3.14it/s]
 64%|██████▍   | 3702/5758 [21:03<11:40,  2.94it/s]
                                                   

 64%|██████▍   | 3702/5758 [21:03<11:40,  2.94it/s]
 64%|██████▍   | 3703/5758 [21:03<11:56,  2.87it/s]
                                                   

 64%|██████▍   | 3703/5758 [21:03<11:56,  2.87it/s]
 64%|██████▍   | 3704/5758 [21:04<11:24,  3.00it/s]
                                                   

 64%|██████▍   | 3704/5758 [21:04<11:24,  3.00it/s]
 64%|██████▍   | 3705/5758 [21:04<10:55,  3.13it/s]
                                                   

 64%|██████▍   | 3705/5758 [21:04<10:55,  3.13it/s]
 64%|██████▍   | 3706/5758 [21:04<11:18,  3.03it/s]
                                                   

 64%|██████▍   | 3706/5758 [21:04<11:18,  3.03it/s]
 64%|██████▍   | 3707/5758 [21:05<11:38,  2.94it/s]
                                                   

 64%|██████▍   | 3707/5758 [21:05<11:38,  2.94it/s]
 64%|██████▍   | 3708/5758 [21:05<11:17,  3.02it/s]
                                                   

 64%|██████▍   | 3708/5758 [21:05<11:17,  3.02it/s]
 64%|██████▍   | 3709/5758 [21:05<11:35,  2.95it/s]
                                                   

 64%|██████▍   | 3709/5758 [21:05<11:35,  2.95it/s]
 64%|██████▍   | 3710/5758 [21:06<11:11,  3.05it/s]
                                                   

 64%|██████▍   | 3710/5758 [21:06<11:11,  3.05it/s]
 64%|██████▍   | 3711/5758 [21:06<11:29,  2.97it/s]
                                                   

 64%|██████▍   | 3711/5758 [21:06<11:29,  2.97it/s]
 64%|██████▍   | 3712/5758 [21:06<11:06,  3.07it/s]
                                                   

 64%|██████▍   | 3712/5758 [21:06<11:06,  3.07it/s]
 64%|██████▍   | 3713/5758 [21:07<12:16,  2.78it/s]
                                                   

 64%|██████▍   | 3713/5758 [21:07<12:16,  2.78it/s]
 65%|██████▍   | 3714/5758 [21:07<11:38,  2.93it/s]
                                                   

 65%|██████▍   | 3714/5758 [21:07<11:38,  2.93it/s]
 65%|██████▍   | 3715/5758 [21:08<11:54,  2.86it/s]
                                                   

 65%|██████▍   | 3715/5758 [21:08<11:54,  2.86it/s]
 65%|██████▍   | 3716/5758 [21:08<11:18,  3.01it/s]
                                                   

 65%|██████▍   | 3716/5758 [21:08<11:18,  3.01it/s]
 65%|██████▍   | 3717/5758 [21:08<11:23,  2.99it/s]
                                                   

 65%|██████▍   | 3717/5758 [21:08<11:23,  2.99it/s]
 65%|██████▍   | 3718/5758 [21:09<11:42,  2.90it/s]
                                                   

 65%|██████▍   | 3718/5758 [21:09<11:42,  2.90it/s]
 65%|██████▍   | 3719/5758 [21:09<11:09,  3.05it/s]
                                                   

 65%|██████▍   | 3719/5758 [21:09<11:09,  3.05it/s]
 65%|██████▍   | 3720/5758 [21:09<10:50,  3.13it/s]
                                                   

 65%|██████▍   | 3720/5758 [21:09<10:50,  3.13it/s]
 65%|██████▍   | 3721/5758 [21:09<11:18,  3.00it/s]
                                                   
{'loss': '0.6322', 'grad_norm': '17.38', 'learning_rate': '3.635e-06', 'num_tokens': '2.44e+06', 'mean_token_accuracy': '0.8262', 'epoch': '0.6367'}
+{'loss': '1.049', 'grad_norm': '11.12', 'learning_rate': '3.633e-06', 'num_tokens': '2.441e+06', 'mean_token_accuracy': '0.7033', 'epoch': '0.6369'}
+{'loss': '2.063', 'grad_norm': '22.88', 'learning_rate': '3.631e-06', 'num_tokens': '2.441e+06', 'mean_token_accuracy': '0.5211', 'epoch': '0.637'}
+{'loss': '1.518', 'grad_norm': '18.12', 'learning_rate': '3.63e-06', 'num_tokens': '2.442e+06', 'mean_token_accuracy': '0.6675', 'epoch': '0.6372'}
+{'loss': '1.632', 'grad_norm': '13.25', 'learning_rate': '3.628e-06', 'num_tokens': '2.442e+06', 'mean_token_accuracy': '0.652', 'epoch': '0.6374'}
+{'loss': '0.8503', 'grad_norm': '17.12', 'learning_rate': '3.626e-06', 'num_tokens': '2.443e+06', 'mean_token_accuracy': '0.8117', 'epoch': '0.6375'}
+{'loss': '0.5674', 'grad_norm': '7.562', 'learning_rate': '3.625e-06', 'num_tokens': '2.444e+06', 'mean_token_accuracy': '0.8346', 'epoch': '0.6377'}
+{'loss': '1.705', 'grad_norm': '40.25', 'learning_rate': '3.623e-06', 'num_tokens': '2.444e+06', 'mean_token_accuracy': '0.6582', 'epoch': '0.6379'}
+{'loss': '1.889', 'grad_norm': '17.12', 'learning_rate': '3.621e-06', 'num_tokens': '2.445e+06', 'mean_token_accuracy': '0.6099', 'epoch': '0.6381'}
+{'loss': '0.936', 'grad_norm': '12.69', 'learning_rate': '3.619e-06', 'num_tokens': '2.446e+06', 'mean_token_accuracy': '0.6968', 'epoch': '0.6382'}
+{'loss': '1.594', 'grad_norm': '17', 'learning_rate': '3.618e-06', 'num_tokens': '2.446e+06', 'mean_token_accuracy': '0.6967', 'epoch': '0.6384'}
+{'loss': '1.64', 'grad_norm': '15.56', 'learning_rate': '3.616e-06', 'num_tokens': '2.447e+06', 'mean_token_accuracy': '0.6307', 'epoch': '0.6386'}
+{'loss': '1.78', 'grad_norm': '16.25', 'learning_rate': '3.614e-06', 'num_tokens': '2.447e+06', 'mean_token_accuracy': '0.653', 'epoch': '0.6388'}
+{'loss': '1.938', 'grad_norm': '14.12', 'learning_rate': '3.612e-06', 'num_tokens': '2.447e+06', 'mean_token_accuracy': '0.5746', 'epoch': '0.6389'}
+{'loss': '1.899', 'grad_norm': '13.81', 'learning_rate': '3.611e-06', 'num_tokens': '2.448e+06', 'mean_token_accuracy': '0.6035', 'epoch': '0.6391'}
+{'loss': '1.239', 'grad_norm': '16.5', 'learning_rate': '3.609e-06', 'num_tokens': '2.448e+06', 'mean_token_accuracy': '0.7218', 'epoch': '0.6393'}
+{'loss': '1.069', 'grad_norm': '14', 'learning_rate': '3.607e-06', 'num_tokens': '2.449e+06', 'mean_token_accuracy': '0.709', 'epoch': '0.6395'}
+{'loss': '1.011', 'grad_norm': '25.5', 'learning_rate': '3.605e-06', 'num_tokens': '2.449e+06', 'mean_token_accuracy': '0.779', 'epoch': '0.6396'}
+{'loss': '1.876', 'grad_norm': '28.75', 'learning_rate': '3.604e-06', 'num_tokens': '2.449e+06', 'mean_token_accuracy': '0.5843', 'epoch': '0.6398'}
+{'loss': '1.681', 'grad_norm': '15.5', 'learning_rate': '3.602e-06', 'num_tokens': '2.45e+06', 'mean_token_accuracy': '0.6543', 'epoch': '0.64'}
+{'loss': '0.9527', 'grad_norm': '5.844', 'learning_rate': '3.6e-06', 'num_tokens': '2.452e+06', 'mean_token_accuracy': '0.7649', 'epoch': '0.6402'}
+{'loss': '0.6101', 'grad_norm': '7.688', 'learning_rate': '3.598e-06', 'num_tokens': '2.453e+06', 'mean_token_accuracy': '0.8149', 'epoch': '0.6403'}
+{'loss': '2.017', 'grad_norm': '29.5', 'learning_rate': '3.597e-06', 'num_tokens': '2.454e+06', 'mean_token_accuracy': '0.6054', 'epoch': '0.6405'}
+{'loss': '1.123', 'grad_norm': '12.19', 'learning_rate': '3.595e-06', 'num_tokens': '2.455e+06', 'mean_token_accuracy': '0.6892', 'epoch': '0.6407'}
+{'loss': '1.866', 'grad_norm': '28.25', 'learning_rate': '3.593e-06', 'num_tokens': '2.455e+06', 'mean_token_accuracy': '0.6302', 'epoch': '0.6408'}
+{'loss': '2.035', 'grad_norm': '16.75', 'learning_rate': '3.592e-06', 'num_tokens': '2.455e+06', 'mean_token_accuracy': '0.575', 'epoch': '0.641'}
+{'loss': '1.339', 'grad_norm': '16.62', 'learning_rate': '3.59e-06', 'num_tokens': '2.456e+06', 'mean_token_accuracy': '0.6724', 'epoch': '0.6412'}
+{'loss': '1.47', 'grad_norm': '21.62', 'learning_rate': '3.588e-06', 'num_tokens': '2.456e+06', 'mean_token_accuracy': '0.7092', 'epoch': '0.6414'}
+{'loss': '1.926', 'grad_norm': '16.75', 'learning_rate': '3.586e-06', 'num_tokens': '2.456e+06', 'mean_token_accuracy': '0.622', 'epoch': '0.6415'}
+{'loss': '1.154', 'grad_norm': '12.5', 'learning_rate': '3.585e-06', 'num_tokens': '2.457e+06', 'mean_token_accuracy': '0.6929', 'epoch': '0.6417'}
+{'loss': '1.278', 'grad_norm': '12.94', 'learning_rate': '3.583e-06', 'num_tokens': '2.458e+06', 'mean_token_accuracy': '0.6396', 'epoch': '0.6419'}
+{'loss': '2.037', 'grad_norm': '10.75', 'learning_rate': '3.581e-06', 'num_tokens': '2.459e+06', 'mean_token_accuracy': '0.6086', 'epoch': '0.6421'}
+{'loss': '1.294', 'grad_norm': '11.88', 'learning_rate': '3.579e-06', 'num_tokens': '2.459e+06', 'mean_token_accuracy': '0.6325', 'epoch': '0.6422'}
+{'loss': '1.556', 'grad_norm': '8.438', 'learning_rate': '3.578e-06', 'num_tokens': '2.461e+06', 'mean_token_accuracy': '0.6752', 'epoch': '0.6424'}
+{'loss': '1.118', 'grad_norm': '36', 'learning_rate': '3.576e-06', 'num_tokens': '2.461e+06', 'mean_token_accuracy': '0.7273', 'epoch': '0.6426'}
+{'loss': '2.087', 'grad_norm': '16.62', 'learning_rate': '3.574e-06', 'num_tokens': '2.462e+06', 'mean_token_accuracy': '0.6256', 'epoch': '0.6428'}
+{'loss': '1.616', 'grad_norm': '14.25', 'learning_rate': '3.572e-06', 'num_tokens': '2.463e+06', 'mean_token_accuracy': '0.6661', 'epoch': '0.6429'}
+{'loss': '1.445', 'grad_norm': '10.88', 'learning_rate': '3.571e-06', 'num_tokens': '2.464e+06', 'mean_token_accuracy': '0.6181', 'epoch': '0.6431'}
+{'loss': '1.712', 'grad_norm': '15.69', 'learning_rate': '3.569e-06', 'num_tokens': '2.464e+06', 'mean_token_accuracy': '0.6271', 'epoch': '0.6433'}
+{'loss': '1.964', 'grad_norm': '8.5', 'learning_rate': '3.567e-06', 'num_tokens': '2.466e+06', 'mean_token_accuracy': '0.5702', 'epoch': '0.6435'}
+{'loss': '1.736', 'grad_norm': '23.38', 'learning_rate': '3.565e-06', 'num_tokens': '2.466e+06', 'mean_token_accuracy': '0.6316', 'epoch': '0.6436'}
+{'loss': '1.866', 'grad_norm': '14', 'learning_rate': '3.564e-06', 'num_tokens': '2.466e+06', 'mean_token_accuracy': '0.585', 'epoch': '0.6438'}
+{'loss': '1.486', 'grad_norm': '12.5', 'learning_rate': '3.562e-06', 'num_tokens': '2.467e+06', 'mean_token_accuracy': '0.6721', 'epoch': '0.644'}
+{'loss': '1.663', 'grad_norm': '19.25', 'learning_rate': '3.56e-06', 'num_tokens': '2.467e+06', 'mean_token_accuracy': '0.6385', 'epoch': '0.6441'}
+{'loss': '0.7874', 'grad_norm': '9.938', 'learning_rate': '3.559e-06', 'num_tokens': '2.468e+06', 'mean_token_accuracy': '0.8117', 'epoch': '0.6443'}
+{'loss': '1.065', 'grad_norm': '10.44', 'learning_rate': '3.557e-06', 'num_tokens': '2.469e+06', 'mean_token_accuracy': '0.6925', 'epoch': '0.6445'}
+{'loss': '1.086', 'grad_norm': '12.31', 'learning_rate': '3.555e-06', 'num_tokens': '2.47e+06', 'mean_token_accuracy': '0.7074', 'epoch': '0.6447'}
+{'loss': '1.142', 'grad_norm': '11.5', 'learning_rate': '3.553e-06', 'num_tokens': '2.471e+06', 'mean_token_accuracy': '0.6872', 'epoch': '0.6448'}
+{'loss': '1.053', 'grad_norm': '11.81', 'learning_rate': '3.552e-06', 'num_tokens': '2.471e+06', 'mean_token_accuracy': '0.7306', 'epoch': '0.645'}
+{'loss': '1.413', 'grad_norm': '10.94', 'learning_rate': '3.55e-06', 'num_tokens': '2.472e+06', 'mean_token_accuracy': '0.6685', 'epoch': '0.6452'}
+{'loss': '0.7408', 'grad_norm': '25.25', 'learning_rate': '3.548e-06', 'num_tokens': '2.473e+06', 'mean_token_accuracy': '0.8375', 'epoch': '0.6454'}
+{'loss': '1.04', 'grad_norm': '14.56', 'learning_rate': '3.546e-06', 'num_tokens': '2.473e+06', 'mean_token_accuracy': '0.7626', 'epoch': '0.6455'}
+{'loss': '1.195', 'grad_norm': '9.875', 'learning_rate': '3.545e-06', 'num_tokens': '2.474e+06', 'mean_token_accuracy': '0.7059', 'epoch': '0.6457'}
+{'loss': '2.158', 'grad_norm': '21.75', 'learning_rate': '3.543e-06', 'num_tokens': '2.474e+06', 'mean_token_accuracy': '0.5503', 'epoch': '0.6459'}
+{'loss': '1.048', 'grad_norm': '13.56', 'learning_rate': '3.541e-06', 'num_tokens': '2.475e+06', 'mean_token_accuracy': '0.6849', 'epoch': '0.6461'}
+
 65%|██████▍   | 3721/5758 [21:09<11:18,  3.00it/s]
 65%|██████▍   | 3722/5758 [21:10<11:00,  3.08it/s]
                                                   

 65%|██████▍   | 3722/5758 [21:10<11:00,  3.08it/s]
 65%|██████▍   | 3723/5758 [21:10<11:01,  3.08it/s]
                                                   

 65%|██████▍   | 3723/5758 [21:10<11:01,  3.08it/s]
 65%|██████▍   | 3724/5758 [21:10<11:35,  2.92it/s]
                                                   

 65%|██████▍   | 3724/5758 [21:10<11:35,  2.92it/s]
 65%|██████▍   | 3725/5758 [21:11<11:20,  2.99it/s]
                                                   

 65%|██████▍   | 3725/5758 [21:11<11:20,  2.99it/s]
 65%|██████▍   | 3726/5758 [21:11<10:52,  3.11it/s]
                                                   

 65%|██████▍   | 3726/5758 [21:11<10:52,  3.11it/s]
 65%|██████▍   | 3727/5758 [21:11<11:24,  2.97it/s]
                                                   

 65%|██████▍   | 3727/5758 [21:11<11:24,  2.97it/s]
 65%|██████▍   | 3728/5758 [21:12<11:12,  3.02it/s]
                                                   

 65%|██████▍   | 3728/5758 [21:12<11:12,  3.02it/s]
 65%|██████▍   | 3729/5758 [21:12<11:10,  3.03it/s]
                                                   

 65%|██████▍   | 3729/5758 [21:12<11:10,  3.03it/s]
 65%|██████▍   | 3730/5758 [21:12<10:56,  3.09it/s]
                                                   

 65%|██████▍   | 3730/5758 [21:12<10:56,  3.09it/s]
 65%|██████▍   | 3731/5758 [21:13<11:19,  2.99it/s]
                                                   

 65%|██████▍   | 3731/5758 [21:13<11:19,  2.99it/s]
 65%|██████▍   | 3732/5758 [21:13<10:51,  3.11it/s]
                                                   

 65%|██████▍   | 3732/5758 [21:13<10:51,  3.11it/s]
 65%|██████▍   | 3733/5758 [21:13<11:12,  3.01it/s]
                                                   

 65%|██████▍   | 3733/5758 [21:13<11:12,  3.01it/s]
 65%|██████▍   | 3734/5758 [21:14<10:49,  3.11it/s]
                                                   

 65%|██████▍   | 3734/5758 [21:14<10:49,  3.11it/s]
 65%|██████▍   | 3735/5758 [21:14<11:20,  2.97it/s]
                                                   

 65%|██████▍   | 3735/5758 [21:14<11:20,  2.97it/s]
 65%|██████▍   | 3736/5758 [21:14<11:08,  3.02it/s]
                                                   

 65%|██████▍   | 3736/5758 [21:14<11:08,  3.02it/s]
 65%|██████▍   | 3737/5758 [21:15<11:36,  2.90it/s]
                                                   

 65%|██████▍   | 3737/5758 [21:15<11:36,  2.90it/s]
 65%|██████▍   | 3738/5758 [21:15<11:08,  3.02it/s]
                                                   

 65%|██████▍   | 3738/5758 [21:15<11:08,  3.02it/s]
 65%|██████▍   | 3739/5758 [21:15<11:00,  3.06it/s]
                                                   

 65%|██████▍   | 3739/5758 [21:15<11:00,  3.06it/s]
 65%|██████▍   | 3740/5758 [21:16<11:18,  2.98it/s]
                                                   

 65%|██████▍   | 3740/5758 [21:16<11:18,  2.98it/s]
 65%|██████▍   | 3741/5758 [21:16<12:11,  2.76it/s]
                                                   

 65%|██████▍   | 3741/5758 [21:16<12:11,  2.76it/s]
 65%|██████▍   | 3742/5758 [21:16<11:27,  2.93it/s]
                                                   

 65%|██████▍   | 3742/5758 [21:16<11:27,  2.93it/s]
 65%|██████▌   | 3743/5758 [21:17<11:04,  3.03it/s]
                                                   

 65%|██████▌   | 3743/5758 [21:17<11:04,  3.03it/s]
 65%|██████▌   | 3744/5758 [21:17<10:47,  3.11it/s]
                                                   

 65%|██████▌   | 3744/5758 [21:17<10:47,  3.11it/s]
 65%|██████▌   | 3745/5758 [21:17<10:45,  3.12it/s]
                                                   

 65%|██████▌   | 3745/5758 [21:17<10:45,  3.12it/s]
 65%|██████▌   | 3746/5758 [21:18<11:02,  3.04it/s]
                                                   

 65%|██████▌   | 3746/5758 [21:18<11:02,  3.04it/s]
 65%|██████▌   | 3747/5758 [21:18<11:10,  3.00it/s]
                                                   

 65%|██████▌   | 3747/5758 [21:18<11:10,  3.00it/s]
 65%|██████▌   | 3748/5758 [21:18<10:49,  3.10it/s]
                                                   

 65%|██████▌   | 3748/5758 [21:18<10:49,  3.10it/s]
 65%|██████▌   | 3749/5758 [21:19<11:10,  2.99it/s]
                                                   

 65%|██████▌   | 3749/5758 [21:19<11:10,  2.99it/s]
 65%|██████▌   | 3750/5758 [21:19<11:33,  2.90it/s]
                                                   

 65%|██████▌   | 3750/5758 [21:19<11:33,  2.90it/s]
 65%|██████▌   | 3751/5758 [21:19<11:14,  2.98it/s]
                                                   

 65%|██████▌   | 3751/5758 [21:19<11:14,  2.98it/s]
 65%|██████▌   | 3752/5758 [21:20<11:17,  2.96it/s]
                                                   

 65%|██████▌   | 3752/5758 [21:20<11:17,  2.96it/s]
 65%|██████▌   | 3753/5758 [21:20<11:13,  2.98it/s]
                                                   

 65%|██████▌   | 3753/5758 [21:20<11:13,  2.98it/s]
 65%|██████▌   | 3754/5758 [21:20<10:58,  3.05it/s]
                                                   

 65%|██████▌   | 3754/5758 [21:20<10:58,  3.05it/s]
 65%|██████▌   | 3755/5758 [21:21<10:38,  3.14it/s]
                                                   

 65%|██████▌   | 3755/5758 [21:21<10:38,  3.14it/s]
 65%|██████▌   | 3756/5758 [21:21<10:26,  3.20it/s]
                                                   

 65%|██████▌   | 3756/5758 [21:21<10:26,  3.20it/s]
 65%|██████▌   | 3757/5758 [21:21<10:14,  3.25it/s]
                                                   

 65%|██████▌   | 3757/5758 [21:21<10:14,  3.25it/s]
 65%|██████▌   | 3758/5758 [21:22<10:57,  3.04it/s]
                                                   

 65%|██████▌   | 3758/5758 [21:22<10:57,  3.04it/s]
 65%|██████▌   | 3759/5758 [21:22<10:39,  3.13it/s]
                                                   

 65%|██████▌   | 3759/5758 [21:22<10:39,  3.13it/s]
 65%|██████▌   | 3760/5758 [21:22<10:55,  3.05it/s]
                                                   

 65%|██████▌   | 3760/5758 [21:22<10:55,  3.05it/s]
 65%|██████▌   | 3761/5758 [21:23<10:34,  3.15it/s]
                                                   

 65%|██████▌   | 3761/5758 [21:23<10:34,  3.15it/s]
 65%|██████▌   | 3762/5758 [21:23<10:24,  3.20it/s]
                                                   

 65%|██████▌   | 3762/5758 [21:23<10:24,  3.20it/s]
 65%|██████▌   | 3763/5758 [21:23<10:59,  3.02it/s]
                                                   

 65%|██████▌   | 3763/5758 [21:23<10:59,  3.02it/s]
 65%|██████▌   | 3764/5758 [21:24<11:22,  2.92it/s]
                                                   

 65%|██████▌   | 3764/5758 [21:24<11:22,  2.92it/s]
 65%|██████▌   | 3765/5758 [21:24<11:43,  2.83it/s]
                                                   

 65%|██████▌   | 3765/5758 [21:24<11:43,  2.83it/s]
 65%|██████▌   | 3766/5758 [21:25<12:50,  2.58it/s]
                                                   

 65%|██████▌   | 3766/5758 [21:25<12:50,  2.58it/s]
 65%|██████▌   | 3767/5758 [21:25<12:53,  2.57it/s]
                                                   

 65%|██████▌   | 3767/5758 [21:25<12:53,  2.57it/s]
 65%|██████▌   | 3768/5758 [21:25<12:22,  2.68it/s]
                                                   

 65%|██████▌   | 3768/5758 [21:25<12:22,  2.68it/s]
 65%|██████▌   | 3769/5758 [21:26<11:40,  2.84it/s]
                                                   

 65%|██████▌   | 3769/5758 [21:26<11:40,  2.84it/s]
 65%|██████▌   | 3770/5758 [21:26<11:31,  2.87it/s]
                                                   

 65%|██████▌   | 3770/5758 [21:26<11:31,  2.87it/s]
 65%|██████▌   | 3771/5758 [21:26<11:32,  2.87it/s]
                                                   

 65%|██████▌   | 3771/5758 [21:26<11:32,  2.87it/s]
 66%|██████▌   | 3772/5758 [21:27<11:22,  2.91it/s]
                                                   

 66%|██████▌   | 3772/5758 [21:27<11:22,  2.91it/s]
 66%|██████▌   | 3773/5758 [21:27<11:21,  2.91it/s]
                                                   

 66%|██████▌   | 3773/5758 [21:27<11:21,  2.91it/s]
 66%|██████▌   | 3774/5758 [21:27<10:53,  3.04it/s]
                                                   

 66%|██████▌   | 3774/5758 [21:27<10:53,  3.04it/s]
 66%|██████▌   | 3775/5758 [21:28<10:53,  3.03it/s]
                                                   

 66%|██████▌   | 3775/5758 [21:28<10:53,  3.03it/s]
 66%|██████▌   | 3776/5758 [21:28<10:48,  3.05it/s]
                                                   
{'loss': '1.107', 'grad_norm': '18.62', 'learning_rate': '3.539e-06', 'num_tokens': '2.475e+06', 'mean_token_accuracy': '0.7595', 'epoch': '0.6462'}
+{'loss': '2.226', 'grad_norm': '28.25', 'learning_rate': '3.538e-06', 'num_tokens': '2.475e+06', 'mean_token_accuracy': '0.56', 'epoch': '0.6464'}
+{'loss': '1.466', 'grad_norm': '16.88', 'learning_rate': '3.536e-06', 'num_tokens': '2.476e+06', 'mean_token_accuracy': '0.6731', 'epoch': '0.6466'}
+{'loss': '0.9706', 'grad_norm': '24.75', 'learning_rate': '3.534e-06', 'num_tokens': '2.476e+06', 'mean_token_accuracy': '0.7812', 'epoch': '0.6468'}
+{'loss': '1.144', 'grad_norm': '12.75', 'learning_rate': '3.532e-06', 'num_tokens': '2.477e+06', 'mean_token_accuracy': '0.7194', 'epoch': '0.6469'}
+{'loss': '1.291', 'grad_norm': '8.938', 'learning_rate': '3.531e-06', 'num_tokens': '2.478e+06', 'mean_token_accuracy': '0.6948', 'epoch': '0.6471'}
+{'loss': '0.5761', 'grad_norm': '27.88', 'learning_rate': '3.529e-06', 'num_tokens': '2.478e+06', 'mean_token_accuracy': '0.8984', 'epoch': '0.6473'}
+{'loss': '1.28', 'grad_norm': '12.38', 'learning_rate': '3.527e-06', 'num_tokens': '2.479e+06', 'mean_token_accuracy': '0.6533', 'epoch': '0.6474'}
+{'loss': '1.973', 'grad_norm': '11.25', 'learning_rate': '3.526e-06', 'num_tokens': '2.48e+06', 'mean_token_accuracy': '0.5978', 'epoch': '0.6476'}
+{'loss': '0.569', 'grad_norm': '6.938', 'learning_rate': '3.524e-06', 'num_tokens': '2.481e+06', 'mean_token_accuracy': '0.8295', 'epoch': '0.6478'}
+{'loss': '1.53', 'grad_norm': '11.38', 'learning_rate': '3.522e-06', 'num_tokens': '2.482e+06', 'mean_token_accuracy': '0.6446', 'epoch': '0.648'}
+{'loss': '1.1', 'grad_norm': '13.75', 'learning_rate': '3.52e-06', 'num_tokens': '2.483e+06', 'mean_token_accuracy': '0.696', 'epoch': '0.6481'}
+{'loss': '2.337', 'grad_norm': '17', 'learning_rate': '3.519e-06', 'num_tokens': '2.483e+06', 'mean_token_accuracy': '0.5615', 'epoch': '0.6483'}
+{'loss': '1.121', 'grad_norm': '13.19', 'learning_rate': '3.517e-06', 'num_tokens': '2.484e+06', 'mean_token_accuracy': '0.677', 'epoch': '0.6485'}
+{'loss': '1.153', 'grad_norm': '9.938', 'learning_rate': '3.515e-06', 'num_tokens': '2.485e+06', 'mean_token_accuracy': '0.696', 'epoch': '0.6487'}
+{'loss': '1.271', 'grad_norm': '12.56', 'learning_rate': '3.513e-06', 'num_tokens': '2.486e+06', 'mean_token_accuracy': '0.6625', 'epoch': '0.6488'}
+{'loss': '1.49', 'grad_norm': '19.62', 'learning_rate': '3.512e-06', 'num_tokens': '2.486e+06', 'mean_token_accuracy': '0.6853', 'epoch': '0.649'}
+{'loss': '0.6092', 'grad_norm': '8.375', 'learning_rate': '3.51e-06', 'num_tokens': '2.488e+06', 'mean_token_accuracy': '0.8115', 'epoch': '0.6492'}
+{'loss': '1.787', 'grad_norm': '19', 'learning_rate': '3.508e-06', 'num_tokens': '2.488e+06', 'mean_token_accuracy': '0.6292', 'epoch': '0.6494'}
+{'loss': '1.282', 'grad_norm': '16.25', 'learning_rate': '3.506e-06', 'num_tokens': '2.488e+06', 'mean_token_accuracy': '0.7375', 'epoch': '0.6495'}
+{'loss': '1.264', 'grad_norm': '9.312', 'learning_rate': '3.505e-06', 'num_tokens': '2.49e+06', 'mean_token_accuracy': '0.6816', 'epoch': '0.6497'}
+{'loss': '2.047', 'grad_norm': '15.75', 'learning_rate': '3.503e-06', 'num_tokens': '2.49e+06', 'mean_token_accuracy': '0.567', 'epoch': '0.6499'}
+{'loss': '1.333', 'grad_norm': '13.06', 'learning_rate': '3.501e-06', 'num_tokens': '2.491e+06', 'mean_token_accuracy': '0.6736', 'epoch': '0.6501'}
+{'loss': '0.6366', 'grad_norm': '8.375', 'learning_rate': '3.499e-06', 'num_tokens': '2.492e+06', 'mean_token_accuracy': '0.8068', 'epoch': '0.6502'}
+{'loss': '1.207', 'grad_norm': '11.31', 'learning_rate': '3.498e-06', 'num_tokens': '2.493e+06', 'mean_token_accuracy': '0.6658', 'epoch': '0.6504'}
+{'loss': '1.794', 'grad_norm': '40', 'learning_rate': '3.496e-06', 'num_tokens': '2.493e+06', 'mean_token_accuracy': '0.6449', 'epoch': '0.6506'}
+{'loss': '1.04', 'grad_norm': '7.062', 'learning_rate': '3.494e-06', 'num_tokens': '2.495e+06', 'mean_token_accuracy': '0.743', 'epoch': '0.6507'}
+{'loss': '1.213', 'grad_norm': '12.75', 'learning_rate': '3.493e-06', 'num_tokens': '2.496e+06', 'mean_token_accuracy': '0.7361', 'epoch': '0.6509'}
+{'loss': '0.6277', 'grad_norm': '6.969', 'learning_rate': '3.491e-06', 'num_tokens': '2.498e+06', 'mean_token_accuracy': '0.8012', 'epoch': '0.6511'}
+{'loss': '1.925', 'grad_norm': '9.188', 'learning_rate': '3.489e-06', 'num_tokens': '2.499e+06', 'mean_token_accuracy': '0.6137', 'epoch': '0.6513'}
+{'loss': '0.8275', 'grad_norm': '11.25', 'learning_rate': '3.487e-06', 'num_tokens': '2.5e+06', 'mean_token_accuracy': '0.761', 'epoch': '0.6514'}
+{'loss': '2.473', 'grad_norm': '23.62', 'learning_rate': '3.486e-06', 'num_tokens': '2.5e+06', 'mean_token_accuracy': '0.5251', 'epoch': '0.6516'}
+{'loss': '1.399', 'grad_norm': '21.12', 'learning_rate': '3.484e-06', 'num_tokens': '2.5e+06', 'mean_token_accuracy': '0.6864', 'epoch': '0.6518'}
+{'loss': '1.267', 'grad_norm': '12.88', 'learning_rate': '3.482e-06', 'num_tokens': '2.501e+06', 'mean_token_accuracy': '0.6538', 'epoch': '0.652'}
+{'loss': '1.248', 'grad_norm': '13.25', 'learning_rate': '3.48e-06', 'num_tokens': '2.502e+06', 'mean_token_accuracy': '0.6531', 'epoch': '0.6521'}
+{'loss': '1.131', 'grad_norm': '13.56', 'learning_rate': '3.479e-06', 'num_tokens': '2.502e+06', 'mean_token_accuracy': '0.6749', 'epoch': '0.6523'}
+{'loss': '1.155', 'grad_norm': '13.06', 'learning_rate': '3.477e-06', 'num_tokens': '2.503e+06', 'mean_token_accuracy': '0.6687', 'epoch': '0.6525'}
+{'loss': '1', 'grad_norm': '14', 'learning_rate': '3.475e-06', 'num_tokens': '2.504e+06', 'mean_token_accuracy': '0.6997', 'epoch': '0.6527'}
+{'loss': '2.167', 'grad_norm': '16.75', 'learning_rate': '3.473e-06', 'num_tokens': '2.504e+06', 'mean_token_accuracy': '0.571', 'epoch': '0.6528'}
+{'loss': '1.317', 'grad_norm': '26.62', 'learning_rate': '3.472e-06', 'num_tokens': '2.504e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.653'}
+{'loss': '0.5655', 'grad_norm': '8.75', 'learning_rate': '3.47e-06', 'num_tokens': '2.506e+06', 'mean_token_accuracy': '0.8363', 'epoch': '0.6532'}
+{'loss': '0.7495', 'grad_norm': '35', 'learning_rate': '3.468e-06', 'num_tokens': '2.506e+06', 'mean_token_accuracy': '0.8734', 'epoch': '0.6534'}
+{'loss': '1.664', 'grad_norm': '20.5', 'learning_rate': '3.466e-06', 'num_tokens': '2.506e+06', 'mean_token_accuracy': '0.6655', 'epoch': '0.6535'}
+{'loss': '0.7104', 'grad_norm': '9.438', 'learning_rate': '3.465e-06', 'num_tokens': '2.507e+06', 'mean_token_accuracy': '0.7806', 'epoch': '0.6537'}
+{'loss': '1.772', 'grad_norm': '20.38', 'learning_rate': '3.463e-06', 'num_tokens': '2.507e+06', 'mean_token_accuracy': '0.6434', 'epoch': '0.6539'}
+{'loss': '1.513', 'grad_norm': '17.25', 'learning_rate': '3.461e-06', 'num_tokens': '2.508e+06', 'mean_token_accuracy': '0.5984', 'epoch': '0.654'}
+{'loss': '1.328', 'grad_norm': '36.75', 'learning_rate': '3.46e-06', 'num_tokens': '2.508e+06', 'mean_token_accuracy': '0.6884', 'epoch': '0.6542'}
+{'loss': '1.058', 'grad_norm': '32.5', 'learning_rate': '3.458e-06', 'num_tokens': '2.508e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.6544'}
+{'loss': '2.328', 'grad_norm': '19.12', 'learning_rate': '3.456e-06', 'num_tokens': '2.509e+06', 'mean_token_accuracy': '0.5382', 'epoch': '0.6546'}
+{'loss': '1.209', 'grad_norm': '22.25', 'learning_rate': '3.454e-06', 'num_tokens': '2.509e+06', 'mean_token_accuracy': '0.7661', 'epoch': '0.6547'}
+{'loss': '1.055', 'grad_norm': '10.44', 'learning_rate': '3.453e-06', 'num_tokens': '2.51e+06', 'mean_token_accuracy': '0.7028', 'epoch': '0.6549'}
+{'loss': '1.226', 'grad_norm': '11.38', 'learning_rate': '3.451e-06', 'num_tokens': '2.511e+06', 'mean_token_accuracy': '0.6584', 'epoch': '0.6551'}
+{'loss': '1.632', 'grad_norm': '6.562', 'learning_rate': '3.449e-06', 'num_tokens': '2.512e+06', 'mean_token_accuracy': '0.657', 'epoch': '0.6553'}
+{'loss': '1.08', 'grad_norm': '13.69', 'learning_rate': '3.447e-06', 'num_tokens': '2.513e+06', 'mean_token_accuracy': '0.6833', 'epoch': '0.6554'}
+{'loss': '1.276', 'grad_norm': '8.062', 'learning_rate': '3.446e-06', 'num_tokens': '2.515e+06', 'mean_token_accuracy': '0.6794', 'epoch': '0.6556'}
+
 66%|██████▌   | 3776/5758 [21:28<10:48,  3.05it/s]
 66%|██████▌   | 3777/5758 [21:28<10:30,  3.14it/s]
                                                   

 66%|██████▌   | 3777/5758 [21:28<10:30,  3.14it/s]
 66%|██████▌   | 3778/5758 [21:28<10:35,  3.11it/s]
                                                   

 66%|██████▌   | 3778/5758 [21:28<10:35,  3.11it/s]
 66%|██████▌   | 3779/5758 [21:29<10:45,  3.07it/s]
                                                   

 66%|██████▌   | 3779/5758 [21:29<10:45,  3.07it/s]
 66%|██████▌   | 3780/5758 [21:29<10:25,  3.16it/s]
                                                   

 66%|██████▌   | 3780/5758 [21:29<10:25,  3.16it/s]
 66%|██████▌   | 3781/5758 [21:29<10:50,  3.04it/s]
                                                   

 66%|██████▌   | 3781/5758 [21:29<10:50,  3.04it/s]
 66%|██████▌   | 3782/5758 [21:30<10:28,  3.14it/s]
                                                   

 66%|██████▌   | 3782/5758 [21:30<10:28,  3.14it/s]
 66%|██████▌   | 3783/5758 [21:30<10:16,  3.20it/s]
                                                   

 66%|██████▌   | 3783/5758 [21:30<10:16,  3.20it/s]
 66%|██████▌   | 3784/5758 [21:30<09:56,  3.31it/s]
                                                   

 66%|██████▌   | 3784/5758 [21:30<09:56,  3.31it/s]
 66%|██████▌   | 3785/5758 [21:31<09:57,  3.30it/s]
                                                   

 66%|██████▌   | 3785/5758 [21:31<09:57,  3.30it/s]
 66%|██████▌   | 3786/5758 [21:31<09:59,  3.29it/s]
                                                   

 66%|██████▌   | 3786/5758 [21:31<09:59,  3.29it/s]
 66%|██████▌   | 3787/5758 [21:31<09:57,  3.30it/s]
                                                   

 66%|██████▌   | 3787/5758 [21:31<09:57,  3.30it/s]
 66%|██████▌   | 3788/5758 [21:32<10:06,  3.25it/s]
                                                   

 66%|██████▌   | 3788/5758 [21:32<10:06,  3.25it/s]
 66%|██████▌   | 3789/5758 [21:32<10:28,  3.14it/s]
                                                   

 66%|██████▌   | 3789/5758 [21:32<10:28,  3.14it/s]
 66%|██████▌   | 3790/5758 [21:32<10:32,  3.11it/s]
                                                   

 66%|██████▌   | 3790/5758 [21:32<10:32,  3.11it/s]
 66%|██████▌   | 3791/5758 [21:33<10:37,  3.08it/s]
                                                   

 66%|██████▌   | 3791/5758 [21:33<10:37,  3.08it/s]
 66%|██████▌   | 3792/5758 [21:33<12:13,  2.68it/s]
                                                   

 66%|██████▌   | 3792/5758 [21:33<12:13,  2.68it/s]
 66%|██████▌   | 3793/5758 [21:33<12:33,  2.61it/s]
                                                   

 66%|██████▌   | 3793/5758 [21:33<12:33,  2.61it/s]
 66%|██████▌   | 3794/5758 [21:34<11:55,  2.74it/s]
                                                   

 66%|██████▌   | 3794/5758 [21:34<11:55,  2.74it/s]
 66%|██████▌   | 3795/5758 [21:34<11:44,  2.78it/s]
                                                   

 66%|██████▌   | 3795/5758 [21:34<11:44,  2.78it/s]
 66%|██████▌   | 3796/5758 [21:35<11:43,  2.79it/s]
                                                   

 66%|██████▌   | 3796/5758 [21:35<11:43,  2.79it/s]
 66%|██████▌   | 3797/5758 [21:35<11:29,  2.84it/s]
                                                   

 66%|██████▌   | 3797/5758 [21:35<11:29,  2.84it/s]
 66%|██████▌   | 3798/5758 [21:35<11:05,  2.94it/s]
                                                   

 66%|██████▌   | 3798/5758 [21:35<11:05,  2.94it/s]
 66%|██████▌   | 3799/5758 [21:35<10:45,  3.03it/s]
                                                   

 66%|██████▌   | 3799/5758 [21:35<10:45,  3.03it/s]
 66%|██████▌   | 3800/5758 [21:36<10:44,  3.04it/s]
                                                   

 66%|██████▌   | 3800/5758 [21:36<10:44,  3.04it/s]
 66%|██████▌   | 3801/5758 [21:36<11:04,  2.95it/s]
                                                   

 66%|██████▌   | 3801/5758 [21:36<11:04,  2.95it/s]
 66%|██████▌   | 3802/5758 [21:37<11:24,  2.86it/s]
                                                   

 66%|██████▌   | 3802/5758 [21:37<11:24,  2.86it/s]
 66%|██████▌   | 3803/5758 [21:37<11:45,  2.77it/s]
                                                   

 66%|██████▌   | 3803/5758 [21:37<11:45,  2.77it/s]
 66%|██████▌   | 3804/5758 [21:37<11:31,  2.83it/s]
                                                   

 66%|██████▌   | 3804/5758 [21:37<11:31,  2.83it/s]
 66%|██████▌   | 3805/5758 [21:38<10:51,  3.00it/s]
                                                   

 66%|██████▌   | 3805/5758 [21:38<10:51,  3.00it/s]
 66%|██████▌   | 3806/5758 [21:38<11:19,  2.87it/s]
                                                   

 66%|██████▌   | 3806/5758 [21:38<11:19,  2.87it/s]
 66%|██████▌   | 3807/5758 [21:38<11:12,  2.90it/s]
                                                   

 66%|██████▌   | 3807/5758 [21:38<11:12,  2.90it/s]
 66%|██████▌   | 3808/5758 [21:39<11:00,  2.95it/s]
                                                   

 66%|██████▌   | 3808/5758 [21:39<11:00,  2.95it/s]
 66%|██████▌   | 3809/5758 [21:39<10:54,  2.98it/s]
                                                   

 66%|██████▌   | 3809/5758 [21:39<10:54,  2.98it/s]
 66%|██████▌   | 3810/5758 [21:39<11:36,  2.80it/s]
                                                   

 66%|██████▌   | 3810/5758 [21:39<11:36,  2.80it/s]
 66%|██████▌   | 3811/5758 [21:40<11:49,  2.75it/s]
                                                   

 66%|██████▌   | 3811/5758 [21:40<11:49,  2.75it/s]
 66%|██████▌   | 3812/5758 [21:40<12:04,  2.68it/s]
                                                   

 66%|██████▌   | 3812/5758 [21:40<12:04,  2.68it/s]
 66%|██████▌   | 3813/5758 [21:40<12:08,  2.67it/s]
                                                   

 66%|██████▌   | 3813/5758 [21:40<12:08,  2.67it/s]
 66%|██████▌   | 3814/5758 [21:41<11:33,  2.80it/s]
                                                   

 66%|██████▌   | 3814/5758 [21:41<11:33,  2.80it/s]
 66%|██████▋   | 3815/5758 [21:41<11:14,  2.88it/s]
                                                   

 66%|██████▋   | 3815/5758 [21:41<11:14,  2.88it/s]
 66%|██████▋   | 3816/5758 [21:41<10:44,  3.01it/s]
                                                   

 66%|██████▋   | 3816/5758 [21:41<10:44,  3.01it/s]
 66%|██████▋   | 3817/5758 [21:42<11:11,  2.89it/s]
                                                   

 66%|██████▋   | 3817/5758 [21:42<11:11,  2.89it/s]
 66%|██████▋   | 3818/5758 [21:42<12:12,  2.65it/s]
                                                   

 66%|██████▋   | 3818/5758 [21:42<12:12,  2.65it/s]
 66%|██████▋   | 3819/5758 [21:43<11:43,  2.76it/s]
                                                   

 66%|██████▋   | 3819/5758 [21:43<11:43,  2.76it/s]
 66%|██████▋   | 3820/5758 [21:43<11:58,  2.70it/s]
                                                   

 66%|██████▋   | 3820/5758 [21:43<11:58,  2.70it/s]
 66%|██████▋   | 3821/5758 [21:43<11:55,  2.71it/s]
                                                   

 66%|██████▋   | 3821/5758 [21:43<11:55,  2.71it/s]
 66%|██████▋   | 3822/5758 [21:44<11:16,  2.86it/s]
                                                   

 66%|██████▋   | 3822/5758 [21:44<11:16,  2.86it/s]
 66%|██████▋   | 3823/5758 [21:44<11:06,  2.90it/s]
                                                   

 66%|██████▋   | 3823/5758 [21:44<11:06,  2.90it/s]
 66%|██████▋   | 3824/5758 [21:44<11:23,  2.83it/s]
                                                   

 66%|██████▋   | 3824/5758 [21:44<11:23,  2.83it/s]
 66%|██████▋   | 3825/5758 [21:45<11:08,  2.89it/s]
                                                   

 66%|██████▋   | 3825/5758 [21:45<11:08,  2.89it/s]
 66%|██████▋   | 3826/5758 [21:45<11:28,  2.80it/s]
                                                   

 66%|██████▋   | 3826/5758 [21:45<11:28,  2.80it/s]
 66%|██████▋   | 3827/5758 [21:45<11:34,  2.78it/s]
                                                   

 66%|██████▋   | 3827/5758 [21:45<11:34,  2.78it/s]
 66%|██████▋   | 3828/5758 [21:46<11:05,  2.90it/s]
                                                   

 66%|██████▋   | 3828/5758 [21:46<11:05,  2.90it/s]
 66%|██████▋   | 3829/5758 [21:46<10:54,  2.95it/s]
                                                   

 66%|██████▋   | 3829/5758 [21:46<10:54,  2.95it/s]
 67%|██████▋   | 3830/5758 [21:46<10:29,  3.06it/s]
                                                   

 67%|██████▋   | 3830/5758 [21:46<10:29,  3.06it/s]
 67%|██████▋   | 3831/5758 [21:47<10:24,  3.08it/s]
                                                   
{'loss': '1.034', 'grad_norm': '21.62', 'learning_rate': '3.444e-06', 'num_tokens': '2.515e+06', 'mean_token_accuracy': '0.7744', 'epoch': '0.6558'}
+{'loss': '0.518', 'grad_norm': '14.44', 'learning_rate': '3.442e-06', 'num_tokens': '2.515e+06', 'mean_token_accuracy': '0.8769', 'epoch': '0.656'}
+{'loss': '2.13', 'grad_norm': '12.75', 'learning_rate': '3.44e-06', 'num_tokens': '2.516e+06', 'mean_token_accuracy': '0.5622', 'epoch': '0.6561'}
+{'loss': '0.6515', 'grad_norm': '6.719', 'learning_rate': '3.439e-06', 'num_tokens': '2.518e+06', 'mean_token_accuracy': '0.8099', 'epoch': '0.6563'}
+{'loss': '0.8752', 'grad_norm': '22.5', 'learning_rate': '3.437e-06', 'num_tokens': '2.518e+06', 'mean_token_accuracy': '0.8354', 'epoch': '0.6565'}
+{'loss': '1.127', 'grad_norm': '11.81', 'learning_rate': '3.435e-06', 'num_tokens': '2.519e+06', 'mean_token_accuracy': '0.685', 'epoch': '0.6567'}
+{'loss': '1.307', 'grad_norm': '13.38', 'learning_rate': '3.433e-06', 'num_tokens': '2.519e+06', 'mean_token_accuracy': '0.6946', 'epoch': '0.6568'}
+{'loss': '1.228', 'grad_norm': '13.25', 'learning_rate': '3.432e-06', 'num_tokens': '2.52e+06', 'mean_token_accuracy': '0.6581', 'epoch': '0.657'}
+{'loss': '1.063', 'grad_norm': '41.5', 'learning_rate': '3.43e-06', 'num_tokens': '2.52e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.6572'}
+{'loss': '1.397', 'grad_norm': '15.06', 'learning_rate': '3.428e-06', 'num_tokens': '2.521e+06', 'mean_token_accuracy': '0.6437', 'epoch': '0.6573'}
+{'loss': '0.6605', 'grad_norm': '7.875', 'learning_rate': '3.427e-06', 'num_tokens': '2.522e+06', 'mean_token_accuracy': '0.805', 'epoch': '0.6575'}
+{'loss': '2.41', 'grad_norm': '12.81', 'learning_rate': '3.425e-06', 'num_tokens': '2.523e+06', 'mean_token_accuracy': '0.5395', 'epoch': '0.6577'}
+{'loss': '1.708', 'grad_norm': '10.5', 'learning_rate': '3.423e-06', 'num_tokens': '2.524e+06', 'mean_token_accuracy': '0.6557', 'epoch': '0.6579'}
+{'loss': '1.405', 'grad_norm': '11.56', 'learning_rate': '3.421e-06', 'num_tokens': '2.525e+06', 'mean_token_accuracy': '0.6038', 'epoch': '0.658'}
+{'loss': '1.324', 'grad_norm': '11.69', 'learning_rate': '3.42e-06', 'num_tokens': '2.525e+06', 'mean_token_accuracy': '0.6464', 'epoch': '0.6582'}
+{'loss': '1.17', 'grad_norm': '12.19', 'learning_rate': '3.418e-06', 'num_tokens': '2.526e+06', 'mean_token_accuracy': '0.6885', 'epoch': '0.6584'}
+{'loss': '1.244', 'grad_norm': '17', 'learning_rate': '3.416e-06', 'num_tokens': '2.526e+06', 'mean_token_accuracy': '0.7452', 'epoch': '0.6586'}
+{'loss': '0.8116', 'grad_norm': '28.88', 'learning_rate': '3.414e-06', 'num_tokens': '2.527e+06', 'mean_token_accuracy': '0.8201', 'epoch': '0.6587'}
+{'loss': '1.064', 'grad_norm': '13.38', 'learning_rate': '3.413e-06', 'num_tokens': '2.527e+06', 'mean_token_accuracy': '0.6871', 'epoch': '0.6589'}
+{'loss': '1.707', 'grad_norm': '19.12', 'learning_rate': '3.411e-06', 'num_tokens': '2.528e+06', 'mean_token_accuracy': '0.6234', 'epoch': '0.6591'}
+{'loss': '1.641', 'grad_norm': '34', 'learning_rate': '3.409e-06', 'num_tokens': '2.528e+06', 'mean_token_accuracy': '0.6373', 'epoch': '0.6593'}
+{'loss': '1.639', 'grad_norm': '22.62', 'learning_rate': '3.407e-06', 'num_tokens': '2.528e+06', 'mean_token_accuracy': '0.6991', 'epoch': '0.6594'}
+{'loss': '1.358', 'grad_norm': '13.38', 'learning_rate': '3.406e-06', 'num_tokens': '2.529e+06', 'mean_token_accuracy': '0.6278', 'epoch': '0.6596'}
+{'loss': '0.6359', 'grad_norm': '14.31', 'learning_rate': '3.404e-06', 'num_tokens': '2.529e+06', 'mean_token_accuracy': '0.864', 'epoch': '0.6598'}
+{'loss': '0.5897', 'grad_norm': '6.344', 'learning_rate': '3.402e-06', 'num_tokens': '2.531e+06', 'mean_token_accuracy': '0.8264', 'epoch': '0.66'}
+{'loss': '2.083', 'grad_norm': '16.75', 'learning_rate': '3.4e-06', 'num_tokens': '2.531e+06', 'mean_token_accuracy': '0.5465', 'epoch': '0.6601'}
+{'loss': '0.5669', 'grad_norm': '18.75', 'learning_rate': '3.399e-06', 'num_tokens': '2.532e+06', 'mean_token_accuracy': '0.8745', 'epoch': '0.6603'}
+{'loss': '1.91', 'grad_norm': '15.75', 'learning_rate': '3.397e-06', 'num_tokens': '2.532e+06', 'mean_token_accuracy': '0.5961', 'epoch': '0.6605'}
+{'loss': '1.297', 'grad_norm': '13.19', 'learning_rate': '3.395e-06', 'num_tokens': '2.533e+06', 'mean_token_accuracy': '0.6783', 'epoch': '0.6606'}
+{'loss': '0.6575', 'grad_norm': '35.25', 'learning_rate': '3.394e-06', 'num_tokens': '2.533e+06', 'mean_token_accuracy': '0.8462', 'epoch': '0.6608'}
+{'loss': '2.145', 'grad_norm': '13.25', 'learning_rate': '3.392e-06', 'num_tokens': '2.534e+06', 'mean_token_accuracy': '0.5461', 'epoch': '0.661'}
+{'loss': '2.108', 'grad_norm': '20.75', 'learning_rate': '3.39e-06', 'num_tokens': '2.534e+06', 'mean_token_accuracy': '0.5619', 'epoch': '0.6612'}
+{'loss': '1.164', 'grad_norm': '33', 'learning_rate': '3.388e-06', 'num_tokens': '2.534e+06', 'mean_token_accuracy': '0.7159', 'epoch': '0.6613'}
+{'loss': '1.693', 'grad_norm': '11.25', 'learning_rate': '3.387e-06', 'num_tokens': '2.535e+06', 'mean_token_accuracy': '0.6305', 'epoch': '0.6615'}
+{'loss': '1.497', 'grad_norm': '29.12', 'learning_rate': '3.385e-06', 'num_tokens': '2.535e+06', 'mean_token_accuracy': '0.7014', 'epoch': '0.6617'}
+{'loss': '1.041', 'grad_norm': '13.94', 'learning_rate': '3.383e-06', 'num_tokens': '2.536e+06', 'mean_token_accuracy': '0.7063', 'epoch': '0.6619'}
+{'loss': '1.746', 'grad_norm': '38', 'learning_rate': '3.381e-06', 'num_tokens': '2.536e+06', 'mean_token_accuracy': '0.5814', 'epoch': '0.662'}
+{'loss': '1.334', 'grad_norm': '14.31', 'learning_rate': '3.38e-06', 'num_tokens': '2.536e+06', 'mean_token_accuracy': '0.7169', 'epoch': '0.6622'}
+{'loss': '1.629', 'grad_norm': '11.44', 'learning_rate': '3.378e-06', 'num_tokens': '2.537e+06', 'mean_token_accuracy': '0.6444', 'epoch': '0.6624'}
+{'loss': '1.685', 'grad_norm': '25.5', 'learning_rate': '3.376e-06', 'num_tokens': '2.537e+06', 'mean_token_accuracy': '0.6321', 'epoch': '0.6626'}
+{'loss': '1.213', 'grad_norm': '14.12', 'learning_rate': '3.374e-06', 'num_tokens': '2.538e+06', 'mean_token_accuracy': '0.6644', 'epoch': '0.6627'}
+{'loss': '1.195', 'grad_norm': '10.31', 'learning_rate': '3.373e-06', 'num_tokens': '2.539e+06', 'mean_token_accuracy': '0.733', 'epoch': '0.6629'}
+{'loss': '1.038', 'grad_norm': '11.56', 'learning_rate': '3.371e-06', 'num_tokens': '2.54e+06', 'mean_token_accuracy': '0.6985', 'epoch': '0.6631'}
+{'loss': '1.123', 'grad_norm': '12.94', 'learning_rate': '3.369e-06', 'num_tokens': '2.54e+06', 'mean_token_accuracy': '0.6786', 'epoch': '0.6633'}
+{'loss': '1.429', 'grad_norm': '20.12', 'learning_rate': '3.367e-06', 'num_tokens': '2.541e+06', 'mean_token_accuracy': '0.7235', 'epoch': '0.6634'}
+{'loss': '1.371', 'grad_norm': '9.688', 'learning_rate': '3.366e-06', 'num_tokens': '2.543e+06', 'mean_token_accuracy': '0.684', 'epoch': '0.6636'}
+{'loss': '1.208', 'grad_norm': '31.38', 'learning_rate': '3.364e-06', 'num_tokens': '2.543e+06', 'mean_token_accuracy': '0.7114', 'epoch': '0.6638'}
+{'loss': '1.319', 'grad_norm': '25.5', 'learning_rate': '3.362e-06', 'num_tokens': '2.543e+06', 'mean_token_accuracy': '0.7081', 'epoch': '0.6639'}
+{'loss': '1.87', 'grad_norm': '18.75', 'learning_rate': '3.361e-06', 'num_tokens': '2.543e+06', 'mean_token_accuracy': '0.6023', 'epoch': '0.6641'}
+{'loss': '1.756', 'grad_norm': '15.5', 'learning_rate': '3.359e-06', 'num_tokens': '2.544e+06', 'mean_token_accuracy': '0.6346', 'epoch': '0.6643'}
+{'loss': '1.344', 'grad_norm': '30.75', 'learning_rate': '3.357e-06', 'num_tokens': '2.544e+06', 'mean_token_accuracy': '0.7615', 'epoch': '0.6645'}
+{'loss': '1.214', 'grad_norm': '18', 'learning_rate': '3.355e-06', 'num_tokens': '2.544e+06', 'mean_token_accuracy': '0.749', 'epoch': '0.6646'}
+{'loss': '2.038', 'grad_norm': '15.56', 'learning_rate': '3.354e-06', 'num_tokens': '2.545e+06', 'mean_token_accuracy': '0.5809', 'epoch': '0.6648'}
+{'loss': '0.9113', 'grad_norm': '19.5', 'learning_rate': '3.352e-06', 'num_tokens': '2.545e+06', 'mean_token_accuracy': '0.7773', 'epoch': '0.665'}
+{'loss': '1.782', 'grad_norm': '12.88', 'learning_rate': '3.35e-06', 'num_tokens': '2.546e+06', 'mean_token_accuracy': '0.6182', 'epoch': '0.6652'}
+
 67%|██████▋   | 3831/5758 [21:47<10:24,  3.08it/s]
 67%|██████▋   | 3832/5758 [21:47<10:37,  3.02it/s]
                                                   

 67%|██████▋   | 3832/5758 [21:47<10:37,  3.02it/s]
 67%|██████▋   | 3833/5758 [21:47<10:06,  3.17it/s]
                                                   

 67%|██████▋   | 3833/5758 [21:47<10:06,  3.17it/s]
 67%|██████▋   | 3834/5758 [21:48<09:54,  3.24it/s]
                                                   

 67%|██████▋   | 3834/5758 [21:48<09:54,  3.24it/s]
 67%|██████▋   | 3835/5758 [21:48<10:27,  3.07it/s]
                                                   

 67%|██████▋   | 3835/5758 [21:48<10:27,  3.07it/s]
 67%|██████▋   | 3836/5758 [21:48<10:46,  2.97it/s]
                                                   

 67%|██████▋   | 3836/5758 [21:48<10:46,  2.97it/s]
 67%|██████▋   | 3837/5758 [21:49<10:20,  3.09it/s]
                                                   

 67%|██████▋   | 3837/5758 [21:49<10:20,  3.09it/s]
 67%|██████▋   | 3838/5758 [21:49<10:01,  3.19it/s]
                                                   

 67%|██████▋   | 3838/5758 [21:49<10:01,  3.19it/s]
 67%|██████▋   | 3839/5758 [21:49<10:26,  3.06it/s]
                                                   

 67%|██████▋   | 3839/5758 [21:49<10:26,  3.06it/s]
 67%|██████▋   | 3840/5758 [21:50<10:18,  3.10it/s]
                                                   

 67%|██████▋   | 3840/5758 [21:50<10:18,  3.10it/s]
 67%|██████▋   | 3841/5758 [21:50<10:11,  3.13it/s]
                                                   

 67%|██████▋   | 3841/5758 [21:50<10:11,  3.13it/s]
 67%|██████▋   | 3842/5758 [21:50<10:10,  3.14it/s]
                                                   

 67%|██████▋   | 3842/5758 [21:50<10:10,  3.14it/s]
 67%|██████▋   | 3843/5758 [21:51<10:08,  3.15it/s]
                                                   

 67%|██████▋   | 3843/5758 [21:51<10:08,  3.15it/s]
 67%|██████▋   | 3844/5758 [21:51<11:37,  2.75it/s]
                                                   

 67%|██████▋   | 3844/5758 [21:51<11:37,  2.75it/s]
 67%|██████▋   | 3845/5758 [21:51<11:18,  2.82it/s]
                                                   

 67%|██████▋   | 3845/5758 [21:51<11:18,  2.82it/s]
 67%|██████▋   | 3846/5758 [21:52<11:21,  2.81it/s]
                                                   

 67%|██████▋   | 3846/5758 [21:52<11:21,  2.81it/s]
 67%|██████▋   | 3847/5758 [21:52<11:25,  2.79it/s]
                                                   

 67%|██████▋   | 3847/5758 [21:52<11:25,  2.79it/s]
 67%|██████▋   | 3848/5758 [21:52<11:19,  2.81it/s]
                                                   

 67%|██████▋   | 3848/5758 [21:52<11:19,  2.81it/s]
 67%|██████▋   | 3849/5758 [21:53<10:59,  2.90it/s]
                                                   

 67%|██████▋   | 3849/5758 [21:53<10:59,  2.90it/s]
 67%|██████▋   | 3850/5758 [21:53<11:31,  2.76it/s]
                                                   

 67%|██████▋   | 3850/5758 [21:53<11:31,  2.76it/s]
 67%|██████▋   | 3851/5758 [21:53<10:51,  2.93it/s]
                                                   

 67%|██████▋   | 3851/5758 [21:53<10:51,  2.93it/s]
 67%|██████▋   | 3852/5758 [21:54<10:30,  3.02it/s]
                                                   

 67%|██████▋   | 3852/5758 [21:54<10:30,  3.02it/s]
 67%|██████▋   | 3853/5758 [21:54<10:46,  2.95it/s]
                                                   

 67%|██████▋   | 3853/5758 [21:54<10:46,  2.95it/s]
 67%|██████▋   | 3854/5758 [21:54<10:26,  3.04it/s]
                                                   

 67%|██████▋   | 3854/5758 [21:54<10:26,  3.04it/s]
 67%|██████▋   | 3855/5758 [21:55<10:06,  3.14it/s]
                                                   

 67%|██████▋   | 3855/5758 [21:55<10:06,  3.14it/s]
 67%|██████▋   | 3856/5758 [21:55<10:02,  3.16it/s]
                                                   

 67%|██████▋   | 3856/5758 [21:55<10:02,  3.16it/s]
 67%|██████▋   | 3857/5758 [21:55<09:49,  3.23it/s]
                                                   

 67%|██████▋   | 3857/5758 [21:55<09:49,  3.23it/s]
 67%|██████▋   | 3858/5758 [21:56<09:43,  3.25it/s]
                                                   

 67%|██████▋   | 3858/5758 [21:56<09:43,  3.25it/s]
 67%|██████▋   | 3859/5758 [21:56<09:46,  3.24it/s]
                                                   

 67%|██████▋   | 3859/5758 [21:56<09:46,  3.24it/s]
 67%|██████▋   | 3860/5758 [21:56<10:01,  3.15it/s]
                                                   

 67%|██████▋   | 3860/5758 [21:56<10:01,  3.15it/s]
 67%|██████▋   | 3861/5758 [21:57<09:55,  3.19it/s]
                                                   

 67%|██████▋   | 3861/5758 [21:57<09:55,  3.19it/s]
 67%|██████▋   | 3862/5758 [21:57<09:53,  3.19it/s]
                                                   

 67%|██████▋   | 3862/5758 [21:57<09:53,  3.19it/s]
 67%|██████▋   | 3863/5758 [21:57<10:16,  3.07it/s]
                                                   

 67%|██████▋   | 3863/5758 [21:57<10:16,  3.07it/s]
 67%|██████▋   | 3864/5758 [21:58<10:37,  2.97it/s]
                                                   

 67%|██████▋   | 3864/5758 [21:58<10:37,  2.97it/s]
 67%|██████▋   | 3865/5758 [21:58<10:58,  2.87it/s]
                                                   

 67%|██████▋   | 3865/5758 [21:58<10:58,  2.87it/s]
 67%|██████▋   | 3866/5758 [21:58<10:33,  2.99it/s]
                                                   

 67%|██████▋   | 3866/5758 [21:58<10:33,  2.99it/s]
 67%|██████▋   | 3867/5758 [21:59<11:04,  2.85it/s]
                                                   

 67%|██████▋   | 3867/5758 [21:59<11:04,  2.85it/s]
 67%|██████▋   | 3868/5758 [21:59<11:25,  2.76it/s]
                                                   

 67%|██████▋   | 3868/5758 [21:59<11:25,  2.76it/s]
 67%|██████▋   | 3869/5758 [21:59<11:29,  2.74it/s]
                                                   

 67%|██████▋   | 3869/5758 [21:59<11:29,  2.74it/s]
 67%|██████▋   | 3870/5758 [22:00<12:18,  2.56it/s]
                                                   

 67%|██████▋   | 3870/5758 [22:00<12:18,  2.56it/s]
 67%|██████▋   | 3871/5758 [22:00<11:32,  2.72it/s]
                                                   

 67%|██████▋   | 3871/5758 [22:00<11:32,  2.72it/s]
 67%|██████▋   | 3872/5758 [22:00<11:18,  2.78it/s]
                                                   

 67%|██████▋   | 3872/5758 [22:00<11:18,  2.78it/s]
 67%|██████▋   | 3873/5758 [22:01<11:06,  2.83it/s]
                                                   

 67%|██████▋   | 3873/5758 [22:01<11:06,  2.83it/s]
 67%|██████▋   | 3874/5758 [22:01<10:53,  2.88it/s]
                                                   

 67%|██████▋   | 3874/5758 [22:01<10:53,  2.88it/s]
 67%|██████▋   | 3875/5758 [22:02<11:10,  2.81it/s]
                                                   

 67%|██████▋   | 3875/5758 [22:02<11:10,  2.81it/s]
 67%|██████▋   | 3876/5758 [22:02<10:51,  2.89it/s]
                                                   

 67%|██████▋   | 3876/5758 [22:02<10:51,  2.89it/s]
 67%|██████▋   | 3877/5758 [22:02<10:21,  3.03it/s]
                                                   

 67%|██████▋   | 3877/5758 [22:02<10:21,  3.03it/s]
 67%|██████▋   | 3878/5758 [22:03<10:38,  2.94it/s]
                                                   

 67%|██████▋   | 3878/5758 [22:03<10:38,  2.94it/s]
 67%|██████▋   | 3879/5758 [22:03<10:17,  3.05it/s]
                                                   

 67%|██████▋   | 3879/5758 [22:03<10:17,  3.05it/s]
 67%|██████▋   | 3880/5758 [22:03<09:58,  3.14it/s]
                                                   

 67%|██████▋   | 3880/5758 [22:03<09:58,  3.14it/s]
 67%|██████▋   | 3881/5758 [22:03<10:03,  3.11it/s]
                                                   

 67%|██████▋   | 3881/5758 [22:03<10:03,  3.11it/s]
 67%|██████▋   | 3882/5758 [22:04<09:51,  3.17it/s]
                                                   

 67%|██████▋   | 3882/5758 [22:04<09:51,  3.17it/s]
 67%|██████▋   | 3883/5758 [22:04<10:21,  3.02it/s]
                                                   

 67%|██████▋   | 3883/5758 [22:04<10:21,  3.02it/s]
 67%|██████▋   | 3884/5758 [22:04<10:39,  2.93it/s]
                                                   

 67%|██████▋   | 3884/5758 [22:04<10:39,  2.93it/s]
 67%|██████▋   | 3885/5758 [22:05<10:37,  2.94it/s]
                                                   

 67%|██████▋   | 3885/5758 [22:05<10:37,  2.94it/s]
 67%|██████▋   | 3886/5758 [22:05<10:19,  3.02it/s]
                                                   
{'loss': '1.616', 'grad_norm': '37', 'learning_rate': '3.348e-06', 'num_tokens': '2.546e+06', 'mean_token_accuracy': '0.7386', 'epoch': '0.6653'}
+{'loss': '0.837', 'grad_norm': '13.62', 'learning_rate': '3.347e-06', 'num_tokens': '2.546e+06', 'mean_token_accuracy': '0.8233', 'epoch': '0.6655'}
+{'loss': '0.8088', 'grad_norm': '50', 'learning_rate': '3.345e-06', 'num_tokens': '2.546e+06', 'mean_token_accuracy': '0.8333', 'epoch': '0.6657'}
+{'loss': '2.147', 'grad_norm': '16.75', 'learning_rate': '3.343e-06', 'num_tokens': '2.547e+06', 'mean_token_accuracy': '0.5743', 'epoch': '0.6659'}
+{'loss': '1.186', 'grad_norm': '14.44', 'learning_rate': '3.341e-06', 'num_tokens': '2.548e+06', 'mean_token_accuracy': '0.6962', 'epoch': '0.666'}
+{'loss': '2.365', 'grad_norm': '15', 'learning_rate': '3.34e-06', 'num_tokens': '2.548e+06', 'mean_token_accuracy': '0.5119', 'epoch': '0.6662'}
+{'loss': '1.157', 'grad_norm': '9.125', 'learning_rate': '3.338e-06', 'num_tokens': '2.549e+06', 'mean_token_accuracy': '0.7394', 'epoch': '0.6664'}
+{'loss': '1.113', 'grad_norm': '13.19', 'learning_rate': '3.336e-06', 'num_tokens': '2.55e+06', 'mean_token_accuracy': '0.6894', 'epoch': '0.6666'}
+{'loss': '1.244', 'grad_norm': '13.19', 'learning_rate': '3.334e-06', 'num_tokens': '2.55e+06', 'mean_token_accuracy': '0.7374', 'epoch': '0.6667'}
+{'loss': '1.211', 'grad_norm': '18.88', 'learning_rate': '3.333e-06', 'num_tokens': '2.551e+06', 'mean_token_accuracy': '0.7123', 'epoch': '0.6669'}
+{'loss': '1.133', 'grad_norm': '11.69', 'learning_rate': '3.331e-06', 'num_tokens': '2.551e+06', 'mean_token_accuracy': '0.7172', 'epoch': '0.6671'}
+{'loss': '1.327', 'grad_norm': '8.062', 'learning_rate': '3.329e-06', 'num_tokens': '2.553e+06', 'mean_token_accuracy': '0.6947', 'epoch': '0.6672'}
+{'loss': '1.175', 'grad_norm': '33.5', 'learning_rate': '3.328e-06', 'num_tokens': '2.553e+06', 'mean_token_accuracy': '0.7333', 'epoch': '0.6674'}
+{'loss': '1.667', 'grad_norm': '13.19', 'learning_rate': '3.326e-06', 'num_tokens': '2.554e+06', 'mean_token_accuracy': '0.6694', 'epoch': '0.6676'}
+{'loss': '1.196', 'grad_norm': '11.31', 'learning_rate': '3.324e-06', 'num_tokens': '2.554e+06', 'mean_token_accuracy': '0.6525', 'epoch': '0.6678'}
+{'loss': '1.091', 'grad_norm': '6.5', 'learning_rate': '3.322e-06', 'num_tokens': '2.556e+06', 'mean_token_accuracy': '0.7357', 'epoch': '0.6679'}
+{'loss': '1.216', 'grad_norm': '15.56', 'learning_rate': '3.321e-06', 'num_tokens': '2.557e+06', 'mean_token_accuracy': '0.6525', 'epoch': '0.6681'}
+{'loss': '2.048', 'grad_norm': '18.62', 'learning_rate': '3.319e-06', 'num_tokens': '2.557e+06', 'mean_token_accuracy': '0.5636', 'epoch': '0.6683'}
+{'loss': '0.4311', 'grad_norm': '16.38', 'learning_rate': '3.317e-06', 'num_tokens': '2.558e+06', 'mean_token_accuracy': '0.9154', 'epoch': '0.6685'}
+{'loss': '0.7744', 'grad_norm': '36', 'learning_rate': '3.315e-06', 'num_tokens': '2.558e+06', 'mean_token_accuracy': '0.7931', 'epoch': '0.6686'}
+{'loss': '1.586', 'grad_norm': '50.5', 'learning_rate': '3.314e-06', 'num_tokens': '2.558e+06', 'mean_token_accuracy': '0.7324', 'epoch': '0.6688'}
+{'loss': '0.9617', 'grad_norm': '12.38', 'learning_rate': '3.312e-06', 'num_tokens': '2.559e+06', 'mean_token_accuracy': '0.7223', 'epoch': '0.669'}
+{'loss': '1.736', 'grad_norm': '16.75', 'learning_rate': '3.31e-06', 'num_tokens': '2.559e+06', 'mean_token_accuracy': '0.6188', 'epoch': '0.6692'}
+{'loss': '1.017', 'grad_norm': '13.06', 'learning_rate': '3.308e-06', 'num_tokens': '2.56e+06', 'mean_token_accuracy': '0.7573', 'epoch': '0.6693'}
+{'loss': '1.306', 'grad_norm': '9.25', 'learning_rate': '3.307e-06', 'num_tokens': '2.561e+06', 'mean_token_accuracy': '0.6839', 'epoch': '0.6695'}
+{'loss': '1.05', 'grad_norm': '11.62', 'learning_rate': '3.305e-06', 'num_tokens': '2.562e+06', 'mean_token_accuracy': '0.6948', 'epoch': '0.6697'}
+{'loss': '1.155', 'grad_norm': '12.75', 'learning_rate': '3.303e-06', 'num_tokens': '2.562e+06', 'mean_token_accuracy': '0.6943', 'epoch': '0.6699'}
+{'loss': '1.14', 'grad_norm': '13.06', 'learning_rate': '3.301e-06', 'num_tokens': '2.563e+06', 'mean_token_accuracy': '0.6721', 'epoch': '0.67'}
+{'loss': '0.6915', 'grad_norm': '19.5', 'learning_rate': '3.3e-06', 'num_tokens': '2.563e+06', 'mean_token_accuracy': '0.8177', 'epoch': '0.6702'}
+{'loss': '1.341', 'grad_norm': '11.69', 'learning_rate': '3.298e-06', 'num_tokens': '2.564e+06', 'mean_token_accuracy': '0.6142', 'epoch': '0.6704'}
+{'loss': '1.261', 'grad_norm': '13.88', 'learning_rate': '3.296e-06', 'num_tokens': '2.565e+06', 'mean_token_accuracy': '0.659', 'epoch': '0.6705'}
+{'loss': '1.168', 'grad_norm': '11.62', 'learning_rate': '3.295e-06', 'num_tokens': '2.566e+06', 'mean_token_accuracy': '0.6895', 'epoch': '0.6707'}
+{'loss': '1.737', 'grad_norm': '17.62', 'learning_rate': '3.293e-06', 'num_tokens': '2.566e+06', 'mean_token_accuracy': '0.5934', 'epoch': '0.6709'}
+{'loss': '1.955', 'grad_norm': '15.94', 'learning_rate': '3.291e-06', 'num_tokens': '2.567e+06', 'mean_token_accuracy': '0.5922', 'epoch': '0.6711'}
+{'loss': '1.108', 'grad_norm': '28.38', 'learning_rate': '3.289e-06', 'num_tokens': '2.567e+06', 'mean_token_accuracy': '0.7951', 'epoch': '0.6712'}
+{'loss': '1.207', 'grad_norm': '14.19', 'learning_rate': '3.288e-06', 'num_tokens': '2.567e+06', 'mean_token_accuracy': '0.6588', 'epoch': '0.6714'}
+{'loss': '1.42', 'grad_norm': '10.5', 'learning_rate': '3.286e-06', 'num_tokens': '2.569e+06', 'mean_token_accuracy': '0.6771', 'epoch': '0.6716'}
+{'loss': '0.5876', 'grad_norm': '8.688', 'learning_rate': '3.284e-06', 'num_tokens': '2.57e+06', 'mean_token_accuracy': '0.8183', 'epoch': '0.6718'}
+{'loss': '2.06', 'grad_norm': '14.31', 'learning_rate': '3.282e-06', 'num_tokens': '2.57e+06', 'mean_token_accuracy': '0.585', 'epoch': '0.6719'}
+{'loss': '1.329', 'grad_norm': '12', 'learning_rate': '3.281e-06', 'num_tokens': '2.571e+06', 'mean_token_accuracy': '0.6101', 'epoch': '0.6721'}
+{'loss': '0.6262', 'grad_norm': '8.312', 'learning_rate': '3.279e-06', 'num_tokens': '2.573e+06', 'mean_token_accuracy': '0.812', 'epoch': '0.6723'}
+{'loss': '2.212', 'grad_norm': '27.88', 'learning_rate': '3.277e-06', 'num_tokens': '2.573e+06', 'mean_token_accuracy': '0.5333', 'epoch': '0.6725'}
+{'loss': '1.07', 'grad_norm': '7.156', 'learning_rate': '3.275e-06', 'num_tokens': '2.574e+06', 'mean_token_accuracy': '0.74', 'epoch': '0.6726'}
+{'loss': '1.247', 'grad_norm': '7.375', 'learning_rate': '3.274e-06', 'num_tokens': '2.576e+06', 'mean_token_accuracy': '0.7229', 'epoch': '0.6728'}
+{'loss': '1.016', 'grad_norm': '13.12', 'learning_rate': '3.272e-06', 'num_tokens': '2.577e+06', 'mean_token_accuracy': '0.7205', 'epoch': '0.673'}
+{'loss': '1.175', 'grad_norm': '12', 'learning_rate': '3.27e-06', 'num_tokens': '2.577e+06', 'mean_token_accuracy': '0.6797', 'epoch': '0.6732'}
+{'loss': '0.8423', 'grad_norm': '13.06', 'learning_rate': '3.268e-06', 'num_tokens': '2.578e+06', 'mean_token_accuracy': '0.7691', 'epoch': '0.6733'}
+{'loss': '1.452', 'grad_norm': '21.62', 'learning_rate': '3.267e-06', 'num_tokens': '2.578e+06', 'mean_token_accuracy': '0.6576', 'epoch': '0.6735'}
+{'loss': '1.174', 'grad_norm': '13.62', 'learning_rate': '3.265e-06', 'num_tokens': '2.579e+06', 'mean_token_accuracy': '0.6947', 'epoch': '0.6737'}
+{'loss': '1.104', 'grad_norm': '14.94', 'learning_rate': '3.263e-06', 'num_tokens': '2.58e+06', 'mean_token_accuracy': '0.6866', 'epoch': '0.6738'}
+{'loss': '2.015', 'grad_norm': '16.75', 'learning_rate': '3.262e-06', 'num_tokens': '2.58e+06', 'mean_token_accuracy': '0.5971', 'epoch': '0.674'}
+{'loss': '1.233', 'grad_norm': '13.5', 'learning_rate': '3.26e-06', 'num_tokens': '2.581e+06', 'mean_token_accuracy': '0.6765', 'epoch': '0.6742'}
+{'loss': '1.942', 'grad_norm': '18.38', 'learning_rate': '3.258e-06', 'num_tokens': '2.581e+06', 'mean_token_accuracy': '0.5807', 'epoch': '0.6744'}
+{'loss': '1.06', 'grad_norm': '6.312', 'learning_rate': '3.256e-06', 'num_tokens': '2.583e+06', 'mean_token_accuracy': '0.7245', 'epoch': '0.6745'}
+{'loss': '1.89', 'grad_norm': '12.25', 'learning_rate': '3.255e-06', 'num_tokens': '2.584e+06', 'mean_token_accuracy': '0.6238', 'epoch': '0.6747'}
+
 67%|██████▋   | 3886/5758 [22:05<10:19,  3.02it/s]
 68%|██████▊   | 3887/5758 [22:05<10:17,  3.03it/s]
                                                   

 68%|██████▊   | 3887/5758 [22:05<10:17,  3.03it/s]
 68%|██████▊   | 3888/5758 [22:06<10:44,  2.90it/s]
                                                   

 68%|██████▊   | 3888/5758 [22:06<10:44,  2.90it/s]
 68%|██████▊   | 3889/5758 [22:06<10:55,  2.85it/s]
                                                   

 68%|██████▊   | 3889/5758 [22:06<10:55,  2.85it/s]
 68%|██████▊   | 3890/5758 [22:07<10:45,  2.89it/s]
                                                   

 68%|██████▊   | 3890/5758 [22:07<10:45,  2.89it/s]
 68%|██████▊   | 3891/5758 [22:07<10:41,  2.91it/s]
                                                   

 68%|██████▊   | 3891/5758 [22:07<10:41,  2.91it/s]
 68%|██████▊   | 3892/5758 [22:07<10:50,  2.87it/s]
                                                   

 68%|██████▊   | 3892/5758 [22:07<10:50,  2.87it/s]
 68%|██████▊   | 3893/5758 [22:08<10:42,  2.90it/s]
                                                   

 68%|██████▊   | 3893/5758 [22:08<10:42,  2.90it/s]
 68%|██████▊   | 3894/5758 [22:08<10:43,  2.90it/s]
                                                   

 68%|██████▊   | 3894/5758 [22:08<10:43,  2.90it/s]
 68%|██████▊   | 3895/5758 [22:08<10:17,  3.01it/s]
                                                   

 68%|██████▊   | 3895/5758 [22:08<10:17,  3.01it/s]
 68%|██████▊   | 3896/5758 [22:09<10:16,  3.02it/s]
                                                   

 68%|██████▊   | 3896/5758 [22:09<10:16,  3.02it/s]
 68%|██████▊   | 3897/5758 [22:09<11:17,  2.75it/s]
                                                   

 68%|██████▊   | 3897/5758 [22:09<11:17,  2.75it/s]
 68%|██████▊   | 3898/5758 [22:09<10:36,  2.92it/s]
                                                   

 68%|██████▊   | 3898/5758 [22:09<10:36,  2.92it/s]
 68%|██████▊   | 3899/5758 [22:10<10:12,  3.04it/s]
                                                   

 68%|██████▊   | 3899/5758 [22:10<10:12,  3.04it/s]
 68%|██████▊   | 3900/5758 [22:10<09:59,  3.10it/s]
                                                   

 68%|██████▊   | 3900/5758 [22:10<09:59,  3.10it/s]
 68%|██████▊   | 3901/5758 [22:10<09:43,  3.18it/s]
                                                   

 68%|██████▊   | 3901/5758 [22:10<09:43,  3.18it/s]
 68%|██████▊   | 3902/5758 [22:10<09:50,  3.15it/s]
                                                   

 68%|██████▊   | 3902/5758 [22:10<09:50,  3.15it/s]
 68%|██████▊   | 3903/5758 [22:11<09:40,  3.19it/s]
                                                   

 68%|██████▊   | 3903/5758 [22:11<09:40,  3.19it/s]
 68%|██████▊   | 3904/5758 [22:11<10:14,  3.01it/s]
                                                   

 68%|██████▊   | 3904/5758 [22:11<10:14,  3.01it/s]
 68%|██████▊   | 3905/5758 [22:11<09:52,  3.13it/s]
                                                   

 68%|██████▊   | 3905/5758 [22:11<09:52,  3.13it/s]
 68%|██████▊   | 3906/5758 [22:12<09:50,  3.14it/s]
                                                   

 68%|██████▊   | 3906/5758 [22:12<09:50,  3.14it/s]
 68%|██████▊   | 3907/5758 [22:12<09:41,  3.18it/s]
                                                   

 68%|██████▊   | 3907/5758 [22:12<09:41,  3.18it/s]
 68%|██████▊   | 3908/5758 [22:12<09:32,  3.23it/s]
                                                   

 68%|██████▊   | 3908/5758 [22:12<09:32,  3.23it/s]
 68%|██████▊   | 3909/5758 [22:13<09:31,  3.23it/s]
                                                   

 68%|██████▊   | 3909/5758 [22:13<09:31,  3.23it/s]
 68%|██████▊   | 3910/5758 [22:13<09:52,  3.12it/s]
                                                   

 68%|██████▊   | 3910/5758 [22:13<09:52,  3.12it/s]
 68%|██████▊   | 3911/5758 [22:13<09:52,  3.12it/s]
                                                   

 68%|██████▊   | 3911/5758 [22:13<09:52,  3.12it/s]
 68%|██████▊   | 3912/5758 [22:14<09:43,  3.16it/s]
                                                   

 68%|██████▊   | 3912/5758 [22:14<09:43,  3.16it/s]
 68%|██████▊   | 3913/5758 [22:14<09:35,  3.20it/s]
                                                   

 68%|██████▊   | 3913/5758 [22:14<09:35,  3.20it/s]
 68%|██████▊   | 3914/5758 [22:14<09:22,  3.28it/s]
                                                   

 68%|██████▊   | 3914/5758 [22:14<09:22,  3.28it/s]
 68%|██████▊   | 3915/5758 [22:15<09:26,  3.25it/s]
                                                   

 68%|██████▊   | 3915/5758 [22:15<09:26,  3.25it/s]
 68%|██████▊   | 3916/5758 [22:15<09:38,  3.18it/s]
                                                   

 68%|██████▊   | 3916/5758 [22:15<09:38,  3.18it/s]
 68%|██████▊   | 3917/5758 [22:15<09:25,  3.26it/s]
                                                   

 68%|██████▊   | 3917/5758 [22:15<09:25,  3.26it/s]
 68%|██████▊   | 3918/5758 [22:16<09:47,  3.13it/s]
                                                   

 68%|██████▊   | 3918/5758 [22:16<09:47,  3.13it/s]
 68%|██████▊   | 3919/5758 [22:16<09:44,  3.15it/s]
                                                   

 68%|██████▊   | 3919/5758 [22:16<09:44,  3.15it/s]
 68%|██████▊   | 3920/5758 [22:16<09:53,  3.09it/s]
                                                   

 68%|██████▊   | 3920/5758 [22:16<09:53,  3.09it/s]
 68%|██████▊   | 3921/5758 [22:17<10:17,  2.98it/s]
                                                   

 68%|██████▊   | 3921/5758 [22:17<10:17,  2.98it/s]
 68%|██████▊   | 3922/5758 [22:17<10:22,  2.95it/s]
                                                   

 68%|██████▊   | 3922/5758 [22:17<10:22,  2.95it/s]
 68%|██████▊   | 3923/5758 [22:17<10:46,  2.84it/s]
                                                   

 68%|██████▊   | 3923/5758 [22:17<10:46,  2.84it/s]
 68%|██████▊   | 3924/5758 [22:18<11:04,  2.76it/s]
                                                   

 68%|██████▊   | 3924/5758 [22:18<11:04,  2.76it/s]
 68%|██████▊   | 3925/5758 [22:18<12:15,  2.49it/s]
                                                   

 68%|██████▊   | 3925/5758 [22:18<12:15,  2.49it/s]
 68%|██████▊   | 3926/5758 [22:19<11:54,  2.57it/s]
                                                   

 68%|██████▊   | 3926/5758 [22:19<11:54,  2.57it/s]
 68%|██████▊   | 3927/5758 [22:19<11:14,  2.71it/s]
                                                   

 68%|██████▊   | 3927/5758 [22:19<11:14,  2.71it/s]
 68%|██████▊   | 3928/5758 [22:19<11:19,  2.69it/s]
                                                   

 68%|██████▊   | 3928/5758 [22:19<11:19,  2.69it/s]
 68%|██████▊   | 3929/5758 [22:20<10:39,  2.86it/s]
                                                   

 68%|██████▊   | 3929/5758 [22:20<10:39,  2.86it/s]
 68%|██████▊   | 3930/5758 [22:20<10:08,  3.01it/s]
                                                   

 68%|██████▊   | 3930/5758 [22:20<10:08,  3.01it/s]
 68%|██████▊   | 3931/5758 [22:20<09:50,  3.09it/s]
                                                   

 68%|██████▊   | 3931/5758 [22:20<09:50,  3.09it/s]
 68%|██████▊   | 3932/5758 [22:20<10:09,  3.00it/s]
                                                   

 68%|██████▊   | 3932/5758 [22:20<10:09,  3.00it/s]
 68%|██████▊   | 3933/5758 [22:21<10:37,  2.86it/s]
                                                   

 68%|██████▊   | 3933/5758 [22:21<10:37,  2.86it/s]
 68%|██████▊   | 3934/5758 [22:21<11:02,  2.75it/s]
                                                   

 68%|██████▊   | 3934/5758 [22:21<11:02,  2.75it/s]
 68%|██████▊   | 3935/5758 [22:22<10:27,  2.90it/s]
                                                   

 68%|██████▊   | 3935/5758 [22:22<10:27,  2.90it/s]
 68%|██████▊   | 3936/5758 [22:22<10:09,  2.99it/s]
                                                   

 68%|██████▊   | 3936/5758 [22:22<10:09,  2.99it/s]
 68%|██████▊   | 3937/5758 [22:22<09:47,  3.10it/s]
                                                   

 68%|██████▊   | 3937/5758 [22:22<09:47,  3.10it/s]
 68%|██████▊   | 3938/5758 [22:23<10:20,  2.93it/s]
                                                   

 68%|██████▊   | 3938/5758 [22:23<10:20,  2.93it/s]
 68%|██████▊   | 3939/5758 [22:23<10:05,  3.01it/s]
                                                   

 68%|██████▊   | 3939/5758 [22:23<10:05,  3.01it/s]
 68%|██████▊   | 3940/5758 [22:23<10:21,  2.93it/s]
                                                   

 68%|██████▊   | 3940/5758 [22:23<10:21,  2.93it/s]
 68%|██████▊   | 3941/5758 [22:24<10:48,  2.80it/s]
                                                   
{'loss': '1.109', 'grad_norm': '13.81', 'learning_rate': '3.253e-06', 'num_tokens': '2.585e+06', 'mean_token_accuracy': '0.6845', 'epoch': '0.6749'}
+{'loss': '0.7288', 'grad_norm': '22', 'learning_rate': '3.251e-06', 'num_tokens': '2.585e+06', 'mean_token_accuracy': '0.8177', 'epoch': '0.6751'}
+{'loss': '1.157', 'grad_norm': '21.25', 'learning_rate': '3.249e-06', 'num_tokens': '2.585e+06', 'mean_token_accuracy': '0.7214', 'epoch': '0.6752'}
+{'loss': '1.354', 'grad_norm': '11.69', 'learning_rate': '3.248e-06', 'num_tokens': '2.586e+06', 'mean_token_accuracy': '0.6534', 'epoch': '0.6754'}
+{'loss': '1.474', 'grad_norm': '7.562', 'learning_rate': '3.246e-06', 'num_tokens': '2.588e+06', 'mean_token_accuracy': '0.6778', 'epoch': '0.6756'}
+{'loss': '1.127', 'grad_norm': '6.688', 'learning_rate': '3.244e-06', 'num_tokens': '2.59e+06', 'mean_token_accuracy': '0.7245', 'epoch': '0.6758'}
+{'loss': '1.849', 'grad_norm': '19.5', 'learning_rate': '3.242e-06', 'num_tokens': '2.591e+06', 'mean_token_accuracy': '0.6248', 'epoch': '0.6759'}
+{'loss': '1.42', 'grad_norm': '19.62', 'learning_rate': '3.241e-06', 'num_tokens': '2.591e+06', 'mean_token_accuracy': '0.7072', 'epoch': '0.6761'}
+{'loss': '1.832', 'grad_norm': '20.75', 'learning_rate': '3.239e-06', 'num_tokens': '2.591e+06', 'mean_token_accuracy': '0.6118', 'epoch': '0.6763'}
+{'loss': '0.9915', 'grad_norm': '14', 'learning_rate': '3.237e-06', 'num_tokens': '2.592e+06', 'mean_token_accuracy': '0.7012', 'epoch': '0.6765'}
+{'loss': '1.264', 'grad_norm': '8.125', 'learning_rate': '3.235e-06', 'num_tokens': '2.594e+06', 'mean_token_accuracy': '0.719', 'epoch': '0.6766'}
+{'loss': '2.373', 'grad_norm': '31.38', 'learning_rate': '3.234e-06', 'num_tokens': '2.594e+06', 'mean_token_accuracy': '0.5178', 'epoch': '0.6768'}
+{'loss': '0.5479', 'grad_norm': '38.25', 'learning_rate': '3.232e-06', 'num_tokens': '2.594e+06', 'mean_token_accuracy': '0.8143', 'epoch': '0.677'}
+{'loss': '0.9981', 'grad_norm': '11.81', 'learning_rate': '3.23e-06', 'num_tokens': '2.595e+06', 'mean_token_accuracy': '0.7161', 'epoch': '0.6771'}
+{'loss': '2.134', 'grad_norm': '18', 'learning_rate': '3.229e-06', 'num_tokens': '2.595e+06', 'mean_token_accuracy': '0.5201', 'epoch': '0.6773'}
+{'loss': '2.098', 'grad_norm': '16.5', 'learning_rate': '3.227e-06', 'num_tokens': '2.595e+06', 'mean_token_accuracy': '0.5632', 'epoch': '0.6775'}
+{'loss': '1.365', 'grad_norm': '11.62', 'learning_rate': '3.225e-06', 'num_tokens': '2.596e+06', 'mean_token_accuracy': '0.6655', 'epoch': '0.6777'}
+{'loss': '1.824', 'grad_norm': '26.62', 'learning_rate': '3.223e-06', 'num_tokens': '2.596e+06', 'mean_token_accuracy': '0.6527', 'epoch': '0.6778'}
+{'loss': '0.6476', 'grad_norm': '19.62', 'learning_rate': '3.222e-06', 'num_tokens': '2.597e+06', 'mean_token_accuracy': '0.864', 'epoch': '0.678'}
+{'loss': '1.167', 'grad_norm': '13.25', 'learning_rate': '3.22e-06', 'num_tokens': '2.597e+06', 'mean_token_accuracy': '0.6987', 'epoch': '0.6782'}
+{'loss': '1.177', 'grad_norm': '14', 'learning_rate': '3.218e-06', 'num_tokens': '2.598e+06', 'mean_token_accuracy': '0.6601', 'epoch': '0.6784'}
+{'loss': '1.102', 'grad_norm': '13.62', 'learning_rate': '3.216e-06', 'num_tokens': '2.599e+06', 'mean_token_accuracy': '0.7145', 'epoch': '0.6785'}
+{'loss': '0.8703', 'grad_norm': '24.62', 'learning_rate': '3.215e-06', 'num_tokens': '2.599e+06', 'mean_token_accuracy': '0.7888', 'epoch': '0.6787'}
+{'loss': '0.9814', 'grad_norm': '7.969', 'learning_rate': '3.213e-06', 'num_tokens': '2.6e+06', 'mean_token_accuracy': '0.7419', 'epoch': '0.6789'}
+{'loss': '1.464', 'grad_norm': '26.38', 'learning_rate': '3.211e-06', 'num_tokens': '2.601e+06', 'mean_token_accuracy': '0.6737', 'epoch': '0.6791'}
+{'loss': '2.21', 'grad_norm': '10.88', 'learning_rate': '3.209e-06', 'num_tokens': '2.601e+06', 'mean_token_accuracy': '0.5483', 'epoch': '0.6792'}
+{'loss': '1.955', 'grad_norm': '24.75', 'learning_rate': '3.208e-06', 'num_tokens': '2.602e+06', 'mean_token_accuracy': '0.6221', 'epoch': '0.6794'}
+{'loss': '1.316', 'grad_norm': '13.38', 'learning_rate': '3.206e-06', 'num_tokens': '2.602e+06', 'mean_token_accuracy': '0.6522', 'epoch': '0.6796'}
+{'loss': '0.7126', 'grad_norm': '35.75', 'learning_rate': '3.204e-06', 'num_tokens': '2.602e+06', 'mean_token_accuracy': '0.8406', 'epoch': '0.6797'}
+{'loss': '1.137', 'grad_norm': '11.25', 'learning_rate': '3.203e-06', 'num_tokens': '2.603e+06', 'mean_token_accuracy': '0.6848', 'epoch': '0.6799'}
+{'loss': '1.237', 'grad_norm': '36.25', 'learning_rate': '3.201e-06', 'num_tokens': '2.603e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.6801'}
+{'loss': '0.5367', 'grad_norm': '14.06', 'learning_rate': '3.199e-06', 'num_tokens': '2.603e+06', 'mean_token_accuracy': '0.8717', 'epoch': '0.6803'}
+{'loss': '2.09', 'grad_norm': '20.25', 'learning_rate': '3.197e-06', 'num_tokens': '2.604e+06', 'mean_token_accuracy': '0.5919', 'epoch': '0.6804'}
+{'loss': '1.368', 'grad_norm': '13.06', 'learning_rate': '3.196e-06', 'num_tokens': '2.605e+06', 'mean_token_accuracy': '0.6306', 'epoch': '0.6806'}
+{'loss': '1.159', 'grad_norm': '7.031', 'learning_rate': '3.194e-06', 'num_tokens': '2.606e+06', 'mean_token_accuracy': '0.717', 'epoch': '0.6808'}
+{'loss': '2.126', 'grad_norm': '17.88', 'learning_rate': '3.192e-06', 'num_tokens': '2.607e+06', 'mean_token_accuracy': '0.5826', 'epoch': '0.681'}
+{'loss': '0.5073', 'grad_norm': '6.375', 'learning_rate': '3.19e-06', 'num_tokens': '2.609e+06', 'mean_token_accuracy': '0.8455', 'epoch': '0.6811'}
+{'loss': '1.497', 'grad_norm': '14.81', 'learning_rate': '3.189e-06', 'num_tokens': '2.609e+06', 'mean_token_accuracy': '0.6431', 'epoch': '0.6813'}
+{'loss': '1.599', 'grad_norm': '12.31', 'learning_rate': '3.187e-06', 'num_tokens': '2.61e+06', 'mean_token_accuracy': '0.656', 'epoch': '0.6815'}
+{'loss': '0.6475', 'grad_norm': '8.062', 'learning_rate': '3.185e-06', 'num_tokens': '2.611e+06', 'mean_token_accuracy': '0.8169', 'epoch': '0.6817'}
+{'loss': '1.481', 'grad_norm': '33.75', 'learning_rate': '3.183e-06', 'num_tokens': '2.612e+06', 'mean_token_accuracy': '0.7699', 'epoch': '0.6818'}
+{'loss': '1.012', 'grad_norm': '15.25', 'learning_rate': '3.182e-06', 'num_tokens': '2.612e+06', 'mean_token_accuracy': '0.7812', 'epoch': '0.682'}
+{'loss': '1.469', 'grad_norm': '10.88', 'learning_rate': '3.18e-06', 'num_tokens': '2.613e+06', 'mean_token_accuracy': '0.6872', 'epoch': '0.6822'}
+{'loss': '1.333', 'grad_norm': '11.31', 'learning_rate': '3.178e-06', 'num_tokens': '2.614e+06', 'mean_token_accuracy': '0.6945', 'epoch': '0.6824'}
+{'loss': '1.95', 'grad_norm': '21', 'learning_rate': '3.176e-06', 'num_tokens': '2.614e+06', 'mean_token_accuracy': '0.6258', 'epoch': '0.6825'}
+{'loss': '1.346', 'grad_norm': '12.56', 'learning_rate': '3.175e-06', 'num_tokens': '2.615e+06', 'mean_token_accuracy': '0.7059', 'epoch': '0.6827'}
+{'loss': '1.116', 'grad_norm': '17.25', 'learning_rate': '3.173e-06', 'num_tokens': '2.615e+06', 'mean_token_accuracy': '0.746', 'epoch': '0.6829'}
+{'loss': '1.125', 'grad_norm': '14.31', 'learning_rate': '3.171e-06', 'num_tokens': '2.616e+06', 'mean_token_accuracy': '0.6825', 'epoch': '0.683'}
+{'loss': '1.96', 'grad_norm': '35.25', 'learning_rate': '3.17e-06', 'num_tokens': '2.616e+06', 'mean_token_accuracy': '0.6522', 'epoch': '0.6832'}
+{'loss': '1.175', 'grad_norm': '12.5', 'learning_rate': '3.168e-06', 'num_tokens': '2.617e+06', 'mean_token_accuracy': '0.6844', 'epoch': '0.6834'}
+{'loss': '1.138', 'grad_norm': '14.38', 'learning_rate': '3.166e-06', 'num_tokens': '2.617e+06', 'mean_token_accuracy': '0.6758', 'epoch': '0.6836'}
+{'loss': '1.252', 'grad_norm': '15.44', 'learning_rate': '3.164e-06', 'num_tokens': '2.618e+06', 'mean_token_accuracy': '0.6505', 'epoch': '0.6837'}
+{'loss': '0.8067', 'grad_norm': '13.5', 'learning_rate': '3.163e-06', 'num_tokens': '2.619e+06', 'mean_token_accuracy': '0.7456', 'epoch': '0.6839'}
+{'loss': '1.377', 'grad_norm': '12.88', 'learning_rate': '3.161e-06', 'num_tokens': '2.619e+06', 'mean_token_accuracy': '0.6343', 'epoch': '0.6841'}
+{'loss': '1.948', 'grad_norm': '19.5', 'learning_rate': '3.159e-06', 'num_tokens': '2.62e+06', 'mean_token_accuracy': '0.6102', 'epoch': '0.6843'}
+
 68%|██████▊   | 3941/5758 [22:24<10:48,  2.80it/s]
 68%|██████▊   | 3942/5758 [22:24<11:01,  2.74it/s]
                                                   

 68%|██████▊   | 3942/5758 [22:24<11:01,  2.74it/s]
 68%|██████▊   | 3943/5758 [22:24<11:15,  2.69it/s]
                                                   

 68%|██████▊   | 3943/5758 [22:24<11:15,  2.69it/s]
 68%|██████▊   | 3944/5758 [22:25<10:52,  2.78it/s]
                                                   

 68%|██████▊   | 3944/5758 [22:25<10:52,  2.78it/s]
 69%|██████▊   | 3945/5758 [22:25<10:16,  2.94it/s]
                                                   

 69%|██████▊   | 3945/5758 [22:25<10:16,  2.94it/s]
 69%|██████▊   | 3946/5758 [22:25<10:13,  2.96it/s]
                                                   

 69%|██████▊   | 3946/5758 [22:25<10:13,  2.96it/s]
 69%|██████▊   | 3947/5758 [22:26<09:46,  3.09it/s]
                                                   

 69%|██████▊   | 3947/5758 [22:26<09:46,  3.09it/s]
 69%|██████▊   | 3948/5758 [22:26<09:39,  3.12it/s]
                                                   

 69%|██████▊   | 3948/5758 [22:26<09:39,  3.12it/s]
 69%|██████▊   | 3949/5758 [22:26<10:12,  2.95it/s]
                                                   

 69%|██████▊   | 3949/5758 [22:26<10:12,  2.95it/s]
 69%|██████▊   | 3950/5758 [22:27<11:15,  2.68it/s]
                                                   

 69%|██████▊   | 3950/5758 [22:27<11:15,  2.68it/s]
 69%|██████▊   | 3951/5758 [22:27<10:51,  2.78it/s]
                                                   

 69%|██████▊   | 3951/5758 [22:27<10:51,  2.78it/s]
 69%|██████▊   | 3952/5758 [22:27<10:16,  2.93it/s]
                                                   

 69%|██████▊   | 3952/5758 [22:27<10:16,  2.93it/s]
 69%|██████▊   | 3953/5758 [22:28<10:26,  2.88it/s]
                                                   

 69%|██████▊   | 3953/5758 [22:28<10:26,  2.88it/s]
 69%|██████▊   | 3954/5758 [22:28<10:36,  2.84it/s]
                                                   

 69%|██████▊   | 3954/5758 [22:28<10:36,  2.84it/s]
 69%|██████▊   | 3955/5758 [22:28<10:06,  2.97it/s]
                                                   

 69%|██████▊   | 3955/5758 [22:28<10:06,  2.97it/s]
 69%|██████▊   | 3956/5758 [22:29<09:52,  3.04it/s]
                                                   

 69%|██████▊   | 3956/5758 [22:29<09:52,  3.04it/s]
 69%|██████▊   | 3957/5758 [22:29<10:19,  2.91it/s]
                                                   

 69%|██████▊   | 3957/5758 [22:29<10:19,  2.91it/s]
 69%|██████▊   | 3958/5758 [22:29<10:10,  2.95it/s]
                                                   

 69%|██████▊   | 3958/5758 [22:29<10:10,  2.95it/s]
 69%|██████▉   | 3959/5758 [22:30<09:57,  3.01it/s]
                                                   

 69%|██████▉   | 3959/5758 [22:30<09:57,  3.01it/s]
 69%|██████▉   | 3960/5758 [22:30<09:40,  3.10it/s]
                                                   

 69%|██████▉   | 3960/5758 [22:30<09:40,  3.10it/s]
 69%|██████▉   | 3961/5758 [22:30<09:43,  3.08it/s]
                                                   

 69%|██████▉   | 3961/5758 [22:30<09:43,  3.08it/s]
 69%|██████▉   | 3962/5758 [22:31<09:57,  3.01it/s]
                                                   

 69%|██████▉   | 3962/5758 [22:31<09:57,  3.01it/s]
 69%|██████▉   | 3963/5758 [22:31<10:04,  2.97it/s]
                                                   

 69%|██████▉   | 3963/5758 [22:31<10:04,  2.97it/s]
 69%|██████▉   | 3964/5758 [22:31<10:24,  2.87it/s]
                                                   

 69%|██████▉   | 3964/5758 [22:31<10:24,  2.87it/s]
 69%|██████▉   | 3965/5758 [22:32<10:06,  2.96it/s]
                                                   

 69%|██████▉   | 3965/5758 [22:32<10:06,  2.96it/s]
 69%|██████▉   | 3966/5758 [22:32<09:52,  3.02it/s]
                                                   

 69%|██████▉   | 3966/5758 [22:32<09:52,  3.02it/s]
 69%|██████▉   | 3967/5758 [22:32<09:29,  3.15it/s]
                                                   

 69%|██████▉   | 3967/5758 [22:32<09:29,  3.15it/s]
 69%|██████▉   | 3968/5758 [22:33<09:13,  3.24it/s]
                                                   

 69%|██████▉   | 3968/5758 [22:33<09:13,  3.24it/s]
 69%|██████▉   | 3969/5758 [22:33<09:12,  3.24it/s]
                                                   

 69%|██████▉   | 3969/5758 [22:33<09:12,  3.24it/s]
 69%|██████▉   | 3970/5758 [22:33<09:44,  3.06it/s]
                                                   

 69%|██████▉   | 3970/5758 [22:33<09:44,  3.06it/s]
 69%|██████▉   | 3971/5758 [22:34<09:43,  3.06it/s]
                                                   

 69%|██████▉   | 3971/5758 [22:34<09:43,  3.06it/s]
 69%|██████▉   | 3972/5758 [22:34<10:04,  2.95it/s]
                                                   

 69%|██████▉   | 3972/5758 [22:34<10:04,  2.95it/s]
 69%|██████▉   | 3973/5758 [22:34<09:41,  3.07it/s]
                                                   

 69%|██████▉   | 3973/5758 [22:34<09:41,  3.07it/s]
 69%|██████▉   | 3974/5758 [22:35<09:25,  3.15it/s]
                                                   

 69%|██████▉   | 3974/5758 [22:35<09:25,  3.15it/s]
 69%|██████▉   | 3975/5758 [22:35<10:02,  2.96it/s]
                                                   

 69%|██████▉   | 3975/5758 [22:35<10:02,  2.96it/s]
 69%|██████▉   | 3976/5758 [22:36<11:22,  2.61it/s]
                                                   

 69%|██████▉   | 3976/5758 [22:36<11:22,  2.61it/s]
 69%|██████▉   | 3977/5758 [22:36<10:42,  2.77it/s]
                                                   

 69%|██████▉   | 3977/5758 [22:36<10:42,  2.77it/s]
 69%|██████▉   | 3978/5758 [22:36<10:09,  2.92it/s]
                                                   

 69%|██████▉   | 3978/5758 [22:36<10:09,  2.92it/s]
 69%|██████▉   | 3979/5758 [22:36<10:30,  2.82it/s]
                                                   

 69%|██████▉   | 3979/5758 [22:37<10:30,  2.82it/s]
 69%|██████▉   | 3980/5758 [22:37<10:46,  2.75it/s]
                                                   

 69%|██████▉   | 3980/5758 [22:37<10:46,  2.75it/s]
 69%|██████▉   | 3981/5758 [22:37<10:23,  2.85it/s]
                                                   

 69%|██████▉   | 3981/5758 [22:37<10:23,  2.85it/s]
 69%|██████▉   | 3982/5758 [22:38<09:56,  2.98it/s]
                                                   

 69%|██████▉   | 3982/5758 [22:38<09:56,  2.98it/s]
 69%|██████▉   | 3983/5758 [22:38<10:04,  2.94it/s]
                                                   

 69%|██████▉   | 3983/5758 [22:38<10:04,  2.94it/s]
 69%|██████▉   | 3984/5758 [22:38<09:44,  3.04it/s]
                                                   

 69%|██████▉   | 3984/5758 [22:38<09:44,  3.04it/s]
 69%|██████▉   | 3985/5758 [22:39<10:08,  2.91it/s]
                                                   

 69%|██████▉   | 3985/5758 [22:39<10:08,  2.91it/s]
 69%|██████▉   | 3986/5758 [22:39<09:45,  3.03it/s]
                                                   

 69%|██████▉   | 3986/5758 [22:39<09:45,  3.03it/s]
 69%|██████▉   | 3987/5758 [22:39<10:03,  2.93it/s]
                                                   

 69%|██████▉   | 3987/5758 [22:39<10:03,  2.93it/s]
 69%|██████▉   | 3988/5758 [22:40<10:03,  2.94it/s]
                                                   

 69%|██████▉   | 3988/5758 [22:40<10:03,  2.94it/s]
 69%|██████▉   | 3989/5758 [22:40<10:14,  2.88it/s]
                                                   

 69%|██████▉   | 3989/5758 [22:40<10:14,  2.88it/s]
 69%|██████▉   | 3990/5758 [22:40<09:53,  2.98it/s]
                                                   

 69%|██████▉   | 3990/5758 [22:40<09:53,  2.98it/s]
 69%|██████▉   | 3991/5758 [22:41<09:56,  2.96it/s]
                                                   

 69%|██████▉   | 3991/5758 [22:41<09:56,  2.96it/s]
 69%|██████▉   | 3992/5758 [22:41<09:50,  2.99it/s]
                                                   

 69%|██████▉   | 3992/5758 [22:41<09:50,  2.99it/s]
 69%|██████▉   | 3993/5758 [22:41<09:28,  3.11it/s]
                                                   

 69%|██████▉   | 3993/5758 [22:41<09:28,  3.11it/s]
 69%|██████▉   | 3994/5758 [22:42<09:55,  2.96it/s]
                                                   

 69%|██████▉   | 3994/5758 [22:42<09:55,  2.96it/s]
 69%|██████▉   | 3995/5758 [22:42<09:27,  3.11it/s]
                                                   

 69%|██████▉   | 3995/5758 [22:42<09:27,  3.11it/s]
 69%|██████▉   | 3996/5758 [22:42<09:19,  3.15it/s]
                                                   
{'loss': '1.541', 'grad_norm': '17.38', 'learning_rate': '3.157e-06', 'num_tokens': '2.62e+06', 'mean_token_accuracy': '0.7167', 'epoch': '0.6844'}
+{'loss': '1.498', 'grad_norm': '15.19', 'learning_rate': '3.156e-06', 'num_tokens': '2.621e+06', 'mean_token_accuracy': '0.6708', 'epoch': '0.6846'}
+{'loss': '0.6611', 'grad_norm': '11.81', 'learning_rate': '3.154e-06', 'num_tokens': '2.621e+06', 'mean_token_accuracy': '0.839', 'epoch': '0.6848'}
+{'loss': '1.36', 'grad_norm': '11.56', 'learning_rate': '3.152e-06', 'num_tokens': '2.622e+06', 'mean_token_accuracy': '0.6655', 'epoch': '0.685'}
+{'loss': '1.433', 'grad_norm': '15', 'learning_rate': '3.15e-06', 'num_tokens': '2.623e+06', 'mean_token_accuracy': '0.6292', 'epoch': '0.6851'}
+{'loss': '0.839', 'grad_norm': '14.31', 'learning_rate': '3.149e-06', 'num_tokens': '2.623e+06', 'mean_token_accuracy': '0.7846', 'epoch': '0.6853'}
+{'loss': '1.715', 'grad_norm': '16.5', 'learning_rate': '3.147e-06', 'num_tokens': '2.624e+06', 'mean_token_accuracy': '0.627', 'epoch': '0.6855'}
+{'loss': '1.276', 'grad_norm': '12.12', 'learning_rate': '3.145e-06', 'num_tokens': '2.624e+06', 'mean_token_accuracy': '0.6751', 'epoch': '0.6857'}
+{'loss': '0.9176', 'grad_norm': '17.38', 'learning_rate': '3.143e-06', 'num_tokens': '2.625e+06', 'mean_token_accuracy': '0.7708', 'epoch': '0.6858'}
+{'loss': '0.978', 'grad_norm': '19.12', 'learning_rate': '3.142e-06', 'num_tokens': '2.625e+06', 'mean_token_accuracy': '0.7876', 'epoch': '0.686'}
+{'loss': '0.992', 'grad_norm': '12.38', 'learning_rate': '3.14e-06', 'num_tokens': '2.626e+06', 'mean_token_accuracy': '0.7069', 'epoch': '0.6862'}
+{'loss': '1.265', 'grad_norm': '13.06', 'learning_rate': '3.138e-06', 'num_tokens': '2.626e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.6863'}
+{'loss': '1.049', 'grad_norm': '18.62', 'learning_rate': '3.137e-06', 'num_tokens': '2.627e+06', 'mean_token_accuracy': '0.7909', 'epoch': '0.6865'}
+{'loss': '0.6918', 'grad_norm': '11.44', 'learning_rate': '3.135e-06', 'num_tokens': '2.627e+06', 'mean_token_accuracy': '0.8271', 'epoch': '0.6867'}
+{'loss': '0.9587', 'grad_norm': '12.38', 'learning_rate': '3.133e-06', 'num_tokens': '2.628e+06', 'mean_token_accuracy': '0.7066', 'epoch': '0.6869'}
+{'loss': '1.176', 'grad_norm': '11.94', 'learning_rate': '3.131e-06', 'num_tokens': '2.628e+06', 'mean_token_accuracy': '0.6797', 'epoch': '0.687'}
+{'loss': '1.545', 'grad_norm': '21.25', 'learning_rate': '3.13e-06', 'num_tokens': '2.629e+06', 'mean_token_accuracy': '0.6864', 'epoch': '0.6872'}
+{'loss': '0.9447', 'grad_norm': '7.312', 'learning_rate': '3.128e-06', 'num_tokens': '2.63e+06', 'mean_token_accuracy': '0.7721', 'epoch': '0.6874'}
+{'loss': '1.126', 'grad_norm': '12.25', 'learning_rate': '3.126e-06', 'num_tokens': '2.631e+06', 'mean_token_accuracy': '0.6684', 'epoch': '0.6876'}
+{'loss': '1.095', 'grad_norm': '14.06', 'learning_rate': '3.124e-06', 'num_tokens': '2.632e+06', 'mean_token_accuracy': '0.6937', 'epoch': '0.6877'}
+{'loss': '1.374', 'grad_norm': '12.06', 'learning_rate': '3.123e-06', 'num_tokens': '2.633e+06', 'mean_token_accuracy': '0.6265', 'epoch': '0.6879'}
+{'loss': '1.249', 'grad_norm': '11', 'learning_rate': '3.121e-06', 'num_tokens': '2.634e+06', 'mean_token_accuracy': '0.6691', 'epoch': '0.6881'}
+{'loss': '1.534', 'grad_norm': '44.5', 'learning_rate': '3.119e-06', 'num_tokens': '2.634e+06', 'mean_token_accuracy': '0.6981', 'epoch': '0.6883'}
+{'loss': '1.009', 'grad_norm': '13.25', 'learning_rate': '3.117e-06', 'num_tokens': '2.634e+06', 'mean_token_accuracy': '0.6949', 'epoch': '0.6884'}
+{'loss': '0.8578', 'grad_norm': '10', 'learning_rate': '3.116e-06', 'num_tokens': '2.635e+06', 'mean_token_accuracy': '0.7834', 'epoch': '0.6886'}
+{'loss': '1.536', 'grad_norm': '16', 'learning_rate': '3.114e-06', 'num_tokens': '2.636e+06', 'mean_token_accuracy': '0.6757', 'epoch': '0.6888'}
+{'loss': '1.443', 'grad_norm': '11.88', 'learning_rate': '3.112e-06', 'num_tokens': '2.636e+06', 'mean_token_accuracy': '0.6976', 'epoch': '0.689'}
+{'loss': '0.6603', 'grad_norm': '8.188', 'learning_rate': '3.11e-06', 'num_tokens': '2.637e+06', 'mean_token_accuracy': '0.8055', 'epoch': '0.6891'}
+{'loss': '1.043', 'grad_norm': '22.5', 'learning_rate': '3.109e-06', 'num_tokens': '2.638e+06', 'mean_token_accuracy': '0.7473', 'epoch': '0.6893'}
+{'loss': '1.144', 'grad_norm': '13.62', 'learning_rate': '3.107e-06', 'num_tokens': '2.638e+06', 'mean_token_accuracy': '0.6866', 'epoch': '0.6895'}
+{'loss': '1.425', 'grad_norm': '26.38', 'learning_rate': '3.105e-06', 'num_tokens': '2.638e+06', 'mean_token_accuracy': '0.7391', 'epoch': '0.6896'}
+{'loss': '1.566', 'grad_norm': '13.31', 'learning_rate': '3.104e-06', 'num_tokens': '2.639e+06', 'mean_token_accuracy': '0.6788', 'epoch': '0.6898'}
+{'loss': '1.587', 'grad_norm': '14.75', 'learning_rate': '3.102e-06', 'num_tokens': '2.639e+06', 'mean_token_accuracy': '0.628', 'epoch': '0.69'}
+{'loss': '1.69', 'grad_norm': '16', 'learning_rate': '3.1e-06', 'num_tokens': '2.64e+06', 'mean_token_accuracy': '0.6435', 'epoch': '0.6902'}
+{'loss': '1.205', 'grad_norm': '13.88', 'learning_rate': '3.098e-06', 'num_tokens': '2.64e+06', 'mean_token_accuracy': '0.6678', 'epoch': '0.6903'}
+{'loss': '2.077', 'grad_norm': '19.75', 'learning_rate': '3.097e-06', 'num_tokens': '2.641e+06', 'mean_token_accuracy': '0.5712', 'epoch': '0.6905'}
+{'loss': '1.279', 'grad_norm': '14.88', 'learning_rate': '3.095e-06', 'num_tokens': '2.642e+06', 'mean_token_accuracy': '0.6572', 'epoch': '0.6907'}
+{'loss': '1.301', 'grad_norm': '24.25', 'learning_rate': '3.093e-06', 'num_tokens': '2.642e+06', 'mean_token_accuracy': '0.7654', 'epoch': '0.6909'}
+{'loss': '1.477', 'grad_norm': '30.88', 'learning_rate': '3.091e-06', 'num_tokens': '2.642e+06', 'mean_token_accuracy': '0.7402', 'epoch': '0.691'}
+{'loss': '1.131', 'grad_norm': '14.81', 'learning_rate': '3.09e-06', 'num_tokens': '2.643e+06', 'mean_token_accuracy': '0.6812', 'epoch': '0.6912'}
+{'loss': '1.165', 'grad_norm': '12.12', 'learning_rate': '3.088e-06', 'num_tokens': '2.643e+06', 'mean_token_accuracy': '0.6872', 'epoch': '0.6914'}
+{'loss': '1.865', 'grad_norm': '21.5', 'learning_rate': '3.086e-06', 'num_tokens': '2.644e+06', 'mean_token_accuracy': '0.5909', 'epoch': '0.6916'}
+{'loss': '1.395', 'grad_norm': '12.56', 'learning_rate': '3.084e-06', 'num_tokens': '2.645e+06', 'mean_token_accuracy': '0.6345', 'epoch': '0.6917'}
+{'loss': '1.071', 'grad_norm': '12.5', 'learning_rate': '3.083e-06', 'num_tokens': '2.645e+06', 'mean_token_accuracy': '0.7062', 'epoch': '0.6919'}
+{'loss': '0.7737', 'grad_norm': '21.88', 'learning_rate': '3.081e-06', 'num_tokens': '2.646e+06', 'mean_token_accuracy': '0.7964', 'epoch': '0.6921'}
+{'loss': '0.9872', 'grad_norm': '11.69', 'learning_rate': '3.079e-06', 'num_tokens': '2.646e+06', 'mean_token_accuracy': '0.7215', 'epoch': '0.6923'}
+{'loss': '1.777', 'grad_norm': '20.62', 'learning_rate': '3.077e-06', 'num_tokens': '2.647e+06', 'mean_token_accuracy': '0.6161', 'epoch': '0.6924'}
+{'loss': '0.9715', 'grad_norm': '10.88', 'learning_rate': '3.076e-06', 'num_tokens': '2.648e+06', 'mean_token_accuracy': '0.725', 'epoch': '0.6926'}
+{'loss': '2.359', 'grad_norm': '11.88', 'learning_rate': '3.074e-06', 'num_tokens': '2.649e+06', 'mean_token_accuracy': '0.5455', 'epoch': '0.6928'}
+{'loss': '0.7246', 'grad_norm': '23', 'learning_rate': '3.072e-06', 'num_tokens': '2.649e+06', 'mean_token_accuracy': '0.8011', 'epoch': '0.6929'}
+{'loss': '1.156', 'grad_norm': '10.56', 'learning_rate': '3.071e-06', 'num_tokens': '2.65e+06', 'mean_token_accuracy': '0.6722', 'epoch': '0.6931'}
+{'loss': '1.703', 'grad_norm': '13.31', 'learning_rate': '3.069e-06', 'num_tokens': '2.651e+06', 'mean_token_accuracy': '0.5875', 'epoch': '0.6933'}
+{'loss': '1.19', 'grad_norm': '12.88', 'learning_rate': '3.067e-06', 'num_tokens': '2.651e+06', 'mean_token_accuracy': '0.7417', 'epoch': '0.6935'}
+{'loss': '1.724', 'grad_norm': '39.25', 'learning_rate': '3.065e-06', 'num_tokens': '2.651e+06', 'mean_token_accuracy': '0.624', 'epoch': '0.6936'}
+{'loss': '0.4978', 'grad_norm': '7.812', 'learning_rate': '3.064e-06', 'num_tokens': '2.652e+06', 'mean_token_accuracy': '0.8404', 'epoch': '0.6938'}
+
 69%|██████▉   | 3996/5758 [22:42<09:19,  3.15it/s]
 69%|██████▉   | 3997/5758 [22:43<09:41,  3.03it/s]
                                                   

 69%|██████▉   | 3997/5758 [22:43<09:41,  3.03it/s]
 69%|██████▉   | 3998/5758 [22:43<09:25,  3.11it/s]
                                                   

 69%|██████▉   | 3998/5758 [22:43<09:25,  3.11it/s]
 69%|██████▉   | 3999/5758 [22:43<09:23,  3.12it/s]
                                                   

 69%|██████▉   | 3999/5758 [22:43<09:23,  3.12it/s]
 69%|██████▉   | 4000/5758 [22:43<09:17,  3.16it/s]
                                                   

 69%|██████▉   | 4000/5758 [22:43<09:17,  3.16it/s]{'loss': '2.665', 'grad_norm': '20.5', 'learning_rate': '3.062e-06', 'num_tokens': '2.653e+06', 'mean_token_accuracy': '0.4993', 'epoch': '0.694'}
+{'loss': '1.336', 'grad_norm': '14.38', 'learning_rate': '3.06e-06', 'num_tokens': '2.654e+06', 'mean_token_accuracy': '0.7072', 'epoch': '0.6942'}
+{'loss': '1.243', 'grad_norm': '13.5', 'learning_rate': '3.058e-06', 'num_tokens': '2.654e+06', 'mean_token_accuracy': '0.6657', 'epoch': '0.6943'}
+{'loss': '1.663', 'grad_norm': '24.75', 'learning_rate': '3.057e-06', 'num_tokens': '2.655e+06', 'mean_token_accuracy': '0.6277', 'epoch': '0.6945'}
+{'loss': '1.047', 'grad_norm': '11.25', 'learning_rate': '3.055e-06', 'num_tokens': '2.655e+06', 'mean_token_accuracy': '0.7766', 'epoch': '0.6947'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.61it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.61it/s]
+
 69%|██████▉   | 4001/5758 [22:46<25:29,  1.15it/s]
                                                   

 69%|██████▉   | 4001/5758 [22:46<25:29,  1.15it/s]
 70%|██████▉   | 4002/5758 [22:46<20:49,  1.41it/s]
                                                   

 70%|██████▉   | 4002/5758 [22:46<20:49,  1.41it/s]
 70%|██████▉   | 4003/5758 [22:46<17:36,  1.66it/s]
                                                   

 70%|██████▉   | 4003/5758 [22:46<17:36,  1.66it/s]
 70%|██████▉   | 4004/5758 [22:47<15:43,  1.86it/s]
                                                   

 70%|██████▉   | 4004/5758 [22:47<15:43,  1.86it/s]
 70%|██████▉   | 4005/5758 [22:47<14:46,  1.98it/s]
                                                   

 70%|██████▉   | 4005/5758 [22:47<14:46,  1.98it/s]
 70%|██████▉   | 4006/5758 [22:47<13:05,  2.23it/s]
                                                   

 70%|██████▉   | 4006/5758 [22:47<13:05,  2.23it/s]
 70%|██████▉   | 4007/5758 [22:48<11:40,  2.50it/s]
                                                   

 70%|██████▉   | 4007/5758 [22:48<11:40,  2.50it/s]
 70%|██████▉   | 4008/5758 [22:48<11:05,  2.63it/s]
                                                   

 70%|██████▉   | 4008/5758 [22:48<11:05,  2.63it/s]
 70%|██████▉   | 4009/5758 [22:48<10:46,  2.70it/s]
                                                   

 70%|██████▉   | 4009/5758 [22:48<10:46,  2.70it/s]
 70%|██████▉   | 4010/5758 [22:49<10:15,  2.84it/s]
                                                   

 70%|██████▉   | 4010/5758 [22:49<10:15,  2.84it/s]
 70%|██████▉   | 4011/5758 [22:49<09:40,  3.01it/s]
                                                   

 70%|██████▉   | 4011/5758 [22:49<09:40,  3.01it/s]
 70%|██████▉   | 4012/5758 [22:49<10:06,  2.88it/s]
                                                   

 70%|██████▉   | 4012/5758 [22:49<10:06,  2.88it/s]
 70%|██████▉   | 4013/5758 [22:50<09:37,  3.02it/s]
                                                   

 70%|██████▉   | 4013/5758 [22:50<09:37,  3.02it/s]
 70%|██████▉   | 4014/5758 [22:50<09:21,  3.11it/s]
                                                   

 70%|██████▉   | 4014/5758 [22:50<09:21,  3.11it/s]
 70%|██████▉   | 4015/5758 [22:50<09:45,  2.98it/s]
                                                   

 70%|██████▉   | 4015/5758 [22:50<09:45,  2.98it/s]
 70%|██████▉   | 4016/5758 [22:51<09:26,  3.08it/s]
                                                   

 70%|██████▉   | 4016/5758 [22:51<09:26,  3.08it/s]
 70%|██████▉   | 4017/5758 [22:51<09:12,  3.15it/s]
                                                   

 70%|██████▉   | 4017/5758 [22:51<09:12,  3.15it/s]
 70%|██████▉   | 4018/5758 [22:51<09:10,  3.16it/s]
                                                   

 70%|██████▉   | 4018/5758 [22:51<09:10,  3.16it/s]
 70%|██████▉   | 4019/5758 [22:52<09:50,  2.95it/s]
                                                   

 70%|██████▉   | 4019/5758 [22:52<09:50,  2.95it/s]
 70%|██████▉   | 4020/5758 [22:52<09:32,  3.03it/s]
                                                   

 70%|██████▉   | 4020/5758 [22:52<09:32,  3.03it/s]
 70%|██████▉   | 4021/5758 [22:52<10:20,  2.80it/s]
                                                   

 70%|██████▉   | 4021/5758 [22:52<10:20,  2.80it/s]
 70%|██████▉   | 4022/5758 [22:53<10:41,  2.70it/s]
                                                   

 70%|██████▉   | 4022/5758 [22:53<10:41,  2.70it/s]
 70%|██████▉   | 4023/5758 [22:53<10:39,  2.71it/s]
                                                   

 70%|██████▉   | 4023/5758 [22:53<10:39,  2.71it/s]
 70%|██████▉   | 4024/5758 [22:53<10:05,  2.86it/s]
                                                   

 70%|██████▉   | 4024/5758 [22:53<10:05,  2.86it/s]
 70%|██████▉   | 4025/5758 [22:54<10:21,  2.79it/s]
                                                   

 70%|██████▉   | 4025/5758 [22:54<10:21,  2.79it/s]
 70%|██████▉   | 4026/5758 [22:54<10:03,  2.87it/s]
                                                   

 70%|██████▉   | 4026/5758 [22:54<10:03,  2.87it/s]
 70%|██████▉   | 4027/5758 [22:55<10:16,  2.81it/s]
                                                   

 70%|██████▉   | 4027/5758 [22:55<10:16,  2.81it/s]
 70%|██████▉   | 4028/5758 [22:55<11:15,  2.56it/s]
                                                   

 70%|██████▉   | 4028/5758 [22:55<11:15,  2.56it/s]
 70%|██████▉   | 4029/5758 [22:55<11:36,  2.48it/s]
                                                   

 70%|██████▉   | 4029/5758 [22:55<11:36,  2.48it/s]
 70%|██████▉   | 4030/5758 [22:56<13:09,  2.19it/s]
                                                   

 70%|██████▉   | 4030/5758 [22:56<13:09,  2.19it/s]
 70%|███████   | 4031/5758 [22:56<12:41,  2.27it/s]
                                                   

 70%|███████   | 4031/5758 [22:56<12:41,  2.27it/s]
 70%|███████   | 4032/5758 [22:57<11:42,  2.46it/s]
                                                   

 70%|███████   | 4032/5758 [22:57<11:42,  2.46it/s]
 70%|███████   | 4033/5758 [22:57<11:12,  2.57it/s]
                                                   

 70%|███████   | 4033/5758 [22:57<11:12,  2.57it/s]
 70%|███████   | 4034/5758 [22:57<10:55,  2.63it/s]
                                                   

 70%|███████   | 4034/5758 [22:57<10:55,  2.63it/s]
 70%|███████   | 4035/5758 [22:58<10:19,  2.78it/s]
                                                   

 70%|███████   | 4035/5758 [22:58<10:19,  2.78it/s]
 70%|███████   | 4036/5758 [22:58<10:17,  2.79it/s]
                                                   

 70%|███████   | 4036/5758 [22:58<10:17,  2.79it/s]
 70%|███████   | 4037/5758 [22:58<10:15,  2.80it/s]
                                                   

 70%|███████   | 4037/5758 [22:58<10:15,  2.80it/s]
 70%|███████   | 4038/5758 [22:59<10:02,  2.86it/s]
                                                   

 70%|███████   | 4038/5758 [22:59<10:02,  2.86it/s]
 70%|███████   | 4039/5758 [22:59<09:38,  2.97it/s]
                                                   

 70%|███████   | 4039/5758 [22:59<09:38,  2.97it/s]
 70%|███████   | 4040/5758 [22:59<09:41,  2.95it/s]
                                                   

 70%|███████   | 4040/5758 [22:59<09:41,  2.95it/s]
 70%|███████   | 4041/5758 [23:00<09:50,  2.91it/s]
                                                   

 70%|███████   | 4041/5758 [23:00<09:50,  2.91it/s]
 70%|███████   | 4042/5758 [23:00<10:14,  2.79it/s]
                                                   

 70%|███████   | 4042/5758 [23:00<10:14,  2.79it/s]
 70%|███████   | 4043/5758 [23:01<10:12,  2.80it/s]
                                                   

 70%|███████   | 4043/5758 [23:01<10:12,  2.80it/s]
 70%|███████   | 4044/5758 [23:01<09:55,  2.88it/s]
                                                   

 70%|███████   | 4044/5758 [23:01<09:55,  2.88it/s]
 70%|███████   | 4045/5758 [23:01<10:02,  2.85it/s]
                                                   

 70%|███████   | 4045/5758 [23:01<10:02,  2.85it/s]
 70%|███████   | 4046/5758 [23:02<10:12,  2.80it/s]
                                                   

 70%|███████   | 4046/5758 [23:02<10:12,  2.80it/s]
 70%|███████   | 4047/5758 [23:02<10:41,  2.67it/s]
                                                   

 70%|███████   | 4047/5758 [23:02<10:41,  2.67it/s]
 70%|███████   | 4048/5758 [23:02<10:23,  2.74it/s]
                                                   

 70%|███████   | 4048/5758 [23:02<10:23,  2.74it/s]
 70%|███████   | 4049/5758 [23:03<10:18,  2.76it/s]
                                                   

 70%|███████   | 4049/5758 [23:03<10:18,  2.76it/s]
 70%|███████   | 4050/5758 [23:03<09:55,  2.87it/s]
                                                   

 70%|███████   | 4050/5758 [23:03<09:55,  2.87it/s]
 70%|███████   | 4051/5758 [23:03<09:34,  2.97it/s]
                                                   

 70%|███████   | 4051/5758 [23:03<09:34,  2.97it/s]
 70%|███████   | 4052/5758 [23:04<09:10,  3.10it/s]
                                                   

 70%|███████   | 4052/5758 [23:04<09:10,  3.10it/s]
 70%|███████   | 4053/5758 [23:04<09:11,  3.09it/s]
                                                   

 70%|███████   | 4053/5758 [23:04<09:11,  3.09it/s]
 70%|███████   | 4054/5758 [23:04<08:56,  3.17it/s]
                                                   

 70%|███████   | 4054/5758 [23:04<08:56,  3.17it/s]
 70%|███████   | 4055/5758 [23:05<09:02,  3.14it/s]
                                                   

 70%|███████   | 4055/5758 [23:05<09:02,  3.14it/s]
 70%|███████   | 4056/5758 [23:05<09:57,  2.85it/s]
                                                   
{'loss': '1.036', 'grad_norm': '13.44', 'learning_rate': '3.053e-06', 'num_tokens': '2.656e+06', 'mean_token_accuracy': '0.7128', 'epoch': '0.6949'}
+{'loss': '2.343', 'grad_norm': '18.5', 'learning_rate': '3.051e-06', 'num_tokens': '2.656e+06', 'mean_token_accuracy': '0.507', 'epoch': '0.695'}
+{'loss': '1.006', 'grad_norm': '16.25', 'learning_rate': '3.05e-06', 'num_tokens': '2.657e+06', 'mean_token_accuracy': '0.6832', 'epoch': '0.6952'}
+{'loss': '1.123', 'grad_norm': '13.69', 'learning_rate': '3.048e-06', 'num_tokens': '2.657e+06', 'mean_token_accuracy': '0.6986', 'epoch': '0.6954'}
+{'loss': '0.8892', 'grad_norm': '23.62', 'learning_rate': '3.046e-06', 'num_tokens': '2.658e+06', 'mean_token_accuracy': '0.8', 'epoch': '0.6956'}
+{'loss': '1.21', 'grad_norm': '13.06', 'learning_rate': '3.044e-06', 'num_tokens': '2.658e+06', 'mean_token_accuracy': '0.7276', 'epoch': '0.6957'}
+{'loss': '1.092', 'grad_norm': '13.31', 'learning_rate': '3.043e-06', 'num_tokens': '2.659e+06', 'mean_token_accuracy': '0.715', 'epoch': '0.6959'}
+{'loss': '0.9429', 'grad_norm': '20', 'learning_rate': '3.041e-06', 'num_tokens': '2.659e+06', 'mean_token_accuracy': '0.8018', 'epoch': '0.6961'}
+{'loss': '1.316', 'grad_norm': '14.06', 'learning_rate': '3.039e-06', 'num_tokens': '2.66e+06', 'mean_token_accuracy': '0.7463', 'epoch': '0.6962'}
+{'loss': '1.169', 'grad_norm': '12.81', 'learning_rate': '3.038e-06', 'num_tokens': '2.66e+06', 'mean_token_accuracy': '0.6726', 'epoch': '0.6964'}
+{'loss': '1.152', 'grad_norm': '13.31', 'learning_rate': '3.036e-06', 'num_tokens': '2.661e+06', 'mean_token_accuracy': '0.6976', 'epoch': '0.6966'}
+{'loss': '1.088', 'grad_norm': '14.06', 'learning_rate': '3.034e-06', 'num_tokens': '2.662e+06', 'mean_token_accuracy': '0.6858', 'epoch': '0.6968'}
+{'loss': '1.238', 'grad_norm': '13.81', 'learning_rate': '3.032e-06', 'num_tokens': '2.662e+06', 'mean_token_accuracy': '0.6497', 'epoch': '0.6969'}
+{'loss': '0.7424', 'grad_norm': '8.75', 'learning_rate': '3.031e-06', 'num_tokens': '2.664e+06', 'mean_token_accuracy': '0.7691', 'epoch': '0.6971'}
+{'loss': '0.5936', 'grad_norm': '27.88', 'learning_rate': '3.029e-06', 'num_tokens': '2.664e+06', 'mean_token_accuracy': '0.8291', 'epoch': '0.6973'}
+{'loss': '1.288', 'grad_norm': '13.94', 'learning_rate': '3.027e-06', 'num_tokens': '2.664e+06', 'mean_token_accuracy': '0.6575', 'epoch': '0.6975'}
+{'loss': '0.6436', 'grad_norm': '47.75', 'learning_rate': '3.025e-06', 'num_tokens': '2.664e+06', 'mean_token_accuracy': '0.8493', 'epoch': '0.6976'}
+{'loss': '1.243', 'grad_norm': '13.19', 'learning_rate': '3.024e-06', 'num_tokens': '2.665e+06', 'mean_token_accuracy': '0.6495', 'epoch': '0.6978'}
+{'loss': '1.735', 'grad_norm': '15', 'learning_rate': '3.022e-06', 'num_tokens': '2.666e+06', 'mean_token_accuracy': '0.643', 'epoch': '0.698'}
+{'loss': '1.269', 'grad_norm': '13.81', 'learning_rate': '3.02e-06', 'num_tokens': '2.666e+06', 'mean_token_accuracy': '0.6578', 'epoch': '0.6982'}
+{'loss': '0.8908', 'grad_norm': '22.38', 'learning_rate': '3.018e-06', 'num_tokens': '2.666e+06', 'mean_token_accuracy': '0.7832', 'epoch': '0.6983'}
+{'loss': '0.8174', 'grad_norm': '14.62', 'learning_rate': '3.017e-06', 'num_tokens': '2.667e+06', 'mean_token_accuracy': '0.8275', 'epoch': '0.6985'}
+{'loss': '1.266', 'grad_norm': '11.94', 'learning_rate': '3.015e-06', 'num_tokens': '2.668e+06', 'mean_token_accuracy': '0.6421', 'epoch': '0.6987'}
+{'loss': '0.4628', 'grad_norm': '7.188', 'learning_rate': '3.013e-06', 'num_tokens': '2.669e+06', 'mean_token_accuracy': '0.8609', 'epoch': '0.6989'}
+{'loss': '1.164', 'grad_norm': '13.94', 'learning_rate': '3.011e-06', 'num_tokens': '2.669e+06', 'mean_token_accuracy': '0.6862', 'epoch': '0.699'}
+{'loss': '0.6699', 'grad_norm': '15.81', 'learning_rate': '3.01e-06', 'num_tokens': '2.67e+06', 'mean_token_accuracy': '0.8529', 'epoch': '0.6992'}
+{'loss': '1.003', 'grad_norm': '17.88', 'learning_rate': '3.008e-06', 'num_tokens': '2.67e+06', 'mean_token_accuracy': '0.7118', 'epoch': '0.6994'}
+{'loss': '1.376', 'grad_norm': '28', 'learning_rate': '3.006e-06', 'num_tokens': '2.67e+06', 'mean_token_accuracy': '0.6929', 'epoch': '0.6995'}
+{'loss': '1.649', 'grad_norm': '15.5', 'learning_rate': '3.005e-06', 'num_tokens': '2.671e+06', 'mean_token_accuracy': '0.6567', 'epoch': '0.6997'}
+{'loss': '1.969', 'grad_norm': '13.31', 'learning_rate': '3.003e-06', 'num_tokens': '2.671e+06', 'mean_token_accuracy': '0.6248', 'epoch': '0.6999'}
+{'loss': '1.574', 'grad_norm': '21.5', 'learning_rate': '3.001e-06', 'num_tokens': '2.672e+06', 'mean_token_accuracy': '0.6364', 'epoch': '0.7001'}
+{'loss': '1.402', 'grad_norm': '46.5', 'learning_rate': '2.999e-06', 'num_tokens': '2.672e+06', 'mean_token_accuracy': '0.6813', 'epoch': '0.7002'}
+{'loss': '0.5639', 'grad_norm': '6.812', 'learning_rate': '2.998e-06', 'num_tokens': '2.674e+06', 'mean_token_accuracy': '0.83', 'epoch': '0.7004'}
+{'loss': '1.566', 'grad_norm': '11.31', 'learning_rate': '2.996e-06', 'num_tokens': '2.675e+06', 'mean_token_accuracy': '0.666', 'epoch': '0.7006'}
+{'loss': '0.6102', 'grad_norm': '7.688', 'learning_rate': '2.994e-06', 'num_tokens': '2.676e+06', 'mean_token_accuracy': '0.7973', 'epoch': '0.7008'}
+{'loss': '0.7367', 'grad_norm': '24.12', 'learning_rate': '2.992e-06', 'num_tokens': '2.676e+06', 'mean_token_accuracy': '0.8596', 'epoch': '0.7009'}
+{'loss': '1.924', 'grad_norm': '19.75', 'learning_rate': '2.991e-06', 'num_tokens': '2.677e+06', 'mean_token_accuracy': '0.5947', 'epoch': '0.7011'}
+{'loss': '1.372', 'grad_norm': '12.12', 'learning_rate': '2.989e-06', 'num_tokens': '2.677e+06', 'mean_token_accuracy': '0.6135', 'epoch': '0.7013'}
+{'loss': '1.934', 'grad_norm': '11', 'learning_rate': '2.987e-06', 'num_tokens': '2.678e+06', 'mean_token_accuracy': '0.5931', 'epoch': '0.7015'}
+{'loss': '0.9128', 'grad_norm': '17.62', 'learning_rate': '2.985e-06', 'num_tokens': '2.678e+06', 'mean_token_accuracy': '0.808', 'epoch': '0.7016'}
+{'loss': '1.66', 'grad_norm': '13.12', 'learning_rate': '2.984e-06', 'num_tokens': '2.679e+06', 'mean_token_accuracy': '0.6484', 'epoch': '0.7018'}
+{'loss': '0.4571', 'grad_norm': '22.12', 'learning_rate': '2.982e-06', 'num_tokens': '2.679e+06', 'mean_token_accuracy': '0.8815', 'epoch': '0.702'}
+{'loss': '1.175', 'grad_norm': '11.38', 'learning_rate': '2.98e-06', 'num_tokens': '2.68e+06', 'mean_token_accuracy': '0.668', 'epoch': '0.7022'}
+{'loss': '0.8141', 'grad_norm': '21.62', 'learning_rate': '2.978e-06', 'num_tokens': '2.68e+06', 'mean_token_accuracy': '0.8088', 'epoch': '0.7023'}
+{'loss': '1.655', 'grad_norm': '19.5', 'learning_rate': '2.977e-06', 'num_tokens': '2.681e+06', 'mean_token_accuracy': '0.6693', 'epoch': '0.7025'}
+{'loss': '2.009', 'grad_norm': '17.38', 'learning_rate': '2.975e-06', 'num_tokens': '2.681e+06', 'mean_token_accuracy': '0.574', 'epoch': '0.7027'}
+{'loss': '0.6609', 'grad_norm': '22.5', 'learning_rate': '2.973e-06', 'num_tokens': '2.681e+06', 'mean_token_accuracy': '0.8562', 'epoch': '0.7028'}
+{'loss': '0.5429', 'grad_norm': '7.031', 'learning_rate': '2.972e-06', 'num_tokens': '2.683e+06', 'mean_token_accuracy': '0.8252', 'epoch': '0.703'}
+{'loss': '1.204', 'grad_norm': '10.88', 'learning_rate': '2.97e-06', 'num_tokens': '2.684e+06', 'mean_token_accuracy': '0.7157', 'epoch': '0.7032'}
+{'loss': '1.633', 'grad_norm': '22.88', 'learning_rate': '2.968e-06', 'num_tokens': '2.684e+06', 'mean_token_accuracy': '0.6848', 'epoch': '0.7034'}
+{'loss': '1.286', 'grad_norm': '29.5', 'learning_rate': '2.966e-06', 'num_tokens': '2.684e+06', 'mean_token_accuracy': '0.7759', 'epoch': '0.7035'}
+{'loss': '1.059', 'grad_norm': '12.69', 'learning_rate': '2.965e-06', 'num_tokens': '2.685e+06', 'mean_token_accuracy': '0.7147', 'epoch': '0.7037'}
+{'loss': '1.122', 'grad_norm': '11.19', 'learning_rate': '2.963e-06', 'num_tokens': '2.686e+06', 'mean_token_accuracy': '0.678', 'epoch': '0.7039'}
+{'loss': '2.089', 'grad_norm': '17.88', 'learning_rate': '2.961e-06', 'num_tokens': '2.686e+06', 'mean_token_accuracy': '0.6288', 'epoch': '0.7041'}
+{'loss': '0.8647', 'grad_norm': '13.5', 'learning_rate': '2.959e-06', 'num_tokens': '2.686e+06', 'mean_token_accuracy': '0.7866', 'epoch': '0.7042'}
+
 70%|███████   | 4056/5758 [23:05<09:57,  2.85it/s]
 70%|███████   | 4057/5758 [23:05<09:41,  2.92it/s]
                                                   

 70%|███████   | 4057/5758 [23:05<09:41,  2.92it/s]
 70%|███████   | 4058/5758 [23:06<09:16,  3.05it/s]
                                                   

 70%|███████   | 4058/5758 [23:06<09:16,  3.05it/s]
 70%|███████   | 4059/5758 [23:06<09:15,  3.06it/s]
                                                   

 70%|███████   | 4059/5758 [23:06<09:15,  3.06it/s]
 71%|███████   | 4060/5758 [23:06<08:48,  3.21it/s]
                                                   

 71%|███████   | 4060/5758 [23:06<08:48,  3.21it/s]
 71%|███████   | 4061/5758 [23:07<08:47,  3.22it/s]
                                                   

 71%|███████   | 4061/5758 [23:07<08:47,  3.22it/s]
 71%|███████   | 4062/5758 [23:07<08:43,  3.24it/s]
                                                   

 71%|███████   | 4062/5758 [23:07<08:43,  3.24it/s]
 71%|███████   | 4063/5758 [23:07<09:35,  2.95it/s]
                                                   

 71%|███████   | 4063/5758 [23:07<09:35,  2.95it/s]
 71%|███████   | 4064/5758 [23:08<10:28,  2.70it/s]
                                                   

 71%|███████   | 4064/5758 [23:08<10:28,  2.70it/s]
 71%|███████   | 4065/5758 [23:08<10:15,  2.75it/s]
                                                   

 71%|███████   | 4065/5758 [23:08<10:15,  2.75it/s]
 71%|███████   | 4066/5758 [23:08<09:42,  2.90it/s]
                                                   

 71%|███████   | 4066/5758 [23:08<09:42,  2.90it/s]
 71%|███████   | 4067/5758 [23:09<09:18,  3.03it/s]
                                                   

 71%|███████   | 4067/5758 [23:09<09:18,  3.03it/s]
 71%|███████   | 4068/5758 [23:09<09:10,  3.07it/s]
                                                   

 71%|███████   | 4068/5758 [23:09<09:10,  3.07it/s]
 71%|███████   | 4069/5758 [23:09<08:59,  3.13it/s]
                                                   

 71%|███████   | 4069/5758 [23:09<08:59,  3.13it/s]
 71%|███████   | 4070/5758 [23:10<09:31,  2.95it/s]
                                                   

 71%|███████   | 4070/5758 [23:10<09:31,  2.95it/s]
 71%|███████   | 4071/5758 [23:10<09:12,  3.05it/s]
                                                   

 71%|███████   | 4071/5758 [23:10<09:12,  3.05it/s]
 71%|███████   | 4072/5758 [23:10<09:02,  3.11it/s]
                                                   

 71%|███████   | 4072/5758 [23:10<09:02,  3.11it/s]
 71%|███████   | 4073/5758 [23:11<08:55,  3.14it/s]
                                                   

 71%|███████   | 4073/5758 [23:11<08:55,  3.14it/s]
 71%|███████   | 4074/5758 [23:11<09:04,  3.09it/s]
                                                   

 71%|███████   | 4074/5758 [23:11<09:04,  3.09it/s]
 71%|███████   | 4075/5758 [23:11<09:18,  3.01it/s]
                                                   

 71%|███████   | 4075/5758 [23:11<09:18,  3.01it/s]
 71%|███████   | 4076/5758 [23:12<09:52,  2.84it/s]
                                                   

 71%|███████   | 4076/5758 [23:12<09:52,  2.84it/s]
 71%|███████   | 4077/5758 [23:12<09:32,  2.94it/s]
                                                   

 71%|███████   | 4077/5758 [23:12<09:32,  2.94it/s]
 71%|███████   | 4078/5758 [23:12<09:27,  2.96it/s]
                                                   

 71%|███████   | 4078/5758 [23:12<09:27,  2.96it/s]
 71%|███████   | 4079/5758 [23:13<09:03,  3.09it/s]
                                                   

 71%|███████   | 4079/5758 [23:13<09:03,  3.09it/s]
 71%|███████   | 4080/5758 [23:13<10:31,  2.66it/s]
                                                   

 71%|███████   | 4080/5758 [23:13<10:31,  2.66it/s]
 71%|███████   | 4081/5758 [23:13<09:49,  2.84it/s]
                                                   

 71%|███████   | 4081/5758 [23:13<09:49,  2.84it/s]
 71%|███████   | 4082/5758 [23:14<09:27,  2.96it/s]
                                                   

 71%|███████   | 4082/5758 [23:14<09:27,  2.96it/s]
 71%|███████   | 4083/5758 [23:14<09:13,  3.02it/s]
                                                   

 71%|███████   | 4083/5758 [23:14<09:13,  3.02it/s]
 71%|███████   | 4084/5758 [23:14<08:59,  3.10it/s]
                                                   

 71%|███████   | 4084/5758 [23:14<08:59,  3.10it/s]
 71%|███████   | 4085/5758 [23:15<08:56,  3.12it/s]
                                                   

 71%|███████   | 4085/5758 [23:15<08:56,  3.12it/s]
 71%|███████   | 4086/5758 [23:15<08:46,  3.18it/s]
                                                   

 71%|███████   | 4086/5758 [23:15<08:46,  3.18it/s]
 71%|███████   | 4087/5758 [23:15<09:10,  3.03it/s]
                                                   

 71%|███████   | 4087/5758 [23:15<09:10,  3.03it/s]
 71%|███████   | 4088/5758 [23:16<08:59,  3.09it/s]
                                                   

 71%|███████   | 4088/5758 [23:16<08:59,  3.09it/s]
 71%|███████   | 4089/5758 [23:16<09:44,  2.86it/s]
                                                   

 71%|███████   | 4089/5758 [23:16<09:44,  2.86it/s]
 71%|███████   | 4090/5758 [23:16<09:17,  2.99it/s]
                                                   

 71%|███████   | 4090/5758 [23:16<09:17,  2.99it/s]
 71%|███████   | 4091/5758 [23:17<08:59,  3.09it/s]
                                                   

 71%|███████   | 4091/5758 [23:17<08:59,  3.09it/s]
 71%|███████   | 4092/5758 [23:17<08:57,  3.10it/s]
                                                   

 71%|███████   | 4092/5758 [23:17<08:57,  3.10it/s]
 71%|███████   | 4093/5758 [23:17<08:55,  3.11it/s]
                                                   

 71%|███████   | 4093/5758 [23:17<08:55,  3.11it/s]
 71%|███████   | 4094/5758 [23:18<08:50,  3.14it/s]
                                                   

 71%|███████   | 4094/5758 [23:18<08:50,  3.14it/s]
 71%|███████   | 4095/5758 [23:18<08:32,  3.24it/s]
                                                   

 71%|███████   | 4095/5758 [23:18<08:32,  3.24it/s]
 71%|███████   | 4096/5758 [23:18<08:26,  3.28it/s]
                                                   

 71%|███████   | 4096/5758 [23:18<08:26,  3.28it/s]
 71%|███████   | 4097/5758 [23:18<08:31,  3.24it/s]
                                                   

 71%|███████   | 4097/5758 [23:18<08:31,  3.24it/s]
 71%|███████   | 4098/5758 [23:19<08:37,  3.21it/s]
                                                   

 71%|███████   | 4098/5758 [23:19<08:37,  3.21it/s]
 71%|███████   | 4099/5758 [23:19<09:03,  3.05it/s]
                                                   

 71%|███████   | 4099/5758 [23:19<09:03,  3.05it/s]
 71%|███████   | 4100/5758 [23:19<09:03,  3.05it/s]
                                                   

 71%|███████   | 4100/5758 [23:19<09:03,  3.05it/s]
 71%|███████   | 4101/5758 [23:20<09:25,  2.93it/s]
                                                   

 71%|███████   | 4101/5758 [23:20<09:25,  2.93it/s]
 71%|███████   | 4102/5758 [23:20<09:53,  2.79it/s]
                                                   

 71%|███████   | 4102/5758 [23:20<09:53,  2.79it/s]
 71%|███████▏  | 4103/5758 [23:21<10:09,  2.72it/s]
                                                   

 71%|███████▏  | 4103/5758 [23:21<10:09,  2.72it/s]
 71%|███████▏  | 4104/5758 [23:21<09:37,  2.86it/s]
                                                   

 71%|███████▏  | 4104/5758 [23:21<09:37,  2.86it/s]
 71%|███████▏  | 4105/5758 [23:21<09:16,  2.97it/s]
                                                   

 71%|███████▏  | 4105/5758 [23:21<09:16,  2.97it/s]
 71%|█████���█▏  | 4106/5758 [23:22<09:53,  2.78it/s]
                                                   

 71%|███████▏  | 4106/5758 [23:22<09:53,  2.78it/s]
 71%|███████▏  | 4107/5758 [23:22<10:22,  2.65it/s]
                                                   

 71%|███████▏  | 4107/5758 [23:22<10:22,  2.65it/s]
 71%|███████▏  | 4108/5758 [23:22<10:16,  2.68it/s]
                                                   

 71%|███████▏  | 4108/5758 [23:22<10:16,  2.68it/s]
 71%|███████▏  | 4109/5758 [23:23<10:25,  2.64it/s]
                                                   

 71%|███████▏  | 4109/5758 [23:23<10:25,  2.64it/s]
 71%|███████▏  | 4110/5758 [23:23<10:17,  2.67it/s]
                                                   

 71%|███████▏  | 4110/5758 [23:23<10:17,  2.67it/s]
 71%|███████▏  | 4111/5758 [23:24<10:19,  2.66it/s]
                                                   
{'loss': '1.154', 'grad_norm': '15.38', 'learning_rate': '2.958e-06', 'num_tokens': '2.687e+06', 'mean_token_accuracy': '0.7222', 'epoch': '0.7044'}
+{'loss': '1.107', 'grad_norm': '14', 'learning_rate': '2.956e-06', 'num_tokens': '2.688e+06', 'mean_token_accuracy': '0.6903', 'epoch': '0.7046'}
+{'loss': '1.279', 'grad_norm': '13.5', 'learning_rate': '2.954e-06', 'num_tokens': '2.689e+06', 'mean_token_accuracy': '0.6501', 'epoch': '0.7048'}
+{'loss': '1.648', 'grad_norm': '29.75', 'learning_rate': '2.952e-06', 'num_tokens': '2.689e+06', 'mean_token_accuracy': '0.6573', 'epoch': '0.7049'}
+{'loss': '1.11', 'grad_norm': '47.75', 'learning_rate': '2.951e-06', 'num_tokens': '2.689e+06', 'mean_token_accuracy': '0.8033', 'epoch': '0.7051'}
+{'loss': '1.543', 'grad_norm': '13.31', 'learning_rate': '2.949e-06', 'num_tokens': '2.69e+06', 'mean_token_accuracy': '0.6056', 'epoch': '0.7053'}
+{'loss': '1.17', 'grad_norm': '13.44', 'learning_rate': '2.947e-06', 'num_tokens': '2.69e+06', 'mean_token_accuracy': '0.7079', 'epoch': '0.7055'}
+{'loss': '1.06', 'grad_norm': '28', 'learning_rate': '2.945e-06', 'num_tokens': '2.69e+06', 'mean_token_accuracy': '0.7973', 'epoch': '0.7056'}
+{'loss': '1.236', 'grad_norm': '5.25', 'learning_rate': '2.944e-06', 'num_tokens': '2.694e+06', 'mean_token_accuracy': '0.7043', 'epoch': '0.7058'}
+{'loss': '0.5955', 'grad_norm': '6.969', 'learning_rate': '2.942e-06', 'num_tokens': '2.696e+06', 'mean_token_accuracy': '0.8143', 'epoch': '0.706'}
+{'loss': '0.6479', 'grad_norm': '7.844', 'learning_rate': '2.94e-06', 'num_tokens': '2.698e+06', 'mean_token_accuracy': '0.8099', 'epoch': '0.7061'}
+{'loss': '1.304', 'grad_norm': '13.19', 'learning_rate': '2.939e-06', 'num_tokens': '2.698e+06', 'mean_token_accuracy': '0.6259', 'epoch': '0.7063'}
+{'loss': '1.053', 'grad_norm': '12.62', 'learning_rate': '2.937e-06', 'num_tokens': '2.699e+06', 'mean_token_accuracy': '0.6854', 'epoch': '0.7065'}
+{'loss': '2.248', 'grad_norm': '15.88', 'learning_rate': '2.935e-06', 'num_tokens': '2.7e+06', 'mean_token_accuracy': '0.5567', 'epoch': '0.7067'}
+{'loss': '1.189', 'grad_norm': '14.19', 'learning_rate': '2.933e-06', 'num_tokens': '2.7e+06', 'mean_token_accuracy': '0.6881', 'epoch': '0.7068'}
+{'loss': '0.5729', 'grad_norm': '7.344', 'learning_rate': '2.932e-06', 'num_tokens': '2.702e+06', 'mean_token_accuracy': '0.8249', 'epoch': '0.707'}
+{'loss': '1.537', 'grad_norm': '8.438', 'learning_rate': '2.93e-06', 'num_tokens': '2.703e+06', 'mean_token_accuracy': '0.6422', 'epoch': '0.7072'}
+{'loss': '0.9684', 'grad_norm': '20', 'learning_rate': '2.928e-06', 'num_tokens': '2.703e+06', 'mean_token_accuracy': '0.8436', 'epoch': '0.7074'}
+{'loss': '1.549', 'grad_norm': '12.12', 'learning_rate': '2.926e-06', 'num_tokens': '2.704e+06', 'mean_token_accuracy': '0.5854', 'epoch': '0.7075'}
+{'loss': '0.5036', 'grad_norm': '6.75', 'learning_rate': '2.925e-06', 'num_tokens': '2.706e+06', 'mean_token_accuracy': '0.8464', 'epoch': '0.7077'}
+{'loss': '1.291', 'grad_norm': '42.5', 'learning_rate': '2.923e-06', 'num_tokens': '2.706e+06', 'mean_token_accuracy': '0.7426', 'epoch': '0.7079'}
+{'loss': '1.985', 'grad_norm': '16.62', 'learning_rate': '2.921e-06', 'num_tokens': '2.707e+06', 'mean_token_accuracy': '0.5981', 'epoch': '0.7081'}
+{'loss': '0.9684', 'grad_norm': '21', 'learning_rate': '2.919e-06', 'num_tokens': '2.707e+06', 'mean_token_accuracy': '0.7696', 'epoch': '0.7082'}
+{'loss': '1.392', 'grad_norm': '49.5', 'learning_rate': '2.918e-06', 'num_tokens': '2.707e+06', 'mean_token_accuracy': '0.6324', 'epoch': '0.7084'}
+{'loss': '1.568', 'grad_norm': '19', 'learning_rate': '2.916e-06', 'num_tokens': '2.708e+06', 'mean_token_accuracy': '0.6517', 'epoch': '0.7086'}
+{'loss': '1.246', 'grad_norm': '13.94', 'learning_rate': '2.914e-06', 'num_tokens': '2.708e+06', 'mean_token_accuracy': '0.6792', 'epoch': '0.7088'}
+{'loss': '1.197', 'grad_norm': '12.44', 'learning_rate': '2.912e-06', 'num_tokens': '2.709e+06', 'mean_token_accuracy': '0.689', 'epoch': '0.7089'}
+{'loss': '1.207', 'grad_norm': '12.44', 'learning_rate': '2.911e-06', 'num_tokens': '2.71e+06', 'mean_token_accuracy': '0.6684', 'epoch': '0.7091'}
+{'loss': '1.037', 'grad_norm': '12.44', 'learning_rate': '2.909e-06', 'num_tokens': '2.71e+06', 'mean_token_accuracy': '0.7064', 'epoch': '0.7093'}
+{'loss': '1.068', 'grad_norm': '36', 'learning_rate': '2.907e-06', 'num_tokens': '2.711e+06', 'mean_token_accuracy': '0.7857', 'epoch': '0.7094'}
+{'loss': '1.045', 'grad_norm': '11.88', 'learning_rate': '2.906e-06', 'num_tokens': '2.711e+06', 'mean_token_accuracy': '0.7007', 'epoch': '0.7096'}
+{'loss': '1.724', 'grad_norm': '22', 'learning_rate': '2.904e-06', 'num_tokens': '2.712e+06', 'mean_token_accuracy': '0.6385', 'epoch': '0.7098'}
+{'loss': '1.201', 'grad_norm': '24.25', 'learning_rate': '2.902e-06', 'num_tokens': '2.712e+06', 'mean_token_accuracy': '0.7457', 'epoch': '0.71'}
+{'loss': '0.7274', 'grad_norm': '25', 'learning_rate': '2.9e-06', 'num_tokens': '2.712e+06', 'mean_token_accuracy': '0.8243', 'epoch': '0.7101'}
+{'loss': '0.8777', 'grad_norm': '11.81', 'learning_rate': '2.899e-06', 'num_tokens': '2.713e+06', 'mean_token_accuracy': '0.7955', 'epoch': '0.7103'}
+{'loss': '0.8631', 'grad_norm': '10.56', 'learning_rate': '2.897e-06', 'num_tokens': '2.713e+06', 'mean_token_accuracy': '0.7744', 'epoch': '0.7105'}
+{'loss': '2.348', 'grad_norm': '49', 'learning_rate': '2.895e-06', 'num_tokens': '2.713e+06', 'mean_token_accuracy': '0.5806', 'epoch': '0.7107'}
+{'loss': '0.5294', 'grad_norm': '13.62', 'learning_rate': '2.893e-06', 'num_tokens': '2.714e+06', 'mean_token_accuracy': '0.855', 'epoch': '0.7108'}
+{'loss': '0.8925', 'grad_norm': '10.44', 'learning_rate': '2.892e-06', 'num_tokens': '2.714e+06', 'mean_token_accuracy': '0.7647', 'epoch': '0.711'}
+{'loss': '2.37', 'grad_norm': '38.75', 'learning_rate': '2.89e-06', 'num_tokens': '2.715e+06', 'mean_token_accuracy': '0.5867', 'epoch': '0.7112'}
+{'loss': '1.032', 'grad_norm': '12.81', 'learning_rate': '2.888e-06', 'num_tokens': '2.715e+06', 'mean_token_accuracy': '0.721', 'epoch': '0.7114'}
+{'loss': '2.827', 'grad_norm': '32', 'learning_rate': '2.886e-06', 'num_tokens': '2.715e+06', 'mean_token_accuracy': '0.5842', 'epoch': '0.7115'}
+{'loss': '1.256', 'grad_norm': '15.25', 'learning_rate': '2.885e-06', 'num_tokens': '2.716e+06', 'mean_token_accuracy': '0.7342', 'epoch': '0.7117'}
+{'loss': '1.107', 'grad_norm': '10', 'learning_rate': '2.883e-06', 'num_tokens': '2.717e+06', 'mean_token_accuracy': '0.677', 'epoch': '0.7119'}
+{'loss': '1.159', 'grad_norm': '12.44', 'learning_rate': '2.881e-06', 'num_tokens': '2.718e+06', 'mean_token_accuracy': '0.6771', 'epoch': '0.7121'}
+{'loss': '1.744', 'grad_norm': '17.62', 'learning_rate': '2.879e-06', 'num_tokens': '2.718e+06', 'mean_token_accuracy': '0.6222', 'epoch': '0.7122'}
+{'loss': '1.238', 'grad_norm': '24.62', 'learning_rate': '2.878e-06', 'num_tokens': '2.718e+06', 'mean_token_accuracy': '0.734', 'epoch': '0.7124'}
+{'loss': '1.182', 'grad_norm': '13.81', 'learning_rate': '2.876e-06', 'num_tokens': '2.719e+06', 'mean_token_accuracy': '0.6661', 'epoch': '0.7126'}
+{'loss': '1.785', 'grad_norm': '8.312', 'learning_rate': '2.874e-06', 'num_tokens': '2.72e+06', 'mean_token_accuracy': '0.6078', 'epoch': '0.7127'}
+{'loss': '1.802', 'grad_norm': '16.5', 'learning_rate': '2.873e-06', 'num_tokens': '2.721e+06', 'mean_token_accuracy': '0.6282', 'epoch': '0.7129'}
+{'loss': '1.807', 'grad_norm': '34.5', 'learning_rate': '2.871e-06', 'num_tokens': '2.721e+06', 'mean_token_accuracy': '0.6986', 'epoch': '0.7131'}
+{'loss': '1.309', 'grad_norm': '15.5', 'learning_rate': '2.869e-06', 'num_tokens': '2.722e+06', 'mean_token_accuracy': '0.6391', 'epoch': '0.7133'}
+{'loss': '1.895', 'grad_norm': '16.62', 'learning_rate': '2.867e-06', 'num_tokens': '2.722e+06', 'mean_token_accuracy': '0.6179', 'epoch': '0.7134'}
+{'loss': '1.058', 'grad_norm': '13.31', 'learning_rate': '2.866e-06', 'num_tokens': '2.723e+06', 'mean_token_accuracy': '0.7083', 'epoch': '0.7136'}
+{'loss': '1.634', 'grad_norm': '20.25', 'learning_rate': '2.864e-06', 'num_tokens': '2.723e+06', 'mean_token_accuracy': '0.6486', 'epoch': '0.7138'}
+
 71%|███████▏  | 4111/5758 [23:24<10:19,  2.66it/s]
 71%|███████▏  | 4112/5758 [23:24<10:21,  2.65it/s]
                                                   

 71%|███████▏  | 4112/5758 [23:24<10:21,  2.65it/s]
 71%|███████▏  | 4113/5758 [23:24<10:09,  2.70it/s]
                                                   

 71%|███████▏  | 4113/5758 [23:24<10:09,  2.70it/s]
 71%|███████▏  | 4114/5758 [23:25<10:05,  2.72it/s]
                                                   

 71%|███████▏  | 4114/5758 [23:25<10:05,  2.72it/s]
 71%|███████▏  | 4115/5758 [23:25<09:37,  2.84it/s]
                                                   

 71%|███████▏  | 4115/5758 [23:25<09:37,  2.84it/s]
 71%|███████▏  | 4116/5758 [23:25<09:14,  2.96it/s]
                                                   

 71%|███████▏  | 4116/5758 [23:25<09:14,  2.96it/s]
 72%|███████▏  | 4117/5758 [23:26<09:26,  2.90it/s]
                                                   

 72%|███████▏  | 4117/5758 [23:26<09:26,  2.90it/s]
 72%|███████▏  | 4118/5758 [23:26<09:01,  3.03it/s]
                                                   

 72%|███████▏  | 4118/5758 [23:26<09:01,  3.03it/s]
 72%|███████▏  | 4119/5758 [23:26<08:52,  3.08it/s]
                                                   

 72%|███████▏  | 4119/5758 [23:26<08:52,  3.08it/s]
 72%|███████▏  | 4120/5758 [23:27<08:54,  3.06it/s]
                                                   

 72%|███████▏  | 4120/5758 [23:27<08:54,  3.06it/s]
 72%|███████▏  | 4121/5758 [23:27<08:40,  3.14it/s]
                                                   

 72%|███████▏  | 4121/5758 [23:27<08:40,  3.14it/s]
 72%|███████▏  | 4122/5758 [23:27<09:03,  3.01it/s]
                                                   

 72%|███████▏  | 4122/5758 [23:27<09:03,  3.01it/s]
 72%|███████▏  | 4123/5758 [23:28<09:10,  2.97it/s]
                                                   

 72%|███████▏  | 4123/5758 [23:28<09:10,  2.97it/s]
 72%|███████▏  | 4124/5758 [23:28<10:10,  2.68it/s]
                                                   

 72%|███████▏  | 4124/5758 [23:28<10:10,  2.68it/s]
 72%|███████▏  | 4125/5758 [23:28<09:53,  2.75it/s]
                                                   

 72%|███████▏  | 4125/5758 [23:28<09:53,  2.75it/s]
 72%|███████▏  | 4126/5758 [23:29<09:59,  2.72it/s]
                                                   

 72%|███████▏  | 4126/5758 [23:29<09:59,  2.72it/s]
 72%|███████▏  | 4127/5758 [23:29<10:23,  2.62it/s]
                                                   

 72%|███████▏  | 4127/5758 [23:29<10:23,  2.62it/s]
 72%|███████▏  | 4128/5758 [23:30<10:06,  2.69it/s]
                                                   

 72%|███████▏  | 4128/5758 [23:30<10:06,  2.69it/s]
 72%|███████▏  | 4129/5758 [23:30<10:08,  2.68it/s]
                                                   

 72%|███████▏  | 4129/5758 [23:30<10:08,  2.68it/s]
 72%|███████▏  | 4130/5758 [23:30<10:03,  2.70it/s]
                                                   

 72%|███████▏  | 4130/5758 [23:30<10:03,  2.70it/s]
 72%|███████▏  | 4131/5758 [23:31<10:36,  2.56it/s]
                                                   

 72%|███████▏  | 4131/5758 [23:31<10:36,  2.56it/s]
 72%|███████▏  | 4132/5758 [23:31<10:34,  2.56it/s]
                                                   

 72%|███████▏  | 4132/5758 [23:31<10:34,  2.56it/s]
 72%|███████▏  | 4133/5758 [23:31<10:26,  2.59it/s]
                                                   

 72%|███████▏  | 4133/5758 [23:31<10:26,  2.59it/s]
 72%|███████▏  | 4134/5758 [23:32<10:17,  2.63it/s]
                                                   

 72%|███████▏  | 4134/5758 [23:32<10:17,  2.63it/s]
 72%|███████▏  | 4135/5758 [23:32<11:59,  2.26it/s]
                                                   

 72%|███████▏  | 4135/5758 [23:32<11:59,  2.26it/s]
 72%|███████▏  | 4136/5758 [23:33<11:47,  2.29it/s]
                                                   

 72%|███████▏  | 4136/5758 [23:33<11:47,  2.29it/s]
 72%|███████▏  | 4137/5758 [23:33<11:35,  2.33it/s]
                                                   

 72%|███████▏  | 4137/5758 [23:33<11:35,  2.33it/s]
 72%|███████▏  | 4138/5758 [23:34<10:31,  2.56it/s]
                                                   

 72%|███████▏  | 4138/5758 [23:34<10:31,  2.56it/s]
 72%|███████▏  | 4139/5758 [23:34<09:55,  2.72it/s]
                                                   

 72%|███████▏  | 4139/5758 [23:34<09:55,  2.72it/s]
 72%|███████▏  | 4140/5758 [23:34<09:39,  2.79it/s]
                                                   

 72%|███████▏  | 4140/5758 [23:34<09:39,  2.79it/s]
 72%|███████▏  | 4141/5758 [23:35<09:36,  2.80it/s]
                                                   

 72%|███████▏  | 4141/5758 [23:35<09:36,  2.80it/s]
 72%|███████▏  | 4142/5758 [23:35<09:19,  2.89it/s]
                                                   

 72%|███████▏  | 4142/5758 [23:35<09:19,  2.89it/s]
 72%|███████▏  | 4143/5758 [23:35<09:36,  2.80it/s]
                                                   

 72%|███████▏  | 4143/5758 [23:35<09:36,  2.80it/s]
 72%|███████▏  | 4144/5758 [23:36<09:53,  2.72it/s]
                                                   

 72%|███████▏  | 4144/5758 [23:36<09:53,  2.72it/s]
 72%|███████▏  | 4145/5758 [23:36<09:40,  2.78it/s]
                                                   

 72%|███████▏  | 4145/5758 [23:36<09:40,  2.78it/s]
 72%|███████▏  | 4146/5758 [23:36<09:28,  2.84it/s]
                                                   

 72%|███████▏  | 4146/5758 [23:36<09:28,  2.84it/s]
 72%|███████▏  | 4147/5758 [23:37<09:15,  2.90it/s]
                                                   

 72%|███████▏  | 4147/5758 [23:37<09:15,  2.90it/s]
 72%|███████▏  | 4148/5758 [23:37<09:20,  2.87it/s]
                                                   

 72%|███████▏  | 4148/5758 [23:37<09:20,  2.87it/s]
 72%|███████▏  | 4149/5758 [23:37<08:54,  3.01it/s]
                                                   

 72%|███████▏  | 4149/5758 [23:37<08:54,  3.01it/s]
 72%|███████▏  | 4150/5758 [23:38<08:42,  3.08it/s]
                                                   

 72%|███████▏  | 4150/5758 [23:38<08:42,  3.08it/s]
 72%|███████▏  | 4151/5758 [23:38<09:10,  2.92it/s]
                                                   

 72%|███████▏  | 4151/5758 [23:38<09:10,  2.92it/s]
 72%|███████▏  | 4152/5758 [23:38<09:11,  2.91it/s]
                                                   

 72%|███████▏  | 4152/5758 [23:38<09:11,  2.91it/s]
 72%|███████▏  | 4153/5758 [23:39<09:07,  2.93it/s]
                                                   

 72%|███████▏  | 4153/5758 [23:39<09:07,  2.93it/s]
 72%|███████▏  | 4154/5758 [23:39<09:24,  2.84it/s]
                                                   

 72%|███████▏  | 4154/5758 [23:39<09:24,  2.84it/s]
 72%|███████▏  | 4155/5758 [23:39<09:22,  2.85it/s]
                                                   

 72%|███████▏  | 4155/5758 [23:39<09:22,  2.85it/s]
 72%|███████▏  | 4156/5758 [23:40<09:03,  2.95it/s]
                                                   

 72%|███████▏  | 4156/5758 [23:40<09:03,  2.95it/s]
 72%|███████▏  | 4157/5758 [23:40<09:16,  2.88it/s]
                                                   

 72%|███████▏  | 4157/5758 [23:40<09:16,  2.88it/s]
 72%|███████▏  | 4158/5758 [23:40<08:50,  3.02it/s]
                                                   

 72%|███████▏  | 4158/5758 [23:40<08:50,  3.02it/s]
 72%|███████▏  | 4159/5758 [23:41<09:10,  2.91it/s]
                                                   

 72%|███████▏  | 4159/5758 [23:41<09:10,  2.91it/s]
 72%|███████▏  | 4160/5758 [23:41<09:03,  2.94it/s]
                                                   

 72%|███████▏  | 4160/5758 [23:41<09:03,  2.94it/s]
 72%|███████▏  | 4161/5758 [23:41<09:01,  2.95it/s]
                                                   

 72%|███████▏  | 4161/5758 [23:41<09:01,  2.95it/s]
 72%|███████▏  | 4162/5758 [23:42<08:47,  3.02it/s]
                                                   

 72%|███████▏  | 4162/5758 [23:42<08:47,  3.02it/s]
 72%|███████▏  | 4163/5758 [23:42<09:17,  2.86it/s]
                                                   

 72%|███████▏  | 4163/5758 [23:42<09:17,  2.86it/s]
 72%|███████▏  | 4164/5758 [23:42<08:50,  3.00it/s]
                                                   

 72%|███████▏  | 4164/5758 [23:42<08:50,  3.00it/s]
 72%|███████▏  | 4165/5758 [23:43<08:35,  3.09it/s]
                                                   

 72%|███████▏  | 4165/5758 [23:43<08:35,  3.09it/s]
 72%|███████▏  | 4166/5758 [23:43<08:28,  3.13it/s]
                                                   
{'loss': '1.178', 'grad_norm': '13.56', 'learning_rate': '2.862e-06', 'num_tokens': '2.723e+06', 'mean_token_accuracy': '0.6927', 'epoch': '0.714'}
+{'loss': '1.624', 'grad_norm': '12.94', 'learning_rate': '2.86e-06', 'num_tokens': '2.724e+06', 'mean_token_accuracy': '0.6696', 'epoch': '0.7141'}
+{'loss': '2.093', 'grad_norm': '24.62', 'learning_rate': '2.859e-06', 'num_tokens': '2.724e+06', 'mean_token_accuracy': '0.5975', 'epoch': '0.7143'}
+{'loss': '0.6102', 'grad_norm': '16.5', 'learning_rate': '2.857e-06', 'num_tokens': '2.724e+06', 'mean_token_accuracy': '0.877', 'epoch': '0.7145'}
+{'loss': '0.8245', 'grad_norm': '12.19', 'learning_rate': '2.855e-06', 'num_tokens': '2.725e+06', 'mean_token_accuracy': '0.7537', 'epoch': '0.7147'}
+{'loss': '0.9759', 'grad_norm': '11.75', 'learning_rate': '2.853e-06', 'num_tokens': '2.726e+06', 'mean_token_accuracy': '0.7449', 'epoch': '0.7148'}
+{'loss': '1.638', 'grad_norm': '18.12', 'learning_rate': '2.852e-06', 'num_tokens': '2.726e+06', 'mean_token_accuracy': '0.6548', 'epoch': '0.715'}
+{'loss': '1.389', 'grad_norm': '12.06', 'learning_rate': '2.85e-06', 'num_tokens': '2.727e+06', 'mean_token_accuracy': '0.7219', 'epoch': '0.7152'}
+{'loss': '2.246', 'grad_norm': '16.88', 'learning_rate': '2.848e-06', 'num_tokens': '2.727e+06', 'mean_token_accuracy': '0.5933', 'epoch': '0.7154'}
+{'loss': '1.379', 'grad_norm': '31.38', 'learning_rate': '2.846e-06', 'num_tokens': '2.728e+06', 'mean_token_accuracy': '0.6965', 'epoch': '0.7155'}
+{'loss': '1.316', 'grad_norm': '13.81', 'learning_rate': '2.845e-06', 'num_tokens': '2.728e+06', 'mean_token_accuracy': '0.6332', 'epoch': '0.7157'}
+{'loss': '0.9749', 'grad_norm': '19.88', 'learning_rate': '2.843e-06', 'num_tokens': '2.729e+06', 'mean_token_accuracy': '0.7817', 'epoch': '0.7159'}
+{'loss': '0.9874', 'grad_norm': '10.44', 'learning_rate': '2.841e-06', 'num_tokens': '2.729e+06', 'mean_token_accuracy': '0.7263', 'epoch': '0.716'}
+{'loss': '1.136', 'grad_norm': '38.5', 'learning_rate': '2.84e-06', 'num_tokens': '2.73e+06', 'mean_token_accuracy': '0.7328', 'epoch': '0.7162'}
+{'loss': '1.465', 'grad_norm': '56.25', 'learning_rate': '2.838e-06', 'num_tokens': '2.73e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.7164'}
+{'loss': '2.337', 'grad_norm': '15.06', 'learning_rate': '2.836e-06', 'num_tokens': '2.73e+06', 'mean_token_accuracy': '0.5216', 'epoch': '0.7166'}
+{'loss': '2.114', 'grad_norm': '17', 'learning_rate': '2.834e-06', 'num_tokens': '2.731e+06', 'mean_token_accuracy': '0.5733', 'epoch': '0.7167'}
+{'loss': '0.7954', 'grad_norm': '10.5', 'learning_rate': '2.833e-06', 'num_tokens': '2.731e+06', 'mean_token_accuracy': '0.789', 'epoch': '0.7169'}
+{'loss': '1.737', 'grad_norm': '42.25', 'learning_rate': '2.831e-06', 'num_tokens': '2.731e+06', 'mean_token_accuracy': '0.619', 'epoch': '0.7171'}
+{'loss': '0.6085', 'grad_norm': '8.188', 'learning_rate': '2.829e-06', 'num_tokens': '2.733e+06', 'mean_token_accuracy': '0.8082', 'epoch': '0.7173'}
+{'loss': '1.172', 'grad_norm': '10.75', 'learning_rate': '2.827e-06', 'num_tokens': '2.734e+06', 'mean_token_accuracy': '0.6715', 'epoch': '0.7174'}
+{'loss': '1.083', 'grad_norm': '7.031', 'learning_rate': '2.826e-06', 'num_tokens': '2.735e+06', 'mean_token_accuracy': '0.7369', 'epoch': '0.7176'}
+{'loss': '0.6039', 'grad_norm': '6.969', 'learning_rate': '2.824e-06', 'num_tokens': '2.737e+06', 'mean_token_accuracy': '0.8134', 'epoch': '0.7178'}
+{'loss': '1.184', 'grad_norm': '27', 'learning_rate': '2.822e-06', 'num_tokens': '2.737e+06', 'mean_token_accuracy': '0.7529', 'epoch': '0.718'}
+{'loss': '1.262', 'grad_norm': '6.469', 'learning_rate': '2.82e-06', 'num_tokens': '2.739e+06', 'mean_token_accuracy': '0.7034', 'epoch': '0.7181'}
+{'loss': '0.5949', 'grad_norm': '9.25', 'learning_rate': '2.819e-06', 'num_tokens': '2.74e+06', 'mean_token_accuracy': '0.8282', 'epoch': '0.7183'}
+{'loss': '0.7873', 'grad_norm': '13.94', 'learning_rate': '2.817e-06', 'num_tokens': '2.74e+06', 'mean_token_accuracy': '0.8311', 'epoch': '0.7185'}
+{'loss': '1.444', 'grad_norm': '44.5', 'learning_rate': '2.815e-06', 'num_tokens': '2.74e+06', 'mean_token_accuracy': '0.6986', 'epoch': '0.7187'}
+{'loss': '0.6647', 'grad_norm': '18.38', 'learning_rate': '2.813e-06', 'num_tokens': '2.741e+06', 'mean_token_accuracy': '0.8556', 'epoch': '0.7188'}
+{'loss': '0.9593', 'grad_norm': '10.62', 'learning_rate': '2.812e-06', 'num_tokens': '2.741e+06', 'mean_token_accuracy': '0.7128', 'epoch': '0.719'}
+{'loss': '0.8738', 'grad_norm': '20.5', 'learning_rate': '2.81e-06', 'num_tokens': '2.742e+06', 'mean_token_accuracy': '0.7759', 'epoch': '0.7192'}
+{'loss': '1.184', 'grad_norm': '11.56', 'learning_rate': '2.808e-06', 'num_tokens': '2.742e+06', 'mean_token_accuracy': '0.708', 'epoch': '0.7193'}
+{'loss': '1.668', 'grad_norm': '16.12', 'learning_rate': '2.807e-06', 'num_tokens': '2.743e+06', 'mean_token_accuracy': '0.632', 'epoch': '0.7195'}
+{'loss': '1.004', 'grad_norm': '27.62', 'learning_rate': '2.805e-06', 'num_tokens': '2.743e+06', 'mean_token_accuracy': '0.7664', 'epoch': '0.7197'}
+{'loss': '0.7153', 'grad_norm': '13', 'learning_rate': '2.803e-06', 'num_tokens': '2.744e+06', 'mean_token_accuracy': '0.8117', 'epoch': '0.7199'}
+{'loss': '1.297', 'grad_norm': '21.88', 'learning_rate': '2.801e-06', 'num_tokens': '2.744e+06', 'mean_token_accuracy': '0.7299', 'epoch': '0.72'}
+{'loss': '0.7474', 'grad_norm': '34.25', 'learning_rate': '2.8e-06', 'num_tokens': '2.744e+06', 'mean_token_accuracy': '0.7979', 'epoch': '0.7202'}
+{'loss': '1.223', 'grad_norm': '10.69', 'learning_rate': '2.798e-06', 'num_tokens': '2.745e+06', 'mean_token_accuracy': '0.6629', 'epoch': '0.7204'}
+{'loss': '1.679', 'grad_norm': '16.75', 'learning_rate': '2.796e-06', 'num_tokens': '2.745e+06', 'mean_token_accuracy': '0.6542', 'epoch': '0.7206'}
+{'loss': '0.6631', 'grad_norm': '8.438', 'learning_rate': '2.794e-06', 'num_tokens': '2.747e+06', 'mean_token_accuracy': '0.8061', 'epoch': '0.7207'}
+{'loss': '1.771', 'grad_norm': '15.25', 'learning_rate': '2.793e-06', 'num_tokens': '2.747e+06', 'mean_token_accuracy': '0.5951', 'epoch': '0.7209'}
+{'loss': '0.7856', 'grad_norm': '34', 'learning_rate': '2.791e-06', 'num_tokens': '2.747e+06', 'mean_token_accuracy': '0.7961', 'epoch': '0.7211'}
+{'loss': '0.6393', 'grad_norm': '7.156', 'learning_rate': '2.789e-06', 'num_tokens': '2.749e+06', 'mean_token_accuracy': '0.795', 'epoch': '0.7213'}
+{'loss': '1.811', 'grad_norm': '9.312', 'learning_rate': '2.787e-06', 'num_tokens': '2.75e+06', 'mean_token_accuracy': '0.6271', 'epoch': '0.7214'}
+{'loss': '1.788', 'grad_norm': '10.06', 'learning_rate': '2.786e-06', 'num_tokens': '2.751e+06', 'mean_token_accuracy': '0.61', 'epoch': '0.7216'}
+{'loss': '1.034', 'grad_norm': '11.06', 'learning_rate': '2.784e-06', 'num_tokens': '2.752e+06', 'mean_token_accuracy': '0.706', 'epoch': '0.7218'}
+{'loss': '0.6351', 'grad_norm': '7.125', 'learning_rate': '2.782e-06', 'num_tokens': '2.754e+06', 'mean_token_accuracy': '0.8062', 'epoch': '0.722'}
+{'loss': '0.4409', 'grad_norm': '14.75', 'learning_rate': '2.78e-06', 'num_tokens': '2.754e+06', 'mean_token_accuracy': '0.8671', 'epoch': '0.7221'}
+{'loss': '1.112', 'grad_norm': '8.312', 'learning_rate': '2.779e-06', 'num_tokens': '2.755e+06', 'mean_token_accuracy': '0.7538', 'epoch': '0.7223'}
+{'loss': '2.412', 'grad_norm': '15.5', 'learning_rate': '2.777e-06', 'num_tokens': '2.756e+06', 'mean_token_accuracy': '0.5349', 'epoch': '0.7225'}
+{'loss': '1.054', 'grad_norm': '8.062', 'learning_rate': '2.775e-06', 'num_tokens': '2.758e+06', 'mean_token_accuracy': '0.7151', 'epoch': '0.7226'}
+{'loss': '1.238', 'grad_norm': '13.25', 'learning_rate': '2.774e-06', 'num_tokens': '2.759e+06', 'mean_token_accuracy': '0.6776', 'epoch': '0.7228'}
+{'loss': '0.7254', 'grad_norm': '35.5', 'learning_rate': '2.772e-06', 'num_tokens': '2.759e+06', 'mean_token_accuracy': '0.8214', 'epoch': '0.723'}
+{'loss': '2.179', 'grad_norm': '10.25', 'learning_rate': '2.77e-06', 'num_tokens': '2.76e+06', 'mean_token_accuracy': '0.5301', 'epoch': '0.7232'}
+{'loss': '1.799', 'grad_norm': '8.812', 'learning_rate': '2.768e-06', 'num_tokens': '2.761e+06', 'mean_token_accuracy': '0.6311', 'epoch': '0.7233'}
+
 72%|███████▏  | 4166/5758 [23:43<08:28,  3.13it/s]
 72%|███████▏  | 4167/5758 [23:43<09:02,  2.93it/s]
                                                   

 72%|███████▏  | 4167/5758 [23:43<09:02,  2.93it/s]
 72%|███████▏  | 4168/5758 [23:44<08:39,  3.06it/s]
                                                   

 72%|███████▏  | 4168/5758 [23:44<08:39,  3.06it/s]
 72%|███████▏  | 4169/5758 [23:44<08:59,  2.95it/s]
                                                   

 72%|███████▏  | 4169/5758 [23:44<08:59,  2.95it/s]
 72%|███████▏  | 4170/5758 [23:44<08:41,  3.05it/s]
                                                   

 72%|███████▏  | 4170/5758 [23:44<08:41,  3.05it/s]
 72%|███████▏  | 4171/5758 [23:45<08:41,  3.05it/s]
                                                   

 72%|███████▏  | 4171/5758 [23:45<08:41,  3.05it/s]
 72%|███████▏  | 4172/5758 [23:45<08:59,  2.94it/s]
                                                   

 72%|███████▏  | 4172/5758 [23:45<08:59,  2.94it/s]
 72%|███████▏  | 4173/5758 [23:45<09:06,  2.90it/s]
                                                   

 72%|███████▏  | 4173/5758 [23:45<09:06,  2.90it/s]
 72%|███████▏  | 4174/5758 [23:46<08:46,  3.01it/s]
                                                   

 72%|███████▏  | 4174/5758 [23:46<08:46,  3.01it/s]
 73%|███████▎  | 4175/5758 [23:46<08:26,  3.13it/s]
                                                   

 73%|███████▎  | 4175/5758 [23:46<08:26,  3.13it/s]
 73%|███████▎  | 4176/5758 [23:46<08:41,  3.03it/s]
                                                   

 73%|███████▎  | 4176/5758 [23:46<08:41,  3.03it/s]
 73%|███████▎  | 4177/5758 [23:47<08:26,  3.12it/s]
                                                   

 73%|███████▎  | 4177/5758 [23:47<08:26,  3.12it/s]
 73%|███████▎  | 4178/5758 [23:47<08:19,  3.17it/s]
                                                   

 73%|███████▎  | 4178/5758 [23:47<08:19,  3.17it/s]
 73%|███████▎  | 4179/5758 [23:47<08:50,  2.97it/s]
                                                   

 73%|███████▎  | 4179/5758 [23:47<08:50,  2.97it/s]
 73%|███████▎  | 4180/5758 [23:48<08:45,  3.00it/s]
                                                   

 73%|███████▎  | 4180/5758 [23:48<08:45,  3.00it/s]
 73%|███████▎  | 4181/5758 [23:48<08:44,  3.01it/s]
                                                   

 73%|███████▎  | 4181/5758 [23:48<08:44,  3.01it/s]
 73%|███████▎  | 4182/5758 [23:48<08:42,  3.02it/s]
                                                   

 73%|███████▎  | 4182/5758 [23:48<08:42,  3.02it/s]
 73%|███████▎  | 4183/5758 [23:49<08:51,  2.96it/s]
                                                   

 73%|███████▎  | 4183/5758 [23:49<08:51,  2.96it/s]
 73%|███████▎  | 4184/5758 [23:49<08:37,  3.04it/s]
                                                   

 73%|███████▎  | 4184/5758 [23:49<08:37,  3.04it/s]
 73%|███████▎  | 4185/5758 [23:49<08:56,  2.93it/s]
                                                   

 73%|███████▎  | 4185/5758 [23:49<08:56,  2.93it/s]
 73%|███████▎  | 4186/5758 [23:50<08:45,  2.99it/s]
                                                   

 73%|███████▎  | 4186/5758 [23:50<08:45,  2.99it/s]
 73%|███████▎  | 4187/5758 [23:50<08:41,  3.01it/s]
                                                   

 73%|███████▎  | 4187/5758 [23:50<08:41,  3.01it/s]
 73%|███████▎  | 4188/5758 [23:50<09:25,  2.78it/s]
                                                   

 73%|███████▎  | 4188/5758 [23:50<09:25,  2.78it/s]
 73%|███████▎  | 4189/5758 [23:51<09:06,  2.87it/s]
                                                   

 73%|███████▎  | 4189/5758 [23:51<09:06,  2.87it/s]
 73%|███████▎  | 4190/5758 [23:51<08:42,  3.00it/s]
                                                   

 73%|███████▎  | 4190/5758 [23:51<08:42,  3.00it/s]
 73%|███████▎  | 4191/5758 [23:51<08:21,  3.12it/s]
                                                   

 73%|███████▎  | 4191/5758 [23:51<08:21,  3.12it/s]
 73%|███████▎  | 4192/5758 [23:52<08:16,  3.15it/s]
                                                   

 73%|███████▎  | 4192/5758 [23:52<08:16,  3.15it/s]
 73%|███████▎  | 4193/5758 [23:52<08:49,  2.96it/s]
                                                   

 73%|███████▎  | 4193/5758 [23:52<08:49,  2.96it/s]
 73%|███████▎  | 4194/5758 [23:52<08:39,  3.01it/s]
                                                   

 73%|███████▎  | 4194/5758 [23:52<08:39,  3.01it/s]
 73%|███████▎  | 4195/5758 [23:53<08:34,  3.04it/s]
                                                   

 73%|███████▎  | 4195/5758 [23:53<08:34,  3.04it/s]
 73%|███████▎  | 4196/5758 [23:53<08:44,  2.98it/s]
                                                   

 73%|███████▎  | 4196/5758 [23:53<08:44,  2.98it/s]
 73%|███████▎  | 4197/5758 [23:53<08:32,  3.05it/s]
                                                   

 73%|███████▎  | 4197/5758 [23:53<08:32,  3.05it/s]
 73%|███████▎  | 4198/5758 [23:54<08:44,  2.97it/s]
                                                   

 73%|███████▎  | 4198/5758 [23:54<08:44,  2.97it/s]
 73%|███████▎  | 4199/5758 [23:54<08:45,  2.97it/s]
                                                   

 73%|███████▎  | 4199/5758 [23:54<08:45,  2.97it/s]
 73%|███████▎  | 4200/5758 [23:54<08:54,  2.91it/s]
                                                   

 73%|███████▎  | 4200/5758 [23:54<08:54,  2.91it/s]
 73%|███████▎  | 4201/5758 [23:55<09:10,  2.83it/s]
                                                   

 73%|███████▎  | 4201/5758 [23:55<09:10,  2.83it/s]
 73%|███████▎  | 4202/5758 [23:55<08:53,  2.92it/s]
                                                   

 73%|███████▎  | 4202/5758 [23:55<08:53,  2.92it/s]
 73%|███████▎  | 4203/5758 [23:55<08:31,  3.04it/s]
                                                   

 73%|███████▎  | 4203/5758 [23:55<08:31,  3.04it/s]
 73%|███████▎  | 4204/5758 [23:56<08:46,  2.95it/s]
                                                   

 73%|███████▎  | 4204/5758 [23:56<08:46,  2.95it/s]
 73%|███████▎  | 4205/5758 [23:56<08:51,  2.92it/s]
                                                   

 73%|███████▎  | 4205/5758 [23:56<08:51,  2.92it/s]
 73%|███████▎  | 4206/5758 [23:56<08:42,  2.97it/s]
                                                   

 73%|███████▎  | 4206/5758 [23:56<08:42,  2.97it/s]
 73%|███████▎  | 4207/5758 [23:57<08:32,  3.03it/s]
                                                   

 73%|███████▎  | 4207/5758 [23:57<08:32,  3.03it/s]
 73%|███████▎  | 4208/5758 [23:57<08:22,  3.08it/s]
                                                   

 73%|███████▎  | 4208/5758 [23:57<08:22,  3.08it/s]
 73%|███████▎  | 4209/5758 [23:57<08:12,  3.15it/s]
                                                   

 73%|███████▎  | 4209/5758 [23:57<08:12,  3.15it/s]
 73%|███████▎  | 4210/5758 [23:58<08:02,  3.21it/s]
                                                   

 73%|███████▎  | 4210/5758 [23:58<08:02,  3.21it/s]
 73%|███████▎  | 4211/5758 [23:58<08:01,  3.21it/s]
                                                   

 73%|███████▎  | 4211/5758 [23:58<08:01,  3.21it/s]
 73%|███████▎  | 4212/5758 [23:58<08:35,  3.00it/s]
                                                   

 73%|███████▎  | 4212/5758 [23:58<08:35,  3.00it/s]
 73%|███████▎  | 4213/5758 [23:59<08:30,  3.03it/s]
                                                   

 73%|███████▎  | 4213/5758 [23:59<08:30,  3.03it/s]
 73%|███████▎  | 4214/5758 [23:59<09:09,  2.81it/s]
                                                   

 73%|███████▎  | 4214/5758 [23:59<09:09,  2.81it/s]
 73%|███████▎  | 4215/5758 [23:59<09:19,  2.76it/s]
                                                   

 73%|███████▎  | 4215/5758 [23:59<09:19,  2.76it/s]
 73%|███████▎  | 4216/5758 [24:00<08:57,  2.87it/s]
                                                   

 73%|███████▎  | 4216/5758 [24:00<08:57,  2.87it/s]
 73%|███████▎  | 4217/5758 [24:00<08:28,  3.03it/s]
                                                   

 73%|███████▎  | 4217/5758 [24:00<08:28,  3.03it/s]
 73%|███████▎  | 4218/5758 [24:00<08:26,  3.04it/s]
                                                   

 73%|███████▎  | 4218/5758 [24:00<08:26,  3.04it/s]
 73%|███████▎  | 4219/5758 [24:01<08:17,  3.09it/s]
                                                   

 73%|███████▎  | 4219/5758 [24:01<08:17,  3.09it/s]
 73%|███████▎  | 4220/5758 [24:01<08:17,  3.09it/s]
                                                   

 73%|███████▎  | 4220/5758 [24:01<08:17,  3.09it/s]
 73%|███████▎  | 4221/5758 [24:01<08:08,  3.15it/s]
                                                   
{'loss': '1.39', 'grad_norm': '12.19', 'learning_rate': '2.767e-06', 'num_tokens': '2.762e+06', 'mean_token_accuracy': '0.6485', 'epoch': '0.7235'}
+{'loss': '1.612', 'grad_norm': '41', 'learning_rate': '2.765e-06', 'num_tokens': '2.762e+06', 'mean_token_accuracy': '0.6642', 'epoch': '0.7237'}
+{'loss': '1.401', 'grad_norm': '28.75', 'learning_rate': '2.763e-06', 'num_tokens': '2.762e+06', 'mean_token_accuracy': '0.6582', 'epoch': '0.7239'}
+{'loss': '1.979', 'grad_norm': '9.688', 'learning_rate': '2.761e-06', 'num_tokens': '2.763e+06', 'mean_token_accuracy': '0.5802', 'epoch': '0.724'}
+{'loss': '0.9948', 'grad_norm': '11.88', 'learning_rate': '2.76e-06', 'num_tokens': '2.764e+06', 'mean_token_accuracy': '0.7166', 'epoch': '0.7242'}
+{'loss': '1.029', 'grad_norm': '4.938', 'learning_rate': '2.758e-06', 'num_tokens': '2.767e+06', 'mean_token_accuracy': '0.742', 'epoch': '0.7244'}
+{'loss': '0.4032', 'grad_norm': '17.12', 'learning_rate': '2.756e-06', 'num_tokens': '2.768e+06', 'mean_token_accuracy': '0.9008', 'epoch': '0.7246'}
+{'loss': '2.315', 'grad_norm': '21', 'learning_rate': '2.754e-06', 'num_tokens': '2.768e+06', 'mean_token_accuracy': '0.5514', 'epoch': '0.7247'}
+{'loss': '1.319', 'grad_norm': '16.88', 'learning_rate': '2.753e-06', 'num_tokens': '2.768e+06', 'mean_token_accuracy': '0.7233', 'epoch': '0.7249'}
+{'loss': '0.5869', 'grad_norm': '7.812', 'learning_rate': '2.751e-06', 'num_tokens': '2.769e+06', 'mean_token_accuracy': '0.8307', 'epoch': '0.7251'}
+{'loss': '1.684', 'grad_norm': '14.88', 'learning_rate': '2.749e-06', 'num_tokens': '2.77e+06', 'mean_token_accuracy': '0.6472', 'epoch': '0.7253'}
+{'loss': '0.5078', 'grad_norm': '20.88', 'learning_rate': '2.747e-06', 'num_tokens': '2.77e+06', 'mean_token_accuracy': '0.8994', 'epoch': '0.7254'}
+{'loss': '0.8919', 'grad_norm': '11.94', 'learning_rate': '2.746e-06', 'num_tokens': '2.771e+06', 'mean_token_accuracy': '0.7605', 'epoch': '0.7256'}
+{'loss': '1.251', 'grad_norm': '14.75', 'learning_rate': '2.744e-06', 'num_tokens': '2.771e+06', 'mean_token_accuracy': '0.6484', 'epoch': '0.7258'}
+{'loss': '1.344', 'grad_norm': '13', 'learning_rate': '2.742e-06', 'num_tokens': '2.772e+06', 'mean_token_accuracy': '0.6217', 'epoch': '0.7259'}
+{'loss': '1.11', 'grad_norm': '20.12', 'learning_rate': '2.741e-06', 'num_tokens': '2.772e+06', 'mean_token_accuracy': '0.7404', 'epoch': '0.7261'}
+{'loss': '0.7831', 'grad_norm': '15.75', 'learning_rate': '2.739e-06', 'num_tokens': '2.773e+06', 'mean_token_accuracy': '0.8165', 'epoch': '0.7263'}
+{'loss': '0.334', 'grad_norm': '14.69', 'learning_rate': '2.737e-06', 'num_tokens': '2.773e+06', 'mean_token_accuracy': '0.9283', 'epoch': '0.7265'}
+{'loss': '1.734', 'grad_norm': '19.62', 'learning_rate': '2.735e-06', 'num_tokens': '2.773e+06', 'mean_token_accuracy': '0.6436', 'epoch': '0.7266'}
+{'loss': '1.15', 'grad_norm': '15.69', 'learning_rate': '2.734e-06', 'num_tokens': '2.774e+06', 'mean_token_accuracy': '0.7076', 'epoch': '0.7268'}
+{'loss': '1.265', 'grad_norm': '12.62', 'learning_rate': '2.732e-06', 'num_tokens': '2.775e+06', 'mean_token_accuracy': '0.6526', 'epoch': '0.727'}
+{'loss': '1.01', 'grad_norm': '16.5', 'learning_rate': '2.73e-06', 'num_tokens': '2.775e+06', 'mean_token_accuracy': '0.7669', 'epoch': '0.7272'}
+{'loss': '1.562', 'grad_norm': '17.12', 'learning_rate': '2.728e-06', 'num_tokens': '2.776e+06', 'mean_token_accuracy': '0.6352', 'epoch': '0.7273'}
+{'loss': '1.152', 'grad_norm': '13.06', 'learning_rate': '2.727e-06', 'num_tokens': '2.776e+06', 'mean_token_accuracy': '0.6751', 'epoch': '0.7275'}
+{'loss': '1.179', 'grad_norm': '12.88', 'learning_rate': '2.725e-06', 'num_tokens': '2.777e+06', 'mean_token_accuracy': '0.6858', 'epoch': '0.7277'}
+{'loss': '1.073', 'grad_norm': '23', 'learning_rate': '2.723e-06', 'num_tokens': '2.777e+06', 'mean_token_accuracy': '0.8133', 'epoch': '0.7279'}
+{'loss': '0.8598', 'grad_norm': '21.5', 'learning_rate': '2.721e-06', 'num_tokens': '2.777e+06', 'mean_token_accuracy': '0.7933', 'epoch': '0.728'}
+{'loss': '1.905', 'grad_norm': '9.25', 'learning_rate': '2.72e-06', 'num_tokens': '2.779e+06', 'mean_token_accuracy': '0.6063', 'epoch': '0.7282'}
+{'loss': '0.9818', 'grad_norm': '38', 'learning_rate': '2.718e-06', 'num_tokens': '2.779e+06', 'mean_token_accuracy': '0.8182', 'epoch': '0.7284'}
+{'loss': '1.749', 'grad_norm': '18.88', 'learning_rate': '2.716e-06', 'num_tokens': '2.779e+06', 'mean_token_accuracy': '0.6348', 'epoch': '0.7286'}
+{'loss': '0.6802', 'grad_norm': '13.88', 'learning_rate': '2.714e-06', 'num_tokens': '2.78e+06', 'mean_token_accuracy': '0.8347', 'epoch': '0.7287'}
+{'loss': '2.058', 'grad_norm': '12.31', 'learning_rate': '2.713e-06', 'num_tokens': '2.78e+06', 'mean_token_accuracy': '0.55', 'epoch': '0.7289'}
+{'loss': '1.186', 'grad_norm': '10.5', 'learning_rate': '2.711e-06', 'num_tokens': '2.781e+06', 'mean_token_accuracy': '0.6575', 'epoch': '0.7291'}
+{'loss': '2.295', 'grad_norm': '22.5', 'learning_rate': '2.709e-06', 'num_tokens': '2.781e+06', 'mean_token_accuracy': '0.5247', 'epoch': '0.7292'}
+{'loss': '0.9424', 'grad_norm': '31.88', 'learning_rate': '2.708e-06', 'num_tokens': '2.782e+06', 'mean_token_accuracy': '0.807', 'epoch': '0.7294'}
+{'loss': '1.474', 'grad_norm': '10.5', 'learning_rate': '2.706e-06', 'num_tokens': '2.783e+06', 'mean_token_accuracy': '0.6103', 'epoch': '0.7296'}
+{'loss': '0.8059', 'grad_norm': '31.5', 'learning_rate': '2.704e-06', 'num_tokens': '2.783e+06', 'mean_token_accuracy': '0.8023', 'epoch': '0.7298'}
+{'loss': '2.102', 'grad_norm': '12.25', 'learning_rate': '2.702e-06', 'num_tokens': '2.783e+06', 'mean_token_accuracy': '0.5459', 'epoch': '0.7299'}
+{'loss': '1.526', 'grad_norm': '11.56', 'learning_rate': '2.701e-06', 'num_tokens': '2.784e+06', 'mean_token_accuracy': '0.5973', 'epoch': '0.7301'}
+{'loss': '2.405', 'grad_norm': '23.88', 'learning_rate': '2.699e-06', 'num_tokens': '2.785e+06', 'mean_token_accuracy': '0.534', 'epoch': '0.7303'}
+{'loss': '1.248', 'grad_norm': '12.62', 'learning_rate': '2.697e-06', 'num_tokens': '2.786e+06', 'mean_token_accuracy': '0.6573', 'epoch': '0.7305'}
+{'loss': '2.011', 'grad_norm': '11.88', 'learning_rate': '2.695e-06', 'num_tokens': '2.786e+06', 'mean_token_accuracy': '0.5785', 'epoch': '0.7306'}
+{'loss': '1.214', 'grad_norm': '32', 'learning_rate': '2.694e-06', 'num_tokens': '2.787e+06', 'mean_token_accuracy': '0.7195', 'epoch': '0.7308'}
+{'loss': '0.7862', 'grad_norm': '40.5', 'learning_rate': '2.692e-06', 'num_tokens': '2.787e+06', 'mean_token_accuracy': '0.8', 'epoch': '0.731'}
+{'loss': '0.9684', 'grad_norm': '12.69', 'learning_rate': '2.69e-06', 'num_tokens': '2.787e+06', 'mean_token_accuracy': '0.7366', 'epoch': '0.7312'}
+{'loss': '0.7142', 'grad_norm': '16.5', 'learning_rate': '2.688e-06', 'num_tokens': '2.788e+06', 'mean_token_accuracy': '0.821', 'epoch': '0.7313'}
+{'loss': '1.259', 'grad_norm': '15.31', 'learning_rate': '2.687e-06', 'num_tokens': '2.788e+06', 'mean_token_accuracy': '0.643', 'epoch': '0.7315'}
+{'loss': '1.177', 'grad_norm': '17.25', 'learning_rate': '2.685e-06', 'num_tokens': '2.789e+06', 'mean_token_accuracy': '0.7108', 'epoch': '0.7317'}
+{'loss': '1.19', 'grad_norm': '33', 'learning_rate': '2.683e-06', 'num_tokens': '2.789e+06', 'mean_token_accuracy': '0.7879', 'epoch': '0.7319'}
+{'loss': '0.8127', 'grad_norm': '11.75', 'learning_rate': '2.681e-06', 'num_tokens': '2.789e+06', 'mean_token_accuracy': '0.827', 'epoch': '0.732'}
+{'loss': '1.692', 'grad_norm': '14.69', 'learning_rate': '2.68e-06', 'num_tokens': '2.79e+06', 'mean_token_accuracy': '0.6253', 'epoch': '0.7322'}
+{'loss': '1.4', 'grad_norm': '40.5', 'learning_rate': '2.678e-06', 'num_tokens': '2.79e+06', 'mean_token_accuracy': '0.6912', 'epoch': '0.7324'}
+{'loss': '1.281', 'grad_norm': '22.12', 'learning_rate': '2.676e-06', 'num_tokens': '2.79e+06', 'mean_token_accuracy': '0.7333', 'epoch': '0.7325'}
+{'loss': '1.924', 'grad_norm': '16.62', 'learning_rate': '2.675e-06', 'num_tokens': '2.791e+06', 'mean_token_accuracy': '0.6102', 'epoch': '0.7327'}
+{'loss': '1.768', 'grad_norm': '35', 'learning_rate': '2.673e-06', 'num_tokens': '2.791e+06', 'mean_token_accuracy': '0.6882', 'epoch': '0.7329'}
+
 73%|███████▎  | 4221/5758 [24:01<08:08,  3.15it/s]
 73%|███████▎  | 4222/5758 [24:02<07:59,  3.21it/s]
                                                   

 73%|███████▎  | 4222/5758 [24:02<07:59,  3.21it/s]
 73%|███████▎  | 4223/5758 [24:02<08:34,  2.99it/s]
                                                   

 73%|███████▎  | 4223/5758 [24:02<08:34,  2.99it/s]
 73%|███████▎  | 4224/5758 [24:02<08:21,  3.06it/s]
                                                   

 73%|███████▎  | 4224/5758 [24:02<08:21,  3.06it/s]
 73%|███████▎  | 4225/5758 [24:03<08:13,  3.10it/s]
                                                   

 73%|███████▎  | 4225/5758 [24:03<08:13,  3.10it/s]
 73%|███████▎  | 4226/5758 [24:03<08:25,  3.03it/s]
                                                   

 73%|███████▎  | 4226/5758 [24:03<08:25,  3.03it/s]
 73%|███████▎  | 4227/5758 [24:03<08:46,  2.91it/s]
                                                   

 73%|███████▎  | 4227/5758 [24:03<08:46,  2.91it/s]
 73%|███████▎  | 4228/5758 [24:04<08:25,  3.03it/s]
                                                   

 73%|███████▎  | 4228/5758 [24:04<08:25,  3.03it/s]
 73%|███████▎  | 4229/5758 [24:04<08:13,  3.10it/s]
                                                   

 73%|███████▎  | 4229/5758 [24:04<08:13,  3.10it/s]
 73%|███████▎  | 4230/5758 [24:04<08:16,  3.07it/s]
                                                   

 73%|███████▎  | 4230/5758 [24:04<08:16,  3.07it/s]
 73%|███████▎  | 4231/5758 [24:05<08:24,  3.02it/s]
                                                   

 73%|███████▎  | 4231/5758 [24:05<08:24,  3.02it/s]
 73%|███████▎  | 4232/5758 [24:05<07:55,  3.21it/s]
                                                   

 73%|███████▎  | 4232/5758 [24:05<07:55,  3.21it/s]
 74%|███████▎  | 4233/5758 [24:05<07:47,  3.26it/s]
                                                   

 74%|███████▎  | 4233/5758 [24:05<07:47,  3.26it/s]
 74%|███████▎  | 4234/5758 [24:06<08:20,  3.05it/s]
                                                   

 74%|███████▎  | 4234/5758 [24:06<08:20,  3.05it/s]
 74%|███████▎  | 4235/5758 [24:06<08:03,  3.15it/s]
                                                   

 74%|███████▎  | 4235/5758 [24:06<08:03,  3.15it/s]
 74%|███████▎  | 4236/5758 [24:06<07:56,  3.20it/s]
                                                   

 74%|███████▎  | 4236/5758 [24:06<07:56,  3.20it/s]
 74%|███████▎  | 4237/5758 [24:07<07:52,  3.22it/s]
                                                   

 74%|███████▎  | 4237/5758 [24:07<07:52,  3.22it/s]
 74%|███████▎  | 4238/5758 [24:07<08:09,  3.11it/s]
                                                   

 74%|███████▎  | 4238/5758 [24:07<08:09,  3.11it/s]
 74%|███████▎  | 4239/5758 [24:07<08:00,  3.16it/s]
                                                   

 74%|███████▎  | 4239/5758 [24:07<08:00,  3.16it/s]
 74%|███████▎  | 4240/5758 [24:08<09:05,  2.78it/s]
                                                   

 74%|███████▎  | 4240/5758 [24:08<09:05,  2.78it/s]
 74%|███████▎  | 4241/5758 [24:08<09:15,  2.73it/s]
                                                   

 74%|███████▎  | 4241/5758 [24:08<09:15,  2.73it/s]
 74%|███████▎  | 4242/5758 [24:08<08:49,  2.86it/s]
                                                   

 74%|███████▎  | 4242/5758 [24:08<08:49,  2.86it/s]
 74%|███████▎  | 4243/5758 [24:09<08:24,  3.00it/s]
                                                   

 74%|███████▎  | 4243/5758 [24:09<08:24,  3.00it/s]
 74%|███████▎  | 4244/5758 [24:09<08:13,  3.07it/s]
                                                   

 74%|███████▎  | 4244/5758 [24:09<08:13,  3.07it/s]
 74%|███████▎  | 4245/5758 [24:09<08:24,  3.00it/s]
                                                   

 74%|███████▎  | 4245/5758 [24:09<08:24,  3.00it/s]
 74%|███████▎  | 4246/5758 [24:10<08:18,  3.04it/s]
                                                   

 74%|███████▎  | 4246/5758 [24:10<08:18,  3.04it/s]
 74%|███████▍  | 4247/5758 [24:10<08:04,  3.12it/s]
                                                   

 74%|███████▍  | 4247/5758 [24:10<08:04,  3.12it/s]
 74%|███████▍  | 4248/5758 [24:10<07:57,  3.16it/s]
                                                   

 74%|███████▍  | 4248/5758 [24:10<07:57,  3.16it/s]
 74%|███████▍  | 4249/5758 [24:10<07:47,  3.23it/s]
                                                   

 74%|███████▍  | 4249/5758 [24:10<07:47,  3.23it/s]
 74%|███████▍  | 4250/5758 [24:11<07:51,  3.20it/s]
                                                   

 74%|███████▍  | 4250/5758 [24:11<07:51,  3.20it/s]
 74%|███████▍  | 4251/5758 [24:11<07:58,  3.15it/s]
                                                   

 74%|███████▍  | 4251/5758 [24:11<07:58,  3.15it/s]
 74%|███████▍  | 4252/5758 [24:11<08:13,  3.05it/s]
                                                   

 74%|███████▍  | 4252/5758 [24:11<08:13,  3.05it/s]
 74%|███████▍  | 4253/5758 [24:12<08:07,  3.09it/s]
                                                   

 74%|███████▍  | 4253/5758 [24:12<08:07,  3.09it/s]
 74%|███████▍  | 4254/5758 [24:12<08:47,  2.85it/s]
                                                   

 74%|███████▍  | 4254/5758 [24:12<08:47,  2.85it/s]
 74%|███████▍  | 4255/5758 [24:13<08:38,  2.90it/s]
                                                   

 74%|███████▍  | 4255/5758 [24:13<08:38,  2.90it/s]
 74%|███████▍  | 4256/5758 [24:13<08:39,  2.89it/s]
                                                   

 74%|███████▍  | 4256/5758 [24:13<08:39,  2.89it/s]
 74%|███████▍  | 4257/5758 [24:13<08:17,  3.02it/s]
                                                   

 74%|███████▍  | 4257/5758 [24:13<08:17,  3.02it/s]
 74%|███████▍  | 4258/5758 [24:13<08:04,  3.09it/s]
                                                   

 74%|███████▍  | 4258/5758 [24:13<08:04,  3.09it/s]
 74%|███████▍  | 4259/5758 [24:14<08:23,  2.98it/s]
                                                   

 74%|███████▍  | 4259/5758 [24:14<08:23,  2.98it/s]
 74%|███████▍  | 4260/5758 [24:14<08:03,  3.10it/s]
                                                   

 74%|███████▍  | 4260/5758 [24:14<08:03,  3.10it/s]
 74%|███████▍  | 4261/5758 [24:14<08:04,  3.09it/s]
                                                   

 74%|███████▍  | 4261/5758 [24:14<08:04,  3.09it/s]
 74%|███████▍  | 4262/5758 [24:15<08:28,  2.94it/s]
                                                   

 74%|███████▍  | 4262/5758 [24:15<08:28,  2.94it/s]
 74%|███████▍  | 4263/5758 [24:15<08:45,  2.84it/s]
                                                   

 74%|███████▍  | 4263/5758 [24:15<08:45,  2.84it/s]
 74%|███████▍  | 4264/5758 [24:16<08:36,  2.89it/s]
                                                   

 74%|███████▍  | 4264/5758 [24:16<08:36,  2.89it/s]
 74%|███████▍  | 4265/5758 [24:16<08:57,  2.78it/s]
                                                   

 74%|███████▍  | 4265/5758 [24:16<08:57,  2.78it/s]
 74%|███████▍  | 4266/5758 [24:16<09:51,  2.52it/s]
                                                   

 74%|███████▍  | 4266/5758 [24:16<09:51,  2.52it/s]
 74%|███████▍  | 4267/5758 [24:17<09:40,  2.57it/s]
                                                   

 74%|███████▍  | 4267/5758 [24:17<09:40,  2.57it/s]
 74%|███████▍  | 4268/5758 [24:17<09:16,  2.68it/s]
                                                   

 74%|███████▍  | 4268/5758 [24:17<09:16,  2.68it/s]
 74%|███████▍  | 4269/5758 [24:18<09:12,  2.70it/s]
                                                   

 74%|███████▍  | 4269/5758 [24:18<09:12,  2.70it/s]
 74%|███████▍  | 4270/5758 [24:18<08:47,  2.82it/s]
                                                   

 74%|███████▍  | 4270/5758 [24:18<08:47,  2.82it/s]
 74%|███████▍  | 4271/5758 [24:18<08:38,  2.87it/s]
                                                   

 74%|███████▍  | 4271/5758 [24:18<08:38,  2.87it/s]
 74%|███████▍  | 4272/5758 [24:18<08:14,  3.00it/s]
                                                   

 74%|███████▍  | 4272/5758 [24:18<08:14,  3.00it/s]
 74%|███████▍  | 4273/5758 [24:19<07:58,  3.10it/s]
                                                   

 74%|███████▍  | 4273/5758 [24:19<07:58,  3.10it/s]
 74%|███████▍  | 4274/5758 [24:19<08:11,  3.02it/s]
                                                   

 74%|███████▍  | 4274/5758 [24:19<08:11,  3.02it/s]
 74%|███████▍  | 4275/5758 [24:19<08:06,  3.05it/s]
                                                   

 74%|███████▍  | 4275/5758 [24:19<08:06,  3.05it/s]
 74%|███████▍  | 4276/5758 [24:20<07:54,  3.12it/s]
                                                   
{'loss': '1.099', 'grad_norm': '12.25', 'learning_rate': '2.671e-06', 'num_tokens': '2.792e+06', 'mean_token_accuracy': '0.7017', 'epoch': '0.7331'}
+{'loss': '1.094', 'grad_norm': '12.94', 'learning_rate': '2.669e-06', 'num_tokens': '2.792e+06', 'mean_token_accuracy': '0.6927', 'epoch': '0.7332'}
+{'loss': '1.291', 'grad_norm': '15.5', 'learning_rate': '2.668e-06', 'num_tokens': '2.793e+06', 'mean_token_accuracy': '0.6616', 'epoch': '0.7334'}
+{'loss': '0.4559', 'grad_norm': '6.562', 'learning_rate': '2.666e-06', 'num_tokens': '2.794e+06', 'mean_token_accuracy': '0.8596', 'epoch': '0.7336'}
+{'loss': '1.014', 'grad_norm': '12.44', 'learning_rate': '2.664e-06', 'num_tokens': '2.795e+06', 'mean_token_accuracy': '0.7036', 'epoch': '0.7338'}
+{'loss': '1.068', 'grad_norm': '10.56', 'learning_rate': '2.662e-06', 'num_tokens': '2.796e+06', 'mean_token_accuracy': '0.753', 'epoch': '0.7339'}
+{'loss': '0.9849', 'grad_norm': '29.5', 'learning_rate': '2.661e-06', 'num_tokens': '2.796e+06', 'mean_token_accuracy': '0.7717', 'epoch': '0.7341'}
+{'loss': '1.816', 'grad_norm': '18.5', 'learning_rate': '2.659e-06', 'num_tokens': '2.796e+06', 'mean_token_accuracy': '0.6144', 'epoch': '0.7343'}
+{'loss': '0.6077', 'grad_norm': '13.12', 'learning_rate': '2.657e-06', 'num_tokens': '2.797e+06', 'mean_token_accuracy': '0.8513', 'epoch': '0.7345'}
+{'loss': '2.095', 'grad_norm': '11.56', 'learning_rate': '2.655e-06', 'num_tokens': '2.798e+06', 'mean_token_accuracy': '0.5714', 'epoch': '0.7346'}
+{'loss': '1.457', 'grad_norm': '12.38', 'learning_rate': '2.654e-06', 'num_tokens': '2.798e+06', 'mean_token_accuracy': '0.5912', 'epoch': '0.7348'}
+{'loss': '0.5622', 'grad_norm': '31.75', 'learning_rate': '2.652e-06', 'num_tokens': '2.798e+06', 'mean_token_accuracy': '0.8654', 'epoch': '0.735'}
+{'loss': '1.091', 'grad_norm': '14.38', 'learning_rate': '2.65e-06', 'num_tokens': '2.799e+06', 'mean_token_accuracy': '0.6614', 'epoch': '0.7352'}
+{'loss': '0.5808', 'grad_norm': '22.75', 'learning_rate': '2.648e-06', 'num_tokens': '2.799e+06', 'mean_token_accuracy': '0.856', 'epoch': '0.7353'}
+{'loss': '1.066', 'grad_norm': '9.688', 'learning_rate': '2.647e-06', 'num_tokens': '2.801e+06', 'mean_token_accuracy': '0.7006', 'epoch': '0.7355'}
+{'loss': '1.947', 'grad_norm': '63', 'learning_rate': '2.645e-06', 'num_tokens': '2.801e+06', 'mean_token_accuracy': '0.6622', 'epoch': '0.7357'}
+{'loss': '1.024', 'grad_norm': '12.88', 'learning_rate': '2.643e-06', 'num_tokens': '2.801e+06', 'mean_token_accuracy': '0.6982', 'epoch': '0.7358'}
+{'loss': '1.974', 'grad_norm': '18.75', 'learning_rate': '2.642e-06', 'num_tokens': '2.802e+06', 'mean_token_accuracy': '0.5926', 'epoch': '0.736'}
+{'loss': '2.51', 'grad_norm': '10.69', 'learning_rate': '2.64e-06', 'num_tokens': '2.803e+06', 'mean_token_accuracy': '0.49', 'epoch': '0.7362'}
+{'loss': '1.046', 'grad_norm': '7.031', 'learning_rate': '2.638e-06', 'num_tokens': '2.805e+06', 'mean_token_accuracy': '0.7368', 'epoch': '0.7364'}
+{'loss': '1.646', 'grad_norm': '13.06', 'learning_rate': '2.636e-06', 'num_tokens': '2.805e+06', 'mean_token_accuracy': '0.6649', 'epoch': '0.7365'}
+{'loss': '0.5964', 'grad_norm': '7.125', 'learning_rate': '2.635e-06', 'num_tokens': '2.807e+06', 'mean_token_accuracy': '0.8187', 'epoch': '0.7367'}
+{'loss': '0.6163', 'grad_norm': '8.312', 'learning_rate': '2.633e-06', 'num_tokens': '2.808e+06', 'mean_token_accuracy': '0.8206', 'epoch': '0.7369'}
+{'loss': '1.223', 'grad_norm': '13.44', 'learning_rate': '2.631e-06', 'num_tokens': '2.809e+06', 'mean_token_accuracy': '0.6776', 'epoch': '0.7371'}
+{'loss': '0.4334', 'grad_norm': '16.25', 'learning_rate': '2.629e-06', 'num_tokens': '2.809e+06', 'mean_token_accuracy': '0.8921', 'epoch': '0.7372'}
+{'loss': '1.279', 'grad_norm': '12.94', 'learning_rate': '2.628e-06', 'num_tokens': '2.81e+06', 'mean_token_accuracy': '0.6543', 'epoch': '0.7374'}
+{'loss': '0.5625', 'grad_norm': '7.375', 'learning_rate': '2.626e-06', 'num_tokens': '2.811e+06', 'mean_token_accuracy': '0.8306', 'epoch': '0.7376'}
+{'loss': '1.375', 'grad_norm': '45.75', 'learning_rate': '2.624e-06', 'num_tokens': '2.811e+06', 'mean_token_accuracy': '0.8', 'epoch': '0.7378'}
+{'loss': '1.068', 'grad_norm': '19.75', 'learning_rate': '2.622e-06', 'num_tokens': '2.811e+06', 'mean_token_accuracy': '0.7792', 'epoch': '0.7379'}
+{'loss': '0.5821', 'grad_norm': '7.219', 'learning_rate': '2.621e-06', 'num_tokens': '2.813e+06', 'mean_token_accuracy': '0.8276', 'epoch': '0.7381'}
+{'loss': '0.6603', 'grad_norm': '7.312', 'learning_rate': '2.619e-06', 'num_tokens': '2.815e+06', 'mean_token_accuracy': '0.8048', 'epoch': '0.7383'}
+{'loss': '0.4276', 'grad_norm': '15.62', 'learning_rate': '2.617e-06', 'num_tokens': '2.815e+06', 'mean_token_accuracy': '0.8771', 'epoch': '0.7385'}
+{'loss': '1.256', 'grad_norm': '12', 'learning_rate': '2.615e-06', 'num_tokens': '2.816e+06', 'mean_token_accuracy': '0.6567', 'epoch': '0.7386'}
+{'loss': '1.142', 'grad_norm': '28.75', 'learning_rate': '2.614e-06', 'num_tokens': '2.816e+06', 'mean_token_accuracy': '0.7687', 'epoch': '0.7388'}
+{'loss': '1.474', 'grad_norm': '13.81', 'learning_rate': '2.612e-06', 'num_tokens': '2.817e+06', 'mean_token_accuracy': '0.6152', 'epoch': '0.739'}
+{'loss': '0.755', 'grad_norm': '35.5', 'learning_rate': '2.61e-06', 'num_tokens': '2.817e+06', 'mean_token_accuracy': '0.8738', 'epoch': '0.7391'}
+{'loss': '0.7554', 'grad_norm': '11.56', 'learning_rate': '2.609e-06', 'num_tokens': '2.817e+06', 'mean_token_accuracy': '0.7922', 'epoch': '0.7393'}
+{'loss': '1.646', 'grad_norm': '14.19', 'learning_rate': '2.607e-06', 'num_tokens': '2.818e+06', 'mean_token_accuracy': '0.6609', 'epoch': '0.7395'}
+{'loss': '1.967', 'grad_norm': '10.06', 'learning_rate': '2.605e-06', 'num_tokens': '2.819e+06', 'mean_token_accuracy': '0.6112', 'epoch': '0.7397'}
+{'loss': '1.916', 'grad_norm': '12.62', 'learning_rate': '2.603e-06', 'num_tokens': '2.82e+06', 'mean_token_accuracy': '0.5997', 'epoch': '0.7398'}
+{'loss': '1.071', 'grad_norm': '15.44', 'learning_rate': '2.602e-06', 'num_tokens': '2.82e+06', 'mean_token_accuracy': '0.7568', 'epoch': '0.74'}
+{'loss': '1.011', 'grad_norm': '9.812', 'learning_rate': '2.6e-06', 'num_tokens': '2.821e+06', 'mean_token_accuracy': '0.7479', 'epoch': '0.7402'}
+{'loss': '1.253', 'grad_norm': '15.62', 'learning_rate': '2.598e-06', 'num_tokens': '2.822e+06', 'mean_token_accuracy': '0.6642', 'epoch': '0.7404'}
+{'loss': '1.844', 'grad_norm': '14.5', 'learning_rate': '2.596e-06', 'num_tokens': '2.822e+06', 'mean_token_accuracy': '0.6407', 'epoch': '0.7405'}
+{'loss': '0.6876', 'grad_norm': '12.06', 'learning_rate': '2.595e-06', 'num_tokens': '2.823e+06', 'mean_token_accuracy': '0.7787', 'epoch': '0.7407'}
+{'loss': '0.9459', 'grad_norm': '23.5', 'learning_rate': '2.593e-06', 'num_tokens': '2.823e+06', 'mean_token_accuracy': '0.8209', 'epoch': '0.7409'}
+{'loss': '0.6036', 'grad_norm': '39.75', 'learning_rate': '2.591e-06', 'num_tokens': '2.823e+06', 'mean_token_accuracy': '0.8375', 'epoch': '0.7411'}
+{'loss': '1.222', 'grad_norm': '13.56', 'learning_rate': '2.589e-06', 'num_tokens': '2.824e+06', 'mean_token_accuracy': '0.6731', 'epoch': '0.7412'}
+{'loss': '1.131', 'grad_norm': '17.5', 'learning_rate': '2.588e-06', 'num_tokens': '2.825e+06', 'mean_token_accuracy': '0.7455', 'epoch': '0.7414'}
+{'loss': '1.795', 'grad_norm': '12.38', 'learning_rate': '2.586e-06', 'num_tokens': '2.825e+06', 'mean_token_accuracy': '0.652', 'epoch': '0.7416'}
+{'loss': '1.987', 'grad_norm': '18', 'learning_rate': '2.584e-06', 'num_tokens': '2.826e+06', 'mean_token_accuracy': '0.6143', 'epoch': '0.7418'}
+{'loss': '0.7356', 'grad_norm': '46.75', 'learning_rate': '2.582e-06', 'num_tokens': '2.826e+06', 'mean_token_accuracy': '0.8429', 'epoch': '0.7419'}
+{'loss': '1.231', 'grad_norm': '13.44', 'learning_rate': '2.581e-06', 'num_tokens': '2.827e+06', 'mean_token_accuracy': '0.662', 'epoch': '0.7421'}
+{'loss': '0.9933', 'grad_norm': '18.75', 'learning_rate': '2.579e-06', 'num_tokens': '2.827e+06', 'mean_token_accuracy': '0.7769', 'epoch': '0.7423'}
+{'loss': '0.4888', 'grad_norm': '5.781', 'learning_rate': '2.577e-06', 'num_tokens': '2.828e+06', 'mean_token_accuracy': '0.8591', 'epoch': '0.7424'}
+
 74%|███████▍  | 4276/5758 [24:20<07:54,  3.12it/s]
 74%|███████▍  | 4277/5758 [24:20<07:47,  3.17it/s]
                                                   

 74%|███████▍  | 4277/5758 [24:20<07:47,  3.17it/s]
 74%|███████▍  | 4278/5758 [24:20<07:47,  3.17it/s]
                                                   

 74%|███████▍  | 4278/5758 [24:20<07:47,  3.17it/s]
 74%|███████▍  | 4279/5758 [24:21<07:44,  3.19it/s]
                                                   

 74%|███████▍  | 4279/5758 [24:21<07:44,  3.19it/s]
 74%|███████▍  | 4280/5758 [24:21<08:01,  3.07it/s]
                                                   

 74%|███████▍  | 4280/5758 [24:21<08:01,  3.07it/s]
 74%|███████▍  | 4281/5758 [24:21<07:58,  3.08it/s]
                                                   

 74%|███████▍  | 4281/5758 [24:21<07:58,  3.08it/s]
 74%|███████▍  | 4282/5758 [24:22<08:13,  2.99it/s]
                                                   

 74%|███████▍  | 4282/5758 [24:22<08:13,  2.99it/s]
 74%|███████▍  | 4283/5758 [24:22<08:38,  2.84it/s]
                                                   

 74%|███████▍  | 4283/5758 [24:22<08:38,  2.84it/s]
 74%|███████▍  | 4284/5758 [24:22<08:27,  2.91it/s]
                                                   

 74%|███████▍  | 4284/5758 [24:22<08:27,  2.91it/s]
 74%|███████▍  | 4285/5758 [24:23<08:03,  3.05it/s]
                                                   

 74%|███████▍  | 4285/5758 [24:23<08:03,  3.05it/s]
 74%|███████▍  | 4286/5758 [24:23<07:58,  3.08it/s]
                                                   

 74%|███████▍  | 4286/5758 [24:23<07:58,  3.08it/s]
 74%|███████▍  | 4287/5758 [24:23<07:58,  3.08it/s]
                                                   

 74%|███████▍  | 4287/5758 [24:23<07:58,  3.08it/s]
 74%|███████▍  | 4288/5758 [24:24<07:48,  3.14it/s]
                                                   

 74%|███████▍  | 4288/5758 [24:24<07:48,  3.14it/s]
 74%|███████▍  | 4289/5758 [24:24<07:44,  3.16it/s]
                                                   

 74%|███████▍  | 4289/5758 [24:24<07:44,  3.16it/s]
 75%|███████▍  | 4290/5758 [24:24<08:14,  2.97it/s]
                                                   

 75%|███████▍  | 4290/5758 [24:24<08:14,  2.97it/s]
 75%|███████▍  | 4291/5758 [24:25<07:52,  3.10it/s]
                                                   

 75%|███████▍  | 4291/5758 [24:25<07:52,  3.10it/s]
 75%|███████▍  | 4292/5758 [24:25<09:12,  2.65it/s]
                                                   

 75%|███████▍  | 4292/5758 [24:25<09:12,  2.65it/s]
 75%|███████▍  | 4293/5758 [24:26<09:07,  2.68it/s]
                                                   

 75%|███████▍  | 4293/5758 [24:26<09:07,  2.68it/s]
 75%|███████▍  | 4294/5758 [24:26<08:59,  2.71it/s]
                                                   

 75%|███████▍  | 4294/5758 [24:26<08:59,  2.71it/s]
 75%|███████▍  | 4295/5758 [24:26<08:50,  2.76it/s]
                                                   

 75%|███████▍  | 4295/5758 [24:26<08:50,  2.76it/s]
 75%|███████▍  | 4296/5758 [24:27<08:43,  2.79it/s]
                                                   

 75%|███████▍  | 4296/5758 [24:27<08:43,  2.79it/s]
 75%|███████▍  | 4297/5758 [24:27<08:24,  2.90it/s]
                                                   

 75%|███████▍  | 4297/5758 [24:27<08:24,  2.90it/s]
 75%|███████▍  | 4298/5758 [24:27<08:08,  2.99it/s]
                                                   

 75%|███████▍  | 4298/5758 [24:27<08:08,  2.99it/s]
 75%|███████▍  | 4299/5758 [24:27<07:54,  3.07it/s]
                                                   

 75%|███████▍  | 4299/5758 [24:27<07:54,  3.07it/s]
 75%|███████▍  | 4300/5758 [24:28<08:23,  2.89it/s]
                                                   

 75%|███████▍  | 4300/5758 [24:28<08:23,  2.89it/s]
 75%|███████▍  | 4301/5758 [24:28<08:02,  3.02it/s]
                                                   

 75%|███████▍  | 4301/5758 [24:28<08:02,  3.02it/s]
 75%|███████▍  | 4302/5758 [24:29<08:30,  2.85it/s]
                                                   

 75%|███████▍  | 4302/5758 [24:29<08:30,  2.85it/s]
 75%|███████▍  | 4303/5758 [24:29<08:06,  2.99it/s]
                                                   

 75%|███████▍  | 4303/5758 [24:29<08:06,  2.99it/s]
 75%|███████▍  | 4304/5758 [24:29<08:29,  2.86it/s]
                                                   

 75%|███████▍  | 4304/5758 [24:29<08:29,  2.86it/s]
 75%|███████▍  | 4305/5758 [24:30<08:09,  2.97it/s]
                                                   

 75%|███████▍  | 4305/5758 [24:30<08:09,  2.97it/s]
 75%|███████▍  | 4306/5758 [24:30<08:12,  2.95it/s]
                                                   

 75%|███████▍  | 4306/5758 [24:30<08:12,  2.95it/s]
 75%|███████▍  | 4307/5758 [24:30<08:01,  3.01it/s]
                                                   

 75%|███████▍  | 4307/5758 [24:30<08:01,  3.01it/s]
 75%|███████▍  | 4308/5758 [24:31<08:22,  2.88it/s]
                                                   

 75%|███████▍  | 4308/5758 [24:31<08:22,  2.88it/s]
 75%|███████▍  | 4309/5758 [24:31<08:07,  2.98it/s]
                                                   

 75%|███████▍  | 4309/5758 [24:31<08:07,  2.98it/s]
 75%|███████▍  | 4310/5758 [24:31<07:53,  3.06it/s]
                                                   

 75%|███████▍  | 4310/5758 [24:31<07:53,  3.06it/s]
 75%|███████▍  | 4311/5758 [24:32<08:11,  2.94it/s]
                                                   

 75%|███████▍  | 4311/5758 [24:32<08:11,  2.94it/s]
 75%|███████▍  | 4312/5758 [24:32<07:49,  3.08it/s]
                                                   

 75%|███████▍  | 4312/5758 [24:32<07:49,  3.08it/s]
 75%|███████▍  | 4313/5758 [24:32<07:37,  3.16it/s]
                                                   

 75%|███████▍  | 4313/5758 [24:32<07:37,  3.16it/s]
 75%|███████▍  | 4314/5758 [24:32<07:38,  3.15it/s]
                                                   

 75%|███████▍  | 4314/5758 [24:32<07:38,  3.15it/s]
 75%|███████▍  | 4315/5758 [24:33<07:28,  3.22it/s]
                                                   

 75%|███████▍  | 4315/5758 [24:33<07:28,  3.22it/s]
 75%|███████▍  | 4316/5758 [24:33<07:25,  3.24it/s]
                                                   

 75%|███████▍  | 4316/5758 [24:33<07:25,  3.24it/s]
 75%|███████▍  | 4317/5758 [24:33<07:28,  3.21it/s]
                                                   

 75%|███████▍  | 4317/5758 [24:33<07:28,  3.21it/s]
 75%|███████▍  | 4318/5758 [24:34<07:28,  3.21it/s]
                                                   

 75%|███████▍  | 4318/5758 [24:34<07:28,  3.21it/s]
 75%|███████▌  | 4319/5758 [24:34<08:24,  2.86it/s]
                                                   

 75%|███████▌  | 4319/5758 [24:34<08:24,  2.86it/s]
 75%|███████▌  | 4320/5758 [24:35<08:39,  2.77it/s]
                                                   

 75%|███████▌  | 4320/5758 [24:35<08:39,  2.77it/s]
 75%|███████▌  | 4321/5758 [24:35<08:48,  2.72it/s]
                                                   

 75%|███████▌  | 4321/5758 [24:35<08:48,  2.72it/s]
 75%|███████▌  | 4322/5758 [24:35<08:52,  2.70it/s]
                                                   

 75%|███████▌  | 4322/5758 [24:35<08:52,  2.70it/s]
 75%|███████▌  | 4323/5758 [24:36<08:28,  2.82it/s]
                                                   

 75%|███████▌  | 4323/5758 [24:36<08:28,  2.82it/s]
 75%|███████▌  | 4324/5758 [24:36<08:06,  2.95it/s]
                                                   

 75%|███████▌  | 4324/5758 [24:36<08:06,  2.95it/s]
 75%|███████▌  | 4325/5758 [24:36<08:04,  2.96it/s]
                                                   

 75%|███████▌  | 4325/5758 [24:36<08:04,  2.96it/s]
 75%|███████▌  | 4326/5758 [24:37<07:55,  3.01it/s]
                                                   

 75%|███████▌  | 4326/5758 [24:37<07:55,  3.01it/s]
 75%|███████▌  | 4327/5758 [24:37<07:38,  3.12it/s]
                                                   

 75%|███████▌  | 4327/5758 [24:37<07:38,  3.12it/s]
 75%|███████▌  | 4328/5758 [24:37<07:50,  3.04it/s]
                                                   

 75%|███████▌  | 4328/5758 [24:37<07:50,  3.04it/s]
 75%|███████▌  | 4329/5758 [24:38<07:50,  3.04it/s]
                                                   

 75%|███████▌  | 4329/5758 [24:38<07:50,  3.04it/s]
 75%|███████▌  | 4330/5758 [24:38<08:07,  2.93it/s]
                                                   

 75%|███████▌  | 4330/5758 [24:38<08:07,  2.93it/s]
 75%|███████▌  | 4331/5758 [24:38<07:56,  2.99it/s]
                                                   
{'loss': '0.9412', 'grad_norm': '24.88', 'learning_rate': '2.576e-06', 'num_tokens': '2.829e+06', 'mean_token_accuracy': '0.7989', 'epoch': '0.7426'}
+{'loss': '1.133', 'grad_norm': '11.62', 'learning_rate': '2.574e-06', 'num_tokens': '2.829e+06', 'mean_token_accuracy': '0.691', 'epoch': '0.7428'}
+{'loss': '1.239', 'grad_norm': '7.969', 'learning_rate': '2.572e-06', 'num_tokens': '2.831e+06', 'mean_token_accuracy': '0.7108', 'epoch': '0.743'}
+{'loss': '1.249', 'grad_norm': '13.69', 'learning_rate': '2.57e-06', 'num_tokens': '2.831e+06', 'mean_token_accuracy': '0.6517', 'epoch': '0.7431'}
+{'loss': '1.484', 'grad_norm': '37.75', 'learning_rate': '2.569e-06', 'num_tokens': '2.832e+06', 'mean_token_accuracy': '0.6893', 'epoch': '0.7433'}
+{'loss': '1.223', 'grad_norm': '12.56', 'learning_rate': '2.567e-06', 'num_tokens': '2.832e+06', 'mean_token_accuracy': '0.6767', 'epoch': '0.7435'}
+{'loss': '1.644', 'grad_norm': '17.62', 'learning_rate': '2.565e-06', 'num_tokens': '2.833e+06', 'mean_token_accuracy': '0.6313', 'epoch': '0.7437'}
+{'loss': '1.554', 'grad_norm': '16.38', 'learning_rate': '2.563e-06', 'num_tokens': '2.833e+06', 'mean_token_accuracy': '0.6966', 'epoch': '0.7438'}
+{'loss': '1.406', 'grad_norm': '11.81', 'learning_rate': '2.562e-06', 'num_tokens': '2.834e+06', 'mean_token_accuracy': '0.6331', 'epoch': '0.744'}
+{'loss': '0.6887', 'grad_norm': '31.75', 'learning_rate': '2.56e-06', 'num_tokens': '2.834e+06', 'mean_token_accuracy': '0.8116', 'epoch': '0.7442'}
+{'loss': '2.062', 'grad_norm': '16.38', 'learning_rate': '2.558e-06', 'num_tokens': '2.835e+06', 'mean_token_accuracy': '0.5737', 'epoch': '0.7444'}
+{'loss': '2.039', 'grad_norm': '26.5', 'learning_rate': '2.556e-06', 'num_tokens': '2.835e+06', 'mean_token_accuracy': '0.5672', 'epoch': '0.7445'}
+{'loss': '0.9336', 'grad_norm': '12.31', 'learning_rate': '2.555e-06', 'num_tokens': '2.836e+06', 'mean_token_accuracy': '0.7887', 'epoch': '0.7447'}
+{'loss': '2.324', 'grad_norm': '18.88', 'learning_rate': '2.553e-06', 'num_tokens': '2.837e+06', 'mean_token_accuracy': '0.5425', 'epoch': '0.7449'}
+{'loss': '1.907', 'grad_norm': '17.12', 'learning_rate': '2.551e-06', 'num_tokens': '2.837e+06', 'mean_token_accuracy': '0.5928', 'epoch': '0.7451'}
+{'loss': '0.9707', 'grad_norm': '14.38', 'learning_rate': '2.549e-06', 'num_tokens': '2.838e+06', 'mean_token_accuracy': '0.7886', 'epoch': '0.7452'}
+{'loss': '1.1', 'grad_norm': '13.88', 'learning_rate': '2.548e-06', 'num_tokens': '2.838e+06', 'mean_token_accuracy': '0.6976', 'epoch': '0.7454'}
+{'loss': '1.346', 'grad_norm': '11.38', 'learning_rate': '2.546e-06', 'num_tokens': '2.839e+06', 'mean_token_accuracy': '0.636', 'epoch': '0.7456'}
+{'loss': '1.114', 'grad_norm': '6.469', 'learning_rate': '2.544e-06', 'num_tokens': '2.841e+06', 'mean_token_accuracy': '0.745', 'epoch': '0.7457'}
+{'loss': '0.4565', 'grad_norm': '6.344', 'learning_rate': '2.543e-06', 'num_tokens': '2.843e+06', 'mean_token_accuracy': '0.8609', 'epoch': '0.7459'}
+{'loss': '0.8027', 'grad_norm': '13.62', 'learning_rate': '2.541e-06', 'num_tokens': '2.843e+06', 'mean_token_accuracy': '0.8248', 'epoch': '0.7461'}
+{'loss': '1.268', 'grad_norm': '12.94', 'learning_rate': '2.539e-06', 'num_tokens': '2.844e+06', 'mean_token_accuracy': '0.6662', 'epoch': '0.7463'}
+{'loss': '1.312', 'grad_norm': '13.19', 'learning_rate': '2.537e-06', 'num_tokens': '2.845e+06', 'mean_token_accuracy': '0.6318', 'epoch': '0.7464'}
+{'loss': '1.338', 'grad_norm': '13.12', 'learning_rate': '2.536e-06', 'num_tokens': '2.846e+06', 'mean_token_accuracy': '0.6344', 'epoch': '0.7466'}
+{'loss': '1.687', 'grad_norm': '17.75', 'learning_rate': '2.534e-06', 'num_tokens': '2.846e+06', 'mean_token_accuracy': '0.6735', 'epoch': '0.7468'}
+{'loss': '1.13', 'grad_norm': '25.75', 'learning_rate': '2.532e-06', 'num_tokens': '2.846e+06', 'mean_token_accuracy': '0.7895', 'epoch': '0.747'}
+{'loss': '1.419', 'grad_norm': '22.62', 'learning_rate': '2.53e-06', 'num_tokens': '2.846e+06', 'mean_token_accuracy': '0.6858', 'epoch': '0.7471'}
+{'loss': '1.551', 'grad_norm': '13.56', 'learning_rate': '2.529e-06', 'num_tokens': '2.847e+06', 'mean_token_accuracy': '0.6765', 'epoch': '0.7473'}
+{'loss': '1.87', 'grad_norm': '18.5', 'learning_rate': '2.527e-06', 'num_tokens': '2.848e+06', 'mean_token_accuracy': '0.6081', 'epoch': '0.7475'}
+{'loss': '1.096', 'grad_norm': '12.56', 'learning_rate': '2.525e-06', 'num_tokens': '2.848e+06', 'mean_token_accuracy': '0.7139', 'epoch': '0.7477'}
+{'loss': '2.175', 'grad_norm': '15.19', 'learning_rate': '2.523e-06', 'num_tokens': '2.849e+06', 'mean_token_accuracy': '0.5698', 'epoch': '0.7478'}
+{'loss': '0.6602', 'grad_norm': '7.688', 'learning_rate': '2.522e-06', 'num_tokens': '2.85e+06', 'mean_token_accuracy': '0.8074', 'epoch': '0.748'}
+{'loss': '1.53', 'grad_norm': '15.31', 'learning_rate': '2.52e-06', 'num_tokens': '2.851e+06', 'mean_token_accuracy': '0.6394', 'epoch': '0.7482'}
+{'loss': '1.097', 'grad_norm': '12.38', 'learning_rate': '2.518e-06', 'num_tokens': '2.851e+06', 'mean_token_accuracy': '0.6979', 'epoch': '0.7484'}
+{'loss': '1.402', 'grad_norm': '6.375', 'learning_rate': '2.516e-06', 'num_tokens': '2.854e+06', 'mean_token_accuracy': '0.7062', 'epoch': '0.7485'}
+{'loss': '1.657', 'grad_norm': '10.44', 'learning_rate': '2.515e-06', 'num_tokens': '2.855e+06', 'mean_token_accuracy': '0.6297', 'epoch': '0.7487'}
+{'loss': '1.321', 'grad_norm': '14.44', 'learning_rate': '2.513e-06', 'num_tokens': '2.856e+06', 'mean_token_accuracy': '0.6465', 'epoch': '0.7489'}
+{'loss': '0.7104', 'grad_norm': '46.25', 'learning_rate': '2.511e-06', 'num_tokens': '2.856e+06', 'mean_token_accuracy': '0.8493', 'epoch': '0.749'}
+{'loss': '0.4317', 'grad_norm': '25.5', 'learning_rate': '2.51e-06', 'num_tokens': '2.856e+06', 'mean_token_accuracy': '0.8778', 'epoch': '0.7492'}
+{'loss': '1.44', 'grad_norm': '11.94', 'learning_rate': '2.508e-06', 'num_tokens': '2.857e+06', 'mean_token_accuracy': '0.6961', 'epoch': '0.7494'}
+{'loss': '1.781', 'grad_norm': '14.88', 'learning_rate': '2.506e-06', 'num_tokens': '2.857e+06', 'mean_token_accuracy': '0.6327', 'epoch': '0.7496'}
+{'loss': '1.107', 'grad_norm': '11.88', 'learning_rate': '2.504e-06', 'num_tokens': '2.858e+06', 'mean_token_accuracy': '0.7019', 'epoch': '0.7497'}
+{'loss': '1.975', 'grad_norm': '12.5', 'learning_rate': '2.503e-06', 'num_tokens': '2.859e+06', 'mean_token_accuracy': '0.5931', 'epoch': '0.7499'}
+{'loss': '1.496', 'grad_norm': '14.12', 'learning_rate': '2.501e-06', 'num_tokens': '2.859e+06', 'mean_token_accuracy': '0.6124', 'epoch': '0.7501'}
+{'loss': '2.043', 'grad_norm': '16.5', 'learning_rate': '2.499e-06', 'num_tokens': '2.86e+06', 'mean_token_accuracy': '0.5874', 'epoch': '0.7503'}
+{'loss': '1.111', 'grad_norm': '14.25', 'learning_rate': '2.497e-06', 'num_tokens': '2.861e+06', 'mean_token_accuracy': '0.7315', 'epoch': '0.7504'}
+{'loss': '0.7402', 'grad_norm': '24.38', 'learning_rate': '2.496e-06', 'num_tokens': '2.861e+06', 'mean_token_accuracy': '0.8071', 'epoch': '0.7506'}
+{'loss': '1.302', 'grad_norm': '14.12', 'learning_rate': '2.494e-06', 'num_tokens': '2.862e+06', 'mean_token_accuracy': '0.6584', 'epoch': '0.7508'}
+{'loss': '1.267', 'grad_norm': '13.5', 'learning_rate': '2.492e-06', 'num_tokens': '2.862e+06', 'mean_token_accuracy': '0.6608', 'epoch': '0.751'}
+{'loss': '1.271', 'grad_norm': '11.31', 'learning_rate': '2.49e-06', 'num_tokens': '2.863e+06', 'mean_token_accuracy': '0.6412', 'epoch': '0.7511'}
+{'loss': '1.822', 'grad_norm': '15.25', 'learning_rate': '2.489e-06', 'num_tokens': '2.864e+06', 'mean_token_accuracy': '0.6183', 'epoch': '0.7513'}
+{'loss': '2.3', 'grad_norm': '19', 'learning_rate': '2.487e-06', 'num_tokens': '2.864e+06', 'mean_token_accuracy': '0.5556', 'epoch': '0.7515'}
+{'loss': '0.8152', 'grad_norm': '31.75', 'learning_rate': '2.485e-06', 'num_tokens': '2.864e+06', 'mean_token_accuracy': '0.8302', 'epoch': '0.7516'}
+{'loss': '1.161', 'grad_norm': '10.81', 'learning_rate': '2.484e-06', 'num_tokens': '2.865e+06', 'mean_token_accuracy': '0.6864', 'epoch': '0.7518'}
+{'loss': '1.631', 'grad_norm': '15.38', 'learning_rate': '2.482e-06', 'num_tokens': '2.866e+06', 'mean_token_accuracy': '0.5981', 'epoch': '0.752'}
+
 75%|███████▌  | 4331/5758 [24:38<07:56,  2.99it/s]
 75%|███████▌  | 4332/5758 [24:39<07:48,  3.05it/s]
                                                   

 75%|███████▌  | 4332/5758 [24:39<07:48,  3.05it/s]
 75%|███████▌  | 4333/5758 [24:39<08:12,  2.89it/s]
                                                   

 75%|███████▌  | 4333/5758 [24:39<08:12,  2.89it/s]
 75%|███████▌  | 4334/5758 [24:39<07:41,  3.08it/s]
                                                   

 75%|███████▌  | 4334/5758 [24:39<07:41,  3.08it/s]
 75%|███████▌  | 4335/5758 [24:40<07:36,  3.12it/s]
                                                   

 75%|███████▌  | 4335/5758 [24:40<07:36,  3.12it/s]
 75%|███████▌  | 4336/5758 [24:40<07:27,  3.18it/s]
                                                   

 75%|███████▌  | 4336/5758 [24:40<07:27,  3.18it/s]
 75%|███████▌  | 4337/5758 [24:40<07:31,  3.15it/s]
                                                   

 75%|███████▌  | 4337/5758 [24:40<07:31,  3.15it/s]
 75%|███████▌  | 4338/5758 [24:40<07:37,  3.10it/s]
                                                   

 75%|███████▌  | 4338/5758 [24:40<07:37,  3.10it/s]
 75%|███████▌  | 4339/5758 [24:41<07:31,  3.15it/s]
                                                   

 75%|███████▌  | 4339/5758 [24:41<07:31,  3.15it/s]
 75%|███████▌  | 4340/5758 [24:41<07:30,  3.15it/s]
                                                   

 75%|███████▌  | 4340/5758 [24:41<07:30,  3.15it/s]
 75%|███████▌  | 4341/5758 [24:41<07:33,  3.12it/s]
                                                   

 75%|███████▌  | 4341/5758 [24:41<07:33,  3.12it/s]
 75%|███████▌  | 4342/5758 [24:42<07:29,  3.15it/s]
                                                   

 75%|███████▌  | 4342/5758 [24:42<07:29,  3.15it/s]
 75%|███████▌  | 4343/5758 [24:42<07:27,  3.17it/s]
                                                   

 75%|███████▌  | 4343/5758 [24:42<07:27,  3.17it/s]
 75%|███████▌  | 4344/5758 [24:42<07:25,  3.17it/s]
                                                   

 75%|███████▌  | 4344/5758 [24:42<07:25,  3.17it/s]
 75%|███████▌  | 4345/5758 [24:43<07:31,  3.13it/s]
                                                   

 75%|███████▌  | 4345/5758 [24:43<07:31,  3.13it/s]
 75%|███████▌  | 4346/5758 [24:43<07:40,  3.07it/s]
                                                   

 75%|███████▌  | 4346/5758 [24:43<07:40,  3.07it/s]
 75%|███████▌  | 4347/5758 [24:44<08:39,  2.72it/s]
                                                   

 75%|███████▌  | 4347/5758 [24:44<08:39,  2.72it/s]
 76%|███████▌  | 4348/5758 [24:44<08:14,  2.85it/s]
                                                   

 76%|███████▌  | 4348/5758 [24:44<08:14,  2.85it/s]
 76%|███████▌  | 4349/5758 [24:44<07:49,  3.00it/s]
                                                   

 76%|███████▌  | 4349/5758 [24:44<07:49,  3.00it/s]
 76%|███████▌  | 4350/5758 [24:44<07:29,  3.13it/s]
                                                   

 76%|███████▌  | 4350/5758 [24:44<07:29,  3.13it/s]
 76%|███████▌  | 4351/5758 [24:45<07:19,  3.20it/s]
                                                   

 76%|███████▌  | 4351/5758 [24:45<07:19,  3.20it/s]
 76%|███████▌  | 4352/5758 [24:45<07:45,  3.02it/s]
                                                   

 76%|███████▌  | 4352/5758 [24:45<07:45,  3.02it/s]
 76%|███████▌  | 4353/5758 [24:45<07:43,  3.03it/s]
                                                   

 76%|███████▌  | 4353/5758 [24:45<07:43,  3.03it/s]
 76%|███████▌  | 4354/5758 [24:46<07:29,  3.13it/s]
                                                   

 76%|███████▌  | 4354/5758 [24:46<07:29,  3.13it/s]
 76%|███████▌  | 4355/5758 [24:46<07:55,  2.95it/s]
                                                   

 76%|███████▌  | 4355/5758 [24:46<07:55,  2.95it/s]
 76%|███████▌  | 4356/5758 [24:46<07:50,  2.98it/s]
                                                   

 76%|███████▌  | 4356/5758 [24:46<07:50,  2.98it/s]
 76%|███████▌  | 4357/5758 [24:47<07:31,  3.10it/s]
                                                   

 76%|███████▌  | 4357/5758 [24:47<07:31,  3.10it/s]
 76%|███████▌  | 4358/5758 [24:47<07:41,  3.03it/s]
                                                   

 76%|███████▌  | 4358/5758 [24:47<07:41,  3.03it/s]
 76%|███████▌  | 4359/5758 [24:47<07:31,  3.10it/s]
                                                   

 76%|███████▌  | 4359/5758 [24:47<07:31,  3.10it/s]
 76%|███████▌  | 4360/5758 [24:48<07:57,  2.93it/s]
                                                   

 76%|███████▌  | 4360/5758 [24:48<07:57,  2.93it/s]
 76%|███████▌  | 4361/5758 [24:48<07:36,  3.06it/s]
                                                   

 76%|███████▌  | 4361/5758 [24:48<07:36,  3.06it/s]
 76%|███████▌  | 4362/5758 [24:48<07:54,  2.94it/s]
                                                   

 76%|███████▌  | 4362/5758 [24:48<07:54,  2.94it/s]
 76%|███████▌  | 4363/5758 [24:49<07:55,  2.93it/s]
                                                   

 76%|███████▌  | 4363/5758 [24:49<07:55,  2.93it/s]
 76%|███████▌  | 4364/5758 [24:49<08:16,  2.81it/s]
                                                   

 76%|███████▌  | 4364/5758 [24:49<08:16,  2.81it/s]
 76%|███████▌  | 4365/5758 [24:50<08:34,  2.71it/s]
                                                   

 76%|███████▌  | 4365/5758 [24:50<08:34,  2.71it/s]
 76%|███████▌  | 4366/5758 [24:50<08:20,  2.78it/s]
                                                   

 76%|███████▌  | 4366/5758 [24:50<08:20,  2.78it/s]
 76%|███████▌  | 4367/5758 [24:50<08:03,  2.88it/s]
                                                   

 76%|███████▌  | 4367/5758 [24:50<08:03,  2.88it/s]
 76%|███████▌  | 4368/5758 [24:51<07:47,  2.97it/s]
                                                   

 76%|███████▌  | 4368/5758 [24:51<07:47,  2.97it/s]
 76%|███████▌  | 4369/5758 [24:51<07:30,  3.08it/s]
                                                   

 76%|███████▌  | 4369/5758 [24:51<07:30,  3.08it/s]
 76%|███████▌  | 4370/5758 [24:51<07:32,  3.07it/s]
                                                   

 76%|███████▌  | 4370/5758 [24:51<07:32,  3.07it/s]
 76%|███████▌  | 4371/5758 [24:52<07:50,  2.94it/s]
                                                   

 76%|███████▌  | 4371/5758 [24:52<07:50,  2.94it/s]
 76%|███████▌  | 4372/5758 [24:52<07:45,  2.98it/s]
                                                   

 76%|███████▌  | 4372/5758 [24:52<07:45,  2.98it/s]
 76%|███████▌  | 4373/5758 [24:52<08:50,  2.61it/s]
                                                   

 76%|███████▌  | 4373/5758 [24:52<08:50,  2.61it/s]
 76%|███████▌  | 4374/5758 [24:53<08:36,  2.68it/s]
                                                   

 76%|███████▌  | 4374/5758 [24:53<08:36,  2.68it/s]
 76%|███████▌  | 4375/5758 [24:53<08:27,  2.72it/s]
                                                   

 76%|███████▌  | 4375/5758 [24:53<08:27,  2.72it/s]
 76%|███████▌  | 4376/5758 [24:53<08:06,  2.84it/s]
                                                   

 76%|███████▌  | 4376/5758 [24:53<08:06,  2.84it/s]
 76%|███████▌  | 4377/5758 [24:54<08:03,  2.86it/s]
                                                   

 76%|███████▌  | 4377/5758 [24:54<08:03,  2.86it/s]
 76%|███████▌  | 4378/5758 [24:54<08:07,  2.83it/s]
                                                   

 76%|███████▌  | 4378/5758 [24:54<08:07,  2.83it/s]
 76%|███████▌  | 4379/5758 [24:54<07:54,  2.91it/s]
                                                   

 76%|███████▌  | 4379/5758 [24:54<07:54,  2.91it/s]
 76%|███████▌  | 4380/5758 [24:55<08:00,  2.87it/s]
                                                   

 76%|███████▌  | 4380/5758 [24:55<08:00,  2.87it/s]
 76%|███████▌  | 4381/5758 [24:55<07:53,  2.91it/s]
                                                   

 76%|███████▌  | 4381/5758 [24:55<07:53,  2.91it/s]
 76%|███████▌  | 4382/5758 [24:55<08:10,  2.80it/s]
                                                   

 76%|███████▌  | 4382/5758 [24:55<08:10,  2.80it/s]
 76%|███████▌  | 4383/5758 [24:56<08:23,  2.73it/s]
                                                   

 76%|███████▌  | 4383/5758 [24:56<08:23,  2.73it/s]
 76%|███████▌  | 4384/5758 [24:56<07:56,  2.88it/s]
                                                   

 76%|███████▌  | 4384/5758 [24:56<07:56,  2.88it/s]
 76%|███████▌  | 4385/5758 [24:56<07:59,  2.87it/s]
                                                   

 76%|███████▌  | 4385/5758 [24:56<07:59,  2.87it/s]
 76%|███████▌  | 4386/5758 [24:57<08:01,  2.85it/s]
                                                   
{'loss': '1.271', 'grad_norm': '12.81', 'learning_rate': '2.48e-06', 'num_tokens': '2.866e+06', 'mean_token_accuracy': '0.6618', 'epoch': '0.7522'}
+{'loss': '1.195', 'grad_norm': '8.75', 'learning_rate': '2.478e-06', 'num_tokens': '2.868e+06', 'mean_token_accuracy': '0.7187', 'epoch': '0.7523'}
+{'loss': '1.676', 'grad_norm': '19.25', 'learning_rate': '2.477e-06', 'num_tokens': '2.868e+06', 'mean_token_accuracy': '0.6493', 'epoch': '0.7525'}
+{'loss': '0.7495', 'grad_norm': '44.25', 'learning_rate': '2.475e-06', 'num_tokens': '2.868e+06', 'mean_token_accuracy': '0.8305', 'epoch': '0.7527'}
+{'loss': '1.171', 'grad_norm': '16.62', 'learning_rate': '2.473e-06', 'num_tokens': '2.869e+06', 'mean_token_accuracy': '0.7363', 'epoch': '0.7529'}
+{'loss': '1.464', 'grad_norm': '12.75', 'learning_rate': '2.471e-06', 'num_tokens': '2.869e+06', 'mean_token_accuracy': '0.686', 'epoch': '0.753'}
+{'loss': '1.941', 'grad_norm': '21.38', 'learning_rate': '2.47e-06', 'num_tokens': '2.87e+06', 'mean_token_accuracy': '0.5784', 'epoch': '0.7532'}
+{'loss': '0.5924', 'grad_norm': '6.906', 'learning_rate': '2.468e-06', 'num_tokens': '2.871e+06', 'mean_token_accuracy': '0.8154', 'epoch': '0.7534'}
+{'loss': '1.13', 'grad_norm': '13.31', 'learning_rate': '2.466e-06', 'num_tokens': '2.872e+06', 'mean_token_accuracy': '0.6791', 'epoch': '0.7536'}
+{'loss': '0.6215', 'grad_norm': '19.5', 'learning_rate': '2.464e-06', 'num_tokens': '2.872e+06', 'mean_token_accuracy': '0.8686', 'epoch': '0.7537'}
+{'loss': '0.5146', 'grad_norm': '7.188', 'learning_rate': '2.463e-06', 'num_tokens': '2.874e+06', 'mean_token_accuracy': '0.8441', 'epoch': '0.7539'}
+{'loss': '0.8302', 'grad_norm': '38.75', 'learning_rate': '2.461e-06', 'num_tokens': '2.874e+06', 'mean_token_accuracy': '0.7125', 'epoch': '0.7541'}
+{'loss': '1.25', 'grad_norm': '14', 'learning_rate': '2.459e-06', 'num_tokens': '2.874e+06', 'mean_token_accuracy': '0.6511', 'epoch': '0.7543'}
+{'loss': '1.084', 'grad_norm': '11.5', 'learning_rate': '2.457e-06', 'num_tokens': '2.875e+06', 'mean_token_accuracy': '0.7004', 'epoch': '0.7544'}
+{'loss': '1.516', 'grad_norm': '7.625', 'learning_rate': '2.456e-06', 'num_tokens': '2.877e+06', 'mean_token_accuracy': '0.6721', 'epoch': '0.7546'}
+{'loss': '1.303', 'grad_norm': '11.69', 'learning_rate': '2.454e-06', 'num_tokens': '2.878e+06', 'mean_token_accuracy': '0.6515', 'epoch': '0.7548'}
+{'loss': '1.036', 'grad_norm': '13.69', 'learning_rate': '2.452e-06', 'num_tokens': '2.878e+06', 'mean_token_accuracy': '0.7778', 'epoch': '0.7549'}
+{'loss': '1.702', 'grad_norm': '12.25', 'learning_rate': '2.451e-06', 'num_tokens': '2.879e+06', 'mean_token_accuracy': '0.6408', 'epoch': '0.7551'}
+{'loss': '1.238', 'grad_norm': '13.75', 'learning_rate': '2.449e-06', 'num_tokens': '2.879e+06', 'mean_token_accuracy': '0.6705', 'epoch': '0.7553'}
+{'loss': '0.4854', 'grad_norm': '16.5', 'learning_rate': '2.447e-06', 'num_tokens': '2.88e+06', 'mean_token_accuracy': '0.8964', 'epoch': '0.7555'}
+{'loss': '1.388', 'grad_norm': '9.688', 'learning_rate': '2.445e-06', 'num_tokens': '2.881e+06', 'mean_token_accuracy': '0.6382', 'epoch': '0.7556'}
+{'loss': '1.447', 'grad_norm': '10.44', 'learning_rate': '2.444e-06', 'num_tokens': '2.882e+06', 'mean_token_accuracy': '0.6052', 'epoch': '0.7558'}
+{'loss': '1.886', 'grad_norm': '17.25', 'learning_rate': '2.442e-06', 'num_tokens': '2.883e+06', 'mean_token_accuracy': '0.5791', 'epoch': '0.756'}
+{'loss': '1.473', 'grad_norm': '13.75', 'learning_rate': '2.44e-06', 'num_tokens': '2.883e+06', 'mean_token_accuracy': '0.6148', 'epoch': '0.7562'}
+{'loss': '0.7296', 'grad_norm': '9.5', 'learning_rate': '2.438e-06', 'num_tokens': '2.884e+06', 'mean_token_accuracy': '0.7774', 'epoch': '0.7563'}
+{'loss': '1.001', 'grad_norm': '24', 'learning_rate': '2.437e-06', 'num_tokens': '2.885e+06', 'mean_token_accuracy': '0.744', 'epoch': '0.7565'}
+{'loss': '1.973', 'grad_norm': '15.69', 'learning_rate': '2.435e-06', 'num_tokens': '2.885e+06', 'mean_token_accuracy': '0.5848', 'epoch': '0.7567'}
+{'loss': '2.129', 'grad_norm': '13.94', 'learning_rate': '2.433e-06', 'num_tokens': '2.886e+06', 'mean_token_accuracy': '0.5617', 'epoch': '0.7569'}
+{'loss': '0.7522', 'grad_norm': '8.562', 'learning_rate': '2.431e-06', 'num_tokens': '2.887e+06', 'mean_token_accuracy': '0.7685', 'epoch': '0.757'}
+{'loss': '0.9961', 'grad_norm': '13.31', 'learning_rate': '2.43e-06', 'num_tokens': '2.888e+06', 'mean_token_accuracy': '0.7436', 'epoch': '0.7572'}
+{'loss': '0.682', 'grad_norm': '22.38', 'learning_rate': '2.428e-06', 'num_tokens': '2.888e+06', 'mean_token_accuracy': '0.8344', 'epoch': '0.7574'}
+{'loss': '1.192', 'grad_norm': '14.38', 'learning_rate': '2.426e-06', 'num_tokens': '2.888e+06', 'mean_token_accuracy': '0.6832', 'epoch': '0.7576'}
+{'loss': '2.114', 'grad_norm': '16', 'learning_rate': '2.424e-06', 'num_tokens': '2.889e+06', 'mean_token_accuracy': '0.5986', 'epoch': '0.7577'}
+{'loss': '1.59', 'grad_norm': '6.406', 'learning_rate': '2.423e-06', 'num_tokens': '2.892e+06', 'mean_token_accuracy': '0.6657', 'epoch': '0.7579'}
+{'loss': '1.269', 'grad_norm': '14', 'learning_rate': '2.421e-06', 'num_tokens': '2.892e+06', 'mean_token_accuracy': '0.7186', 'epoch': '0.7581'}
+{'loss': '0.5347', 'grad_norm': '6.562', 'learning_rate': '2.419e-06', 'num_tokens': '2.894e+06', 'mean_token_accuracy': '0.8428', 'epoch': '0.7582'}
+{'loss': '1.185', 'grad_norm': '7.25', 'learning_rate': '2.418e-06', 'num_tokens': '2.895e+06', 'mean_token_accuracy': '0.7198', 'epoch': '0.7584'}
+{'loss': '1.128', 'grad_norm': '35', 'learning_rate': '2.416e-06', 'num_tokens': '2.896e+06', 'mean_token_accuracy': '0.7803', 'epoch': '0.7586'}
+{'loss': '1.356', 'grad_norm': '25.75', 'learning_rate': '2.414e-06', 'num_tokens': '2.896e+06', 'mean_token_accuracy': '0.6471', 'epoch': '0.7588'}
+{'loss': '0.944', 'grad_norm': '19.75', 'learning_rate': '2.412e-06', 'num_tokens': '2.896e+06', 'mean_token_accuracy': '0.795', 'epoch': '0.7589'}
+{'loss': '0.6499', 'grad_norm': '18.38', 'learning_rate': '2.411e-06', 'num_tokens': '2.896e+06', 'mean_token_accuracy': '0.849', 'epoch': '0.7591'}
+{'loss': '0.6709', 'grad_norm': '41.75', 'learning_rate': '2.409e-06', 'num_tokens': '2.896e+06', 'mean_token_accuracy': '0.878', 'epoch': '0.7593'}
+{'loss': '1.422', 'grad_norm': '13.31', 'learning_rate': '2.407e-06', 'num_tokens': '2.897e+06', 'mean_token_accuracy': '0.6314', 'epoch': '0.7595'}
+{'loss': '1.815', 'grad_norm': '21', 'learning_rate': '2.405e-06', 'num_tokens': '2.897e+06', 'mean_token_accuracy': '0.6834', 'epoch': '0.7596'}
+{'loss': '0.5535', 'grad_norm': '7', 'learning_rate': '2.404e-06', 'num_tokens': '2.899e+06', 'mean_token_accuracy': '0.8246', 'epoch': '0.7598'}
+{'loss': '1.106', 'grad_norm': '13.94', 'learning_rate': '2.402e-06', 'num_tokens': '2.9e+06', 'mean_token_accuracy': '0.7313', 'epoch': '0.76'}
+{'loss': '1.733', 'grad_norm': '24.62', 'learning_rate': '2.4e-06', 'num_tokens': '2.9e+06', 'mean_token_accuracy': '0.6123', 'epoch': '0.7602'}
+{'loss': '1.242', 'grad_norm': '5.219', 'learning_rate': '2.398e-06', 'num_tokens': '2.904e+06', 'mean_token_accuracy': '0.7238', 'epoch': '0.7603'}
+{'loss': '1.616', 'grad_norm': '28.38', 'learning_rate': '2.397e-06', 'num_tokens': '2.904e+06', 'mean_token_accuracy': '0.6633', 'epoch': '0.7605'}
+{'loss': '1.123', 'grad_norm': '15.5', 'learning_rate': '2.395e-06', 'num_tokens': '2.905e+06', 'mean_token_accuracy': '0.7115', 'epoch': '0.7607'}
+{'loss': '1.451', 'grad_norm': '12.44', 'learning_rate': '2.393e-06', 'num_tokens': '2.906e+06', 'mean_token_accuracy': '0.615', 'epoch': '0.7609'}
+{'loss': '0.4584', 'grad_norm': '7.781', 'learning_rate': '2.391e-06', 'num_tokens': '2.907e+06', 'mean_token_accuracy': '0.8565', 'epoch': '0.761'}
+{'loss': '0.6827', 'grad_norm': '17.5', 'learning_rate': '2.39e-06', 'num_tokens': '2.907e+06', 'mean_token_accuracy': '0.8253', 'epoch': '0.7612'}
+{'loss': '1.253', 'grad_norm': '13.94', 'learning_rate': '2.388e-06', 'num_tokens': '2.908e+06', 'mean_token_accuracy': '0.6745', 'epoch': '0.7614'}
+{'loss': '1.834', 'grad_norm': '7.5', 'learning_rate': '2.386e-06', 'num_tokens': '2.91e+06', 'mean_token_accuracy': '0.6309', 'epoch': '0.7615'}
+
 76%|███████▌  | 4386/5758 [24:57<08:01,  2.85it/s]
 76%|███████▌  | 4387/5758 [24:57<07:39,  2.98it/s]
                                                   

 76%|███████▌  | 4387/5758 [24:57<07:39,  2.98it/s]
 76%|███████▌  | 4388/5758 [24:58<07:49,  2.92it/s]
                                                   

 76%|███████▌  | 4388/5758 [24:58<07:49,  2.92it/s]
 76%|███████▌  | 4389/5758 [24:58<07:49,  2.92it/s]
                                                   

 76%|███████▌  | 4389/5758 [24:58<07:49,  2.92it/s]
 76%|███████▌  | 4390/5758 [24:58<08:00,  2.85it/s]
                                                   

 76%|███████▌  | 4390/5758 [24:58<08:00,  2.85it/s]
 76%|███████▋  | 4391/5758 [24:59<07:38,  2.98it/s]
                                                   

 76%|███████▋  | 4391/5758 [24:59<07:38,  2.98it/s]
 76%|███████▋  | 4392/5758 [24:59<07:22,  3.08it/s]
                                                   

 76%|███████▋  | 4392/5758 [24:59<07:22,  3.08it/s]
 76%|███████▋  | 4393/5758 [24:59<07:06,  3.20it/s]
                                                   

 76%|███████▋  | 4393/5758 [24:59<07:06,  3.20it/s]
 76%|███████▋  | 4394/5758 [24:59<07:24,  3.07it/s]
                                                   

 76%|███████▋  | 4394/5758 [24:59<07:24,  3.07it/s]
 76%|███████▋  | 4395/5758 [25:00<07:23,  3.07it/s]
                                                   

 76%|███████▋  | 4395/5758 [25:00<07:23,  3.07it/s]
 76%|███████▋  | 4396/5758 [25:00<07:44,  2.93it/s]
                                                   

 76%|███████▋  | 4396/5758 [25:00<07:44,  2.93it/s]
 76%|███████▋  | 4397/5758 [25:00<07:22,  3.07it/s]
                                                   

 76%|███████▋  | 4397/5758 [25:00<07:22,  3.07it/s]
 76%|███████▋  | 4398/5758 [25:01<07:37,  2.98it/s]
                                                   

 76%|███████▋  | 4398/5758 [25:01<07:37,  2.98it/s]
 76%|███████▋  | 4399/5758 [25:01<08:04,  2.81it/s]
                                                   

 76%|███████▋  | 4399/5758 [25:01<08:04,  2.81it/s]
 76%|███████▋  | 4400/5758 [25:02<07:43,  2.93it/s]
                                                   

 76%|███████▋  | 4400/5758 [25:02<07:43,  2.93it/s]
 76%|███████▋  | 4401/5758 [25:02<07:27,  3.03it/s]
                                                   

 76%|███████▋  | 4401/5758 [25:02<07:27,  3.03it/s]
 76%|███████▋  | 4402/5758 [25:02<07:10,  3.15it/s]
                                                   

 76%|███████▋  | 4402/5758 [25:02<07:10,  3.15it/s]
 76%|███████▋  | 4403/5758 [25:02<07:06,  3.18it/s]
                                                   

 76%|███████▋  | 4403/5758 [25:02<07:06,  3.18it/s]
 76%|███████▋  | 4404/5758 [25:03<07:42,  2.93it/s]
                                                   

 76%|███████▋  | 4404/5758 [25:03<07:42,  2.93it/s]
 77%|███████▋  | 4405/5758 [25:03<07:35,  2.97it/s]
                                                   

 77%|███████▋  | 4405/5758 [25:03<07:35,  2.97it/s]
 77%|███████▋  | 4406/5758 [25:03<07:29,  3.01it/s]
                                                   

 77%|███████▋  | 4406/5758 [25:03<07:29,  3.01it/s]
 77%|███████▋  | 4407/5758 [25:04<07:22,  3.05it/s]
                                                   

 77%|███████▋  | 4407/5758 [25:04<07:22,  3.05it/s]
 77%|███████▋  | 4408/5758 [25:04<07:33,  2.97it/s]
                                                   

 77%|███████▋  | 4408/5758 [25:04<07:33,  2.97it/s]
 77%|███████▋  | 4409/5758 [25:04<07:22,  3.05it/s]
                                                   

 77%|███████▋  | 4409/5758 [25:04<07:22,  3.05it/s]
 77%|███████▋  | 4410/5758 [25:05<07:24,  3.03it/s]
                                                   

 77%|███████▋  | 4410/5758 [25:05<07:24,  3.03it/s]
 77%|███████▋  | 4411/5758 [25:05<07:14,  3.10it/s]
                                                   

 77%|███████▋  | 4411/5758 [25:05<07:14,  3.10it/s]
 77%|███████▋  | 4412/5758 [25:05<07:03,  3.18it/s]
                                                   

 77%|███████▋  | 4412/5758 [25:05<07:03,  3.18it/s]
 77%|███████▋  | 4413/5758 [25:06<07:27,  3.00it/s]
                                                   

 77%|███████▋  | 4413/5758 [25:06<07:27,  3.00it/s]
 77%|███████▋  | 4414/5758 [25:06<07:13,  3.10it/s]
                                                   

 77%|███████▋  | 4414/5758 [25:06<07:13,  3.10it/s]
 77%|███████▋  | 4415/5758 [25:06<07:15,  3.08it/s]
                                                   

 77%|███████▋  | 4415/5758 [25:06<07:15,  3.08it/s]
 77%|███████▋  | 4416/5758 [25:07<07:21,  3.04it/s]
                                                   

 77%|███████▋  | 4416/5758 [25:07<07:21,  3.04it/s]
 77%|███████▋  | 4417/5758 [25:07<07:14,  3.08it/s]
                                                   

 77%|███████▋  | 4417/5758 [25:07<07:14,  3.08it/s]
 77%|███████▋  | 4418/5758 [25:07<07:13,  3.09it/s]
                                                   

 77%|███████▋  | 4418/5758 [25:07<07:13,  3.09it/s]
 77%|███████▋  | 4419/5758 [25:08<07:11,  3.10it/s]
                                                   

 77%|███████▋  | 4419/5758 [25:08<07:11,  3.10it/s]
 77%|███████▋  | 4420/5758 [25:08<07:01,  3.18it/s]
                                                   

 77%|███████▋  | 4420/5758 [25:08<07:01,  3.18it/s]
 77%|███████▋  | 4421/5758 [25:08<06:57,  3.20it/s]
                                                   

 77%|███████▋  | 4421/5758 [25:08<06:57,  3.20it/s]
 77%|███████▋  | 4422/5758 [25:09<07:22,  3.02it/s]
                                                   

 77%|███████▋  | 4422/5758 [25:09<07:22,  3.02it/s]
 77%|███████▋  | 4423/5758 [25:09<07:35,  2.93it/s]
                                                   

 77%|███████▋  | 4423/5758 [25:09<07:35,  2.93it/s]
 77%|███████▋  | 4424/5758 [25:09<07:19,  3.03it/s]
                                                   

 77%|███████▋  | 4424/5758 [25:09<07:19,  3.03it/s]
 77%|███████▋  | 4425/5758 [25:10<08:15,  2.69it/s]
                                                   

 77%|███████▋  | 4425/5758 [25:10<08:15,  2.69it/s]
 77%|███████▋  | 4426/5758 [25:10<08:07,  2.73it/s]
                                                   

 77%|███████▋  | 4426/5758 [25:10<08:07,  2.73it/s]
 77%|███████▋  | 4427/5758 [25:10<07:50,  2.83it/s]
                                                   

 77%|███████▋  | 4427/5758 [25:10<07:50,  2.83it/s]
 77%|███████▋  | 4428/5758 [25:11<07:57,  2.78it/s]
                                                   

 77%|███████▋  | 4428/5758 [25:11<07:57,  2.78it/s]
 77%|███████▋  | 4429/5758 [25:11<08:15,  2.68it/s]
                                                   

 77%|███████▋  | 4429/5758 [25:11<08:15,  2.68it/s]
 77%|███████▋  | 4430/5758 [25:12<08:02,  2.75it/s]
                                                   

 77%|███████▋  | 4430/5758 [25:12<08:02,  2.75it/s]
 77%|███████▋  | 4431/5758 [25:12<07:53,  2.80it/s]
                                                   

 77%|███████▋  | 4431/5758 [25:12<07:53,  2.80it/s]
 77%|███████▋  | 4432/5758 [25:12<08:04,  2.74it/s]
                                                   

 77%|███████▋  | 4432/5758 [25:12<08:04,  2.74it/s]
 77%|███████▋  | 4433/5758 [25:13<08:13,  2.68it/s]
                                                   

 77%|███████▋  | 4433/5758 [25:13<08:13,  2.68it/s]
 77%|███████▋  | 4434/5758 [25:13<07:56,  2.78it/s]
                                                   

 77%|███████▋  | 4434/5758 [25:13<07:56,  2.78it/s]
 77%|███████▋  | 4435/5758 [25:13<07:57,  2.77it/s]
                                                   

 77%|███████▋  | 4435/5758 [25:13<07:57,  2.77it/s]
 77%|███████▋  | 4436/5758 [25:14<07:40,  2.87it/s]
                                                   

 77%|███████▋  | 4436/5758 [25:14<07:40,  2.87it/s]
 77%|███████▋  | 4437/5758 [25:14<07:22,  2.98it/s]
                                                   

 77%|███████▋  | 4437/5758 [25:14<07:22,  2.98it/s]
 77%|███████▋  | 4438/5758 [25:14<07:18,  3.01it/s]
                                                   

 77%|███████▋  | 4438/5758 [25:14<07:18,  3.01it/s]
 77%|███████▋  | 4439/5758 [25:15<07:25,  2.96it/s]
                                                   

 77%|███████▋  | 4439/5758 [25:15<07:25,  2.96it/s]
 77%|███████▋  | 4440/5758 [25:15<07:17,  3.01it/s]
                                                   

 77%|███████▋  | 4440/5758 [25:15<07:17,  3.01it/s]
 77%|███████▋  | 4441/5758 [25:15<07:03,  3.11it/s]
                                                   
{'loss': '1.755', 'grad_norm': '10.31', 'learning_rate': '2.385e-06', 'num_tokens': '2.911e+06', 'mean_token_accuracy': '0.6524', 'epoch': '0.7617'}
+{'loss': '1.455', 'grad_norm': '14.06', 'learning_rate': '2.383e-06', 'num_tokens': '2.912e+06', 'mean_token_accuracy': '0.6079', 'epoch': '0.7619'}
+{'loss': '1.317', 'grad_norm': '10.75', 'learning_rate': '2.381e-06', 'num_tokens': '2.913e+06', 'mean_token_accuracy': '0.6393', 'epoch': '0.7621'}
+{'loss': '0.6007', 'grad_norm': '6.969', 'learning_rate': '2.379e-06', 'num_tokens': '2.914e+06', 'mean_token_accuracy': '0.8251', 'epoch': '0.7622'}
+{'loss': '0.8056', 'grad_norm': '35', 'learning_rate': '2.378e-06', 'num_tokens': '2.914e+06', 'mean_token_accuracy': '0.8387', 'epoch': '0.7624'}
+{'loss': '1.406', 'grad_norm': '13.38', 'learning_rate': '2.376e-06', 'num_tokens': '2.915e+06', 'mean_token_accuracy': '0.6615', 'epoch': '0.7626'}
+{'loss': '0.933', 'grad_norm': '23.88', 'learning_rate': '2.374e-06', 'num_tokens': '2.915e+06', 'mean_token_accuracy': '0.7875', 'epoch': '0.7628'}
+{'loss': '1.954', 'grad_norm': '16.12', 'learning_rate': '2.372e-06', 'num_tokens': '2.916e+06', 'mean_token_accuracy': '0.5886', 'epoch': '0.7629'}
+{'loss': '1.845', 'grad_norm': '9.75', 'learning_rate': '2.371e-06', 'num_tokens': '2.917e+06', 'mean_token_accuracy': '0.6099', 'epoch': '0.7631'}
+{'loss': '1.992', 'grad_norm': '29.5', 'learning_rate': '2.369e-06', 'num_tokens': '2.917e+06', 'mean_token_accuracy': '0.597', 'epoch': '0.7633'}
+{'loss': '1.921', 'grad_norm': '15.38', 'learning_rate': '2.367e-06', 'num_tokens': '2.918e+06', 'mean_token_accuracy': '0.5907', 'epoch': '0.7635'}
+{'loss': '1.301', 'grad_norm': '14.88', 'learning_rate': '2.365e-06', 'num_tokens': '2.918e+06', 'mean_token_accuracy': '0.6439', 'epoch': '0.7636'}
+{'loss': '1.265', 'grad_norm': '6.938', 'learning_rate': '2.364e-06', 'num_tokens': '2.92e+06', 'mean_token_accuracy': '0.6991', 'epoch': '0.7638'}
+{'loss': '1.341', 'grad_norm': '15.44', 'learning_rate': '2.362e-06', 'num_tokens': '2.921e+06', 'mean_token_accuracy': '0.6298', 'epoch': '0.764'}
+{'loss': '1.155', 'grad_norm': '13.19', 'learning_rate': '2.36e-06', 'num_tokens': '2.922e+06', 'mean_token_accuracy': '0.6657', 'epoch': '0.7642'}
+{'loss': '2.097', 'grad_norm': '9.375', 'learning_rate': '2.358e-06', 'num_tokens': '2.923e+06', 'mean_token_accuracy': '0.5682', 'epoch': '0.7643'}
+{'loss': '1.337', 'grad_norm': '13.81', 'learning_rate': '2.357e-06', 'num_tokens': '2.924e+06', 'mean_token_accuracy': '0.6535', 'epoch': '0.7645'}
+{'loss': '0.694', 'grad_norm': '37.75', 'learning_rate': '2.355e-06', 'num_tokens': '2.924e+06', 'mean_token_accuracy': '0.8519', 'epoch': '0.7647'}
+{'loss': '1.194', 'grad_norm': '29.25', 'learning_rate': '2.353e-06', 'num_tokens': '2.924e+06', 'mean_token_accuracy': '0.7172', 'epoch': '0.7648'}
+{'loss': '1.685', 'grad_norm': '10.81', 'learning_rate': '2.352e-06', 'num_tokens': '2.925e+06', 'mean_token_accuracy': '0.6349', 'epoch': '0.765'}
+{'loss': '1.602', 'grad_norm': '14.88', 'learning_rate': '2.35e-06', 'num_tokens': '2.925e+06', 'mean_token_accuracy': '0.6503', 'epoch': '0.7652'}
+{'loss': '0.9449', 'grad_norm': '12.75', 'learning_rate': '2.348e-06', 'num_tokens': '2.926e+06', 'mean_token_accuracy': '0.7191', 'epoch': '0.7654'}
+{'loss': '0.5373', 'grad_norm': '17', 'learning_rate': '2.346e-06', 'num_tokens': '2.926e+06', 'mean_token_accuracy': '0.876', 'epoch': '0.7655'}
+{'loss': '1.237', 'grad_norm': '13', 'learning_rate': '2.345e-06', 'num_tokens': '2.927e+06', 'mean_token_accuracy': '0.6768', 'epoch': '0.7657'}
+{'loss': '1.31', 'grad_norm': '11.25', 'learning_rate': '2.343e-06', 'num_tokens': '2.928e+06', 'mean_token_accuracy': '0.6521', 'epoch': '0.7659'}
+{'loss': '1.949', 'grad_norm': '13', 'learning_rate': '2.341e-06', 'num_tokens': '2.929e+06', 'mean_token_accuracy': '0.5587', 'epoch': '0.7661'}
+{'loss': '1.451', 'grad_norm': '11', 'learning_rate': '2.339e-06', 'num_tokens': '2.929e+06', 'mean_token_accuracy': '0.6758', 'epoch': '0.7662'}
+{'loss': '1.189', 'grad_norm': '14.81', 'learning_rate': '2.338e-06', 'num_tokens': '2.93e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.7664'}
+{'loss': '0.583', 'grad_norm': '7.656', 'learning_rate': '2.336e-06', 'num_tokens': '2.931e+06', 'mean_token_accuracy': '0.822', 'epoch': '0.7666'}
+{'loss': '0.6672', 'grad_norm': '21.25', 'learning_rate': '2.334e-06', 'num_tokens': '2.931e+06', 'mean_token_accuracy': '0.8404', 'epoch': '0.7668'}
+{'loss': '2.02', 'grad_norm': '16', 'learning_rate': '2.332e-06', 'num_tokens': '2.932e+06', 'mean_token_accuracy': '0.5743', 'epoch': '0.7669'}
+{'loss': '1.932', 'grad_norm': '12.5', 'learning_rate': '2.331e-06', 'num_tokens': '2.933e+06', 'mean_token_accuracy': '0.608', 'epoch': '0.7671'}
+{'loss': '1.312', 'grad_norm': '12.06', 'learning_rate': '2.329e-06', 'num_tokens': '2.933e+06', 'mean_token_accuracy': '0.6427', 'epoch': '0.7673'}
+{'loss': '1.063', 'grad_norm': '18', 'learning_rate': '2.327e-06', 'num_tokens': '2.933e+06', 'mean_token_accuracy': '0.7742', 'epoch': '0.7675'}
+{'loss': '1.241', 'grad_norm': '12.56', 'learning_rate': '2.325e-06', 'num_tokens': '2.934e+06', 'mean_token_accuracy': '0.6627', 'epoch': '0.7676'}
+{'loss': '1.449', 'grad_norm': '10.06', 'learning_rate': '2.324e-06', 'num_tokens': '2.936e+06', 'mean_token_accuracy': '0.665', 'epoch': '0.7678'}
+{'loss': '1.871', 'grad_norm': '17.75', 'learning_rate': '2.322e-06', 'num_tokens': '2.936e+06', 'mean_token_accuracy': '0.6174', 'epoch': '0.768'}
+{'loss': '0.5832', 'grad_norm': '18.38', 'learning_rate': '2.32e-06', 'num_tokens': '2.936e+06', 'mean_token_accuracy': '0.834', 'epoch': '0.7681'}
+{'loss': '0.5792', 'grad_norm': '7.938', 'learning_rate': '2.319e-06', 'num_tokens': '2.938e+06', 'mean_token_accuracy': '0.8192', 'epoch': '0.7683'}
+{'loss': '1.462', 'grad_norm': '24.5', 'learning_rate': '2.317e-06', 'num_tokens': '2.938e+06', 'mean_token_accuracy': '0.6244', 'epoch': '0.7685'}
+{'loss': '1.499', 'grad_norm': '36.5', 'learning_rate': '2.315e-06', 'num_tokens': '2.938e+06', 'mean_token_accuracy': '0.7264', 'epoch': '0.7687'}
+{'loss': '0.5782', 'grad_norm': '7.344', 'learning_rate': '2.313e-06', 'num_tokens': '2.94e+06', 'mean_token_accuracy': '0.8296', 'epoch': '0.7688'}
+{'loss': '2.28', 'grad_norm': '24', 'learning_rate': '2.312e-06', 'num_tokens': '2.94e+06', 'mean_token_accuracy': '0.5418', 'epoch': '0.769'}
+{'loss': '1.077', 'grad_norm': '18', 'learning_rate': '2.31e-06', 'num_tokens': '2.94e+06', 'mean_token_accuracy': '0.7705', 'epoch': '0.7692'}
+{'loss': '1.368', 'grad_norm': '44.5', 'learning_rate': '2.308e-06', 'num_tokens': '2.94e+06', 'mean_token_accuracy': '0.6806', 'epoch': '0.7694'}
+{'loss': '0.8532', 'grad_norm': '12.69', 'learning_rate': '2.306e-06', 'num_tokens': '2.941e+06', 'mean_token_accuracy': '0.7327', 'epoch': '0.7695'}
+{'loss': '0.8817', 'grad_norm': '23.88', 'learning_rate': '2.305e-06', 'num_tokens': '2.941e+06', 'mean_token_accuracy': '0.8', 'epoch': '0.7697'}
+{'loss': '1.165', 'grad_norm': '10.56', 'learning_rate': '2.303e-06', 'num_tokens': '2.942e+06', 'mean_token_accuracy': '0.676', 'epoch': '0.7699'}
+{'loss': '1.933', 'grad_norm': '28.38', 'learning_rate': '2.301e-06', 'num_tokens': '2.942e+06', 'mean_token_accuracy': '0.6378', 'epoch': '0.7701'}
+{'loss': '1.186', 'grad_norm': '18.38', 'learning_rate': '2.299e-06', 'num_tokens': '2.942e+06', 'mean_token_accuracy': '0.7409', 'epoch': '0.7702'}
+{'loss': '1.238', 'grad_norm': '12.81', 'learning_rate': '2.298e-06', 'num_tokens': '2.943e+06', 'mean_token_accuracy': '0.6616', 'epoch': '0.7704'}
+{'loss': '1.112', 'grad_norm': '13.38', 'learning_rate': '2.296e-06', 'num_tokens': '2.944e+06', 'mean_token_accuracy': '0.7067', 'epoch': '0.7706'}
+{'loss': '1.648', 'grad_norm': '15.88', 'learning_rate': '2.294e-06', 'num_tokens': '2.944e+06', 'mean_token_accuracy': '0.6247', 'epoch': '0.7708'}
+{'loss': '1.956', 'grad_norm': '9.625', 'learning_rate': '2.292e-06', 'num_tokens': '2.945e+06', 'mean_token_accuracy': '0.6026', 'epoch': '0.7709'}
+{'loss': '0.6119', 'grad_norm': '14.06', 'learning_rate': '2.291e-06', 'num_tokens': '2.945e+06', 'mean_token_accuracy': '0.8538', 'epoch': '0.7711'}
+
 77%|███████▋  | 4441/5758 [25:15<07:03,  3.11it/s]
 77%|███████▋  | 4442/5758 [25:16<06:48,  3.23it/s]
                                                   

 77%|███████▋  | 4442/5758 [25:16<06:48,  3.23it/s]
 77%|███████▋  | 4443/5758 [25:16<06:41,  3.28it/s]
                                                   

 77%|███████▋  | 4443/5758 [25:16<06:41,  3.28it/s]
 77%|███████▋  | 4444/5758 [25:16<06:40,  3.28it/s]
                                                   

 77%|███████▋  | 4444/5758 [25:16<06:40,  3.28it/s]
 77%|███████▋  | 4445/5758 [25:17<06:48,  3.22it/s]
                                                   

 77%|███████▋  | 4445/5758 [25:17<06:48,  3.22it/s]
 77%|███████▋  | 4446/5758 [25:17<06:39,  3.28it/s]
                                                   

 77%|███████▋  | 4446/5758 [25:17<06:39,  3.28it/s]
 77%|███████▋  | 4447/5758 [25:17<06:53,  3.17it/s]
                                                   

 77%|███████▋  | 4447/5758 [25:17<06:53,  3.17it/s]
 77%|███████▋  | 4448/5758 [25:17<06:39,  3.28it/s]
                                                   

 77%|███████▋  | 4448/5758 [25:17<06:39,  3.28it/s]
 77%|███████▋  | 4449/5758 [25:18<06:35,  3.31it/s]
                                                   

 77%|███████▋  | 4449/5758 [25:18<06:35,  3.31it/s]
 77%|███████▋  | 4450/5758 [25:18<06:49,  3.20it/s]
                                                   

 77%|███████▋  | 4450/5758 [25:18<06:49,  3.20it/s]
 77%|███████▋  | 4451/5758 [25:18<07:10,  3.04it/s]
                                                   

 77%|███████▋  | 4451/5758 [25:18<07:10,  3.04it/s]
 77%|███████▋  | 4452/5758 [25:19<07:53,  2.76it/s]
                                                   

 77%|███████▋  | 4452/5758 [25:19<07:53,  2.76it/s]
 77%|███████▋  | 4453/5758 [25:19<07:31,  2.89it/s]
                                                   

 77%|███████▋  | 4453/5758 [25:19<07:31,  2.89it/s]
 77%|███████▋  | 4454/5758 [25:20<07:41,  2.83it/s]
                                                   

 77%|███████▋  | 4454/5758 [25:20<07:41,  2.83it/s]
 77%|███████▋  | 4455/5758 [25:20<07:44,  2.81it/s]
                                                   

 77%|███████▋  | 4455/5758 [25:20<07:44,  2.81it/s]
 77%|███████▋  | 4456/5758 [25:20<07:13,  3.00it/s]
                                                   

 77%|███████▋  | 4456/5758 [25:20<07:13,  3.00it/s]
 77%|███████▋  | 4457/5758 [25:21<07:07,  3.04it/s]
                                                   

 77%|███████▋  | 4457/5758 [25:21<07:07,  3.04it/s]
 77%|███████▋  | 4458/5758 [25:21<06:50,  3.17it/s]
                                                   

 77%|███████▋  | 4458/5758 [25:21<06:50,  3.17it/s]
 77%|███████▋  | 4459/5758 [25:21<07:25,  2.92it/s]
                                                   

 77%|███████▋  | 4459/5758 [25:21<07:25,  2.92it/s]
 77%|███████▋  | 4460/5758 [25:22<07:19,  2.95it/s]
                                                   

 77%|███████▋  | 4460/5758 [25:22<07:19,  2.95it/s]
 77%|███████▋  | 4461/5758 [25:22<07:37,  2.83it/s]
                                                   

 77%|███████▋  | 4461/5758 [25:22<07:37,  2.83it/s]
 77%|███████▋  | 4462/5758 [25:22<07:41,  2.81it/s]
                                                   

 77%|███████▋  | 4462/5758 [25:22<07:41,  2.81it/s]
 78%|███████▊  | 4463/5758 [25:23<07:27,  2.89it/s]
                                                   

 78%|███████▊  | 4463/5758 [25:23<07:27,  2.89it/s]
 78%|███████▊  | 4464/5758 [25:23<07:33,  2.85it/s]
                                                   

 78%|███████▊  | 4464/5758 [25:23<07:33,  2.85it/s]
 78%|███████▊  | 4465/5758 [25:23<07:21,  2.93it/s]
                                                   

 78%|███████▊  | 4465/5758 [25:23<07:21,  2.93it/s]
 78%|███████▊  | 4466/5758 [25:24<07:01,  3.06it/s]
                                                   

 78%|███████▊  | 4466/5758 [25:24<07:01,  3.06it/s]
 78%|███████▊  | 4467/5758 [25:24<06:52,  3.13it/s]
                                                   

 78%|███████▊  | 4467/5758 [25:24<06:52,  3.13it/s]
 78%|███████▊  | 4468/5758 [25:24<06:41,  3.21it/s]
                                                   

 78%|███████▊  | 4468/5758 [25:24<06:41,  3.21it/s]
 78%|███████▊  | 4469/5758 [25:24<06:34,  3.26it/s]
                                                   

 78%|███████▊  | 4469/5758 [25:24<06:34,  3.26it/s]
 78%|███████▊  | 4470/5758 [25:25<06:50,  3.14it/s]
                                                   

 78%|███████▊  | 4470/5758 [25:25<06:50,  3.14it/s]
 78%|███████▊  | 4471/5758 [25:25<07:07,  3.01it/s]
                                                   

 78%|███████▊  | 4471/5758 [25:25<07:07,  3.01it/s]
 78%|███████▊  | 4472/5758 [25:26<07:09,  3.00it/s]
                                                   

 78%|███████▊  | 4472/5758 [25:26<07:09,  3.00it/s]
 78%|███████▊  | 4473/5758 [25:26<07:28,  2.86it/s]
                                                   

 78%|███████▊  | 4473/5758 [25:26<07:28,  2.86it/s]
 78%|███████▊  | 4474/5758 [25:26<07:39,  2.79it/s]
                                                   

 78%|███████▊  | 4474/5758 [25:26<07:39,  2.79it/s]
 78%|███████▊  | 4475/5758 [25:27<07:35,  2.82it/s]
                                                   

 78%|███████▊  | 4475/5758 [25:27<07:35,  2.82it/s]
 78%|███████▊  | 4476/5758 [25:27<07:25,  2.88it/s]
                                                   

 78%|███████▊  | 4476/5758 [25:27<07:25,  2.88it/s]
 78%|███████▊  | 4477/5758 [25:27<07:13,  2.96it/s]
                                                   

 78%|███████▊  | 4477/5758 [25:27<07:13,  2.96it/s]
 78%|███████▊  | 4478/5758 [25:28<06:58,  3.06it/s]
                                                   

 78%|███████▊  | 4478/5758 [25:28<06:58,  3.06it/s]
 78%|███████▊  | 4479/5758 [25:28<06:57,  3.06it/s]
                                                   

 78%|███████▊  | 4479/5758 [25:28<06:57,  3.06it/s]
 78%|███████▊  | 4480/5758 [25:28<07:59,  2.67it/s]
                                                   

 78%|███████▊  | 4480/5758 [25:28<07:59,  2.67it/s]
 78%|███████▊  | 4481/5758 [25:29<07:47,  2.73it/s]
                                                   

 78%|███████▊  | 4481/5758 [25:29<07:47,  2.73it/s]
 78%|███████▊  | 4482/5758 [25:29<07:27,  2.85it/s]
                                                   

 78%|███████▊  | 4482/5758 [25:29<07:27,  2.85it/s]
 78%|███████▊  | 4483/5758 [25:29<07:12,  2.94it/s]
                                                   

 78%|███████▊  | 4483/5758 [25:29<07:12,  2.94it/s]
 78%|███████▊  | 4484/5758 [25:30<07:17,  2.91it/s]
                                                   

 78%|███████▊  | 4484/5758 [25:30<07:17,  2.91it/s]
 78%|███████▊  | 4485/5758 [25:30<07:25,  2.86it/s]
                                                   

 78%|███████▊  | 4485/5758 [25:30<07:25,  2.86it/s]
 78%|███████▊  | 4486/5758 [25:30<07:22,  2.87it/s]
                                                   

 78%|███████▊  | 4486/5758 [25:30<07:22,  2.87it/s]
 78%|███████▊  | 4487/5758 [25:31<07:23,  2.87it/s]
                                                   

 78%|███████▊  | 4487/5758 [25:31<07:23,  2.87it/s]
 78%|███████▊  | 4488/5758 [25:31<07:37,  2.78it/s]
                                                   

 78%|███████▊  | 4488/5758 [25:31<07:37,  2.78it/s]
 78%|███████▊  | 4489/5758 [25:32<07:47,  2.71it/s]
                                                   

 78%|███████▊  | 4489/5758 [25:32<07:47,  2.71it/s]
 78%|███████▊  | 4490/5758 [25:32<07:45,  2.73it/s]
                                                   

 78%|███████▊  | 4490/5758 [25:32<07:45,  2.73it/s]
 78%|███████▊  | 4491/5758 [25:32<07:25,  2.84it/s]
                                                   

 78%|███████▊  | 4491/5758 [25:32<07:25,  2.84it/s]
 78%|███████▊  | 4492/5758 [25:33<07:02,  3.00it/s]
                                                   

 78%|███████▊  | 4492/5758 [25:33<07:02,  3.00it/s]
 78%|███████▊  | 4493/5758 [25:33<07:01,  3.00it/s]
                                                   

 78%|███████▊  | 4493/5758 [25:33<07:01,  3.00it/s]
 78%|███████▊  | 4494/5758 [25:33<07:00,  3.00it/s]
                                                   

 78%|███████▊  | 4494/5758 [25:33<07:00,  3.00it/s]
 78%|███████▊  | 4495/5758 [25:34<06:56,  3.03it/s]
                                                   

 78%|███████▊  | 4495/5758 [25:34<06:56,  3.03it/s]
 78%|███████▊  | 4496/5758 [25:34<06:43,  3.12it/s]
                                                   
{'loss': '1.147', 'grad_norm': '12.19', 'learning_rate': '2.289e-06', 'num_tokens': '2.946e+06', 'mean_token_accuracy': '0.6712', 'epoch': '0.7713'}
+{'loss': '1.324', 'grad_norm': '14.62', 'learning_rate': '2.287e-06', 'num_tokens': '2.947e+06', 'mean_token_accuracy': '0.6474', 'epoch': '0.7714'}
+{'loss': '0.9855', 'grad_norm': '8.75', 'learning_rate': '2.286e-06', 'num_tokens': '2.948e+06', 'mean_token_accuracy': '0.7738', 'epoch': '0.7716'}
+{'loss': '0.9511', 'grad_norm': '12.06', 'learning_rate': '2.284e-06', 'num_tokens': '2.949e+06', 'mean_token_accuracy': '0.7001', 'epoch': '0.7718'}
+{'loss': '1.183', 'grad_norm': '11.31', 'learning_rate': '2.282e-06', 'num_tokens': '2.95e+06', 'mean_token_accuracy': '0.6663', 'epoch': '0.772'}
+{'loss': '1.626', 'grad_norm': '18.5', 'learning_rate': '2.28e-06', 'num_tokens': '2.95e+06', 'mean_token_accuracy': '0.6698', 'epoch': '0.7721'}
+{'loss': '1.471', 'grad_norm': '40.75', 'learning_rate': '2.279e-06', 'num_tokens': '2.95e+06', 'mean_token_accuracy': '0.6952', 'epoch': '0.7723'}
+{'loss': '1.173', 'grad_norm': '12.56', 'learning_rate': '2.277e-06', 'num_tokens': '2.951e+06', 'mean_token_accuracy': '0.702', 'epoch': '0.7725'}
+{'loss': '1.466', 'grad_norm': '13.88', 'learning_rate': '2.275e-06', 'num_tokens': '2.951e+06', 'mean_token_accuracy': '0.614', 'epoch': '0.7727'}
+{'loss': '1.264', 'grad_norm': '7.969', 'learning_rate': '2.273e-06', 'num_tokens': '2.953e+06', 'mean_token_accuracy': '0.7034', 'epoch': '0.7728'}
+{'loss': '0.6674', 'grad_norm': '9.688', 'learning_rate': '2.272e-06', 'num_tokens': '2.954e+06', 'mean_token_accuracy': '0.8062', 'epoch': '0.773'}
+{'loss': '0.7628', 'grad_norm': '10.75', 'learning_rate': '2.27e-06', 'num_tokens': '2.955e+06', 'mean_token_accuracy': '0.8024', 'epoch': '0.7732'}
+{'loss': '1.179', 'grad_norm': '12.69', 'learning_rate': '2.268e-06', 'num_tokens': '2.956e+06', 'mean_token_accuracy': '0.6739', 'epoch': '0.7734'}
+{'loss': '1.038', 'grad_norm': '15.94', 'learning_rate': '2.266e-06', 'num_tokens': '2.956e+06', 'mean_token_accuracy': '0.7227', 'epoch': '0.7735'}
+{'loss': '1.233', 'grad_norm': '25.62', 'learning_rate': '2.265e-06', 'num_tokens': '2.956e+06', 'mean_token_accuracy': '0.6917', 'epoch': '0.7737'}
+{'loss': '1.109', 'grad_norm': '50.5', 'learning_rate': '2.263e-06', 'num_tokens': '2.956e+06', 'mean_token_accuracy': '0.8065', 'epoch': '0.7739'}
+{'loss': '0.6638', 'grad_norm': '7.656', 'learning_rate': '2.261e-06', 'num_tokens': '2.958e+06', 'mean_token_accuracy': '0.8073', 'epoch': '0.7741'}
+{'loss': '1.136', 'grad_norm': '17.12', 'learning_rate': '2.259e-06', 'num_tokens': '2.959e+06', 'mean_token_accuracy': '0.6838', 'epoch': '0.7742'}
+{'loss': '1.451', 'grad_norm': '32.5', 'learning_rate': '2.258e-06', 'num_tokens': '2.959e+06', 'mean_token_accuracy': '0.6531', 'epoch': '0.7744'}
+{'loss': '0.9461', 'grad_norm': '20.25', 'learning_rate': '2.256e-06', 'num_tokens': '2.959e+06', 'mean_token_accuracy': '0.7972', 'epoch': '0.7746'}
+{'loss': '1.32', 'grad_norm': '17.25', 'learning_rate': '2.254e-06', 'num_tokens': '2.959e+06', 'mean_token_accuracy': '0.7363', 'epoch': '0.7747'}
+{'loss': '1.156', 'grad_norm': '14.44', 'learning_rate': '2.253e-06', 'num_tokens': '2.96e+06', 'mean_token_accuracy': '0.7659', 'epoch': '0.7749'}
+{'loss': '2.105', 'grad_norm': '11.38', 'learning_rate': '2.251e-06', 'num_tokens': '2.96e+06', 'mean_token_accuracy': '0.5709', 'epoch': '0.7751'}
+{'loss': '2.067', 'grad_norm': '19.38', 'learning_rate': '2.249e-06', 'num_tokens': '2.961e+06', 'mean_token_accuracy': '0.5556', 'epoch': '0.7753'}
+{'loss': '1.493', 'grad_norm': '13.06', 'learning_rate': '2.247e-06', 'num_tokens': '2.962e+06', 'mean_token_accuracy': '0.6824', 'epoch': '0.7754'}
+{'loss': '1.766', 'grad_norm': '11.88', 'learning_rate': '2.246e-06', 'num_tokens': '2.962e+06', 'mean_token_accuracy': '0.6212', 'epoch': '0.7756'}
+{'loss': '1.199', 'grad_norm': '6.812', 'learning_rate': '2.244e-06', 'num_tokens': '2.965e+06', 'mean_token_accuracy': '0.7028', 'epoch': '0.7758'}
+{'loss': '1.703', 'grad_norm': '17.62', 'learning_rate': '2.242e-06', 'num_tokens': '2.965e+06', 'mean_token_accuracy': '0.6552', 'epoch': '0.776'}
+{'loss': '1.04', 'grad_norm': '13.5', 'learning_rate': '2.24e-06', 'num_tokens': '2.966e+06', 'mean_token_accuracy': '0.7156', 'epoch': '0.7761'}
+{'loss': '1.589', 'grad_norm': '15.81', 'learning_rate': '2.239e-06', 'num_tokens': '2.967e+06', 'mean_token_accuracy': '0.6549', 'epoch': '0.7763'}
+{'loss': '1.244', 'grad_norm': '6.875', 'learning_rate': '2.237e-06', 'num_tokens': '2.969e+06', 'mean_token_accuracy': '0.7101', 'epoch': '0.7765'}
+{'loss': '1.445', 'grad_norm': '12.19', 'learning_rate': '2.235e-06', 'num_tokens': '2.97e+06', 'mean_token_accuracy': '0.6211', 'epoch': '0.7767'}
+{'loss': '1.151', 'grad_norm': '13.31', 'learning_rate': '2.233e-06', 'num_tokens': '2.97e+06', 'mean_token_accuracy': '0.6995', 'epoch': '0.7768'}
+{'loss': '1.458', 'grad_norm': '20.5', 'learning_rate': '2.232e-06', 'num_tokens': '2.97e+06', 'mean_token_accuracy': '0.6968', 'epoch': '0.777'}
+{'loss': '1.347', 'grad_norm': '9.688', 'learning_rate': '2.23e-06', 'num_tokens': '2.971e+06', 'mean_token_accuracy': '0.7099', 'epoch': '0.7772'}
+{'loss': '1.443', 'grad_norm': '23', 'learning_rate': '2.228e-06', 'num_tokens': '2.971e+06', 'mean_token_accuracy': '0.6946', 'epoch': '0.7774'}
+{'loss': '0.5054', 'grad_norm': '7.062', 'learning_rate': '2.226e-06', 'num_tokens': '2.973e+06', 'mean_token_accuracy': '0.8403', 'epoch': '0.7775'}
+{'loss': '1.075', 'grad_norm': '13.12', 'learning_rate': '2.225e-06', 'num_tokens': '2.974e+06', 'mean_token_accuracy': '0.6843', 'epoch': '0.7777'}
+{'loss': '1.683', 'grad_norm': '16.5', 'learning_rate': '2.223e-06', 'num_tokens': '2.974e+06', 'mean_token_accuracy': '0.6625', 'epoch': '0.7779'}
+{'loss': '0.5253', 'grad_norm': '8.125', 'learning_rate': '2.221e-06', 'num_tokens': '2.975e+06', 'mean_token_accuracy': '0.8402', 'epoch': '0.778'}
+{'loss': '0.7555', 'grad_norm': '13.62', 'learning_rate': '2.22e-06', 'num_tokens': '2.976e+06', 'mean_token_accuracy': '0.8225', 'epoch': '0.7782'}
+{'loss': '2.093', 'grad_norm': '13.62', 'learning_rate': '2.218e-06', 'num_tokens': '2.976e+06', 'mean_token_accuracy': '0.584', 'epoch': '0.7784'}
+{'loss': '1.169', 'grad_norm': '12.12', 'learning_rate': '2.216e-06', 'num_tokens': '2.977e+06', 'mean_token_accuracy': '0.6703', 'epoch': '0.7786'}
+{'loss': '1.187', 'grad_norm': '6.438', 'learning_rate': '2.214e-06', 'num_tokens': '2.979e+06', 'mean_token_accuracy': '0.7188', 'epoch': '0.7787'}
+{'loss': '1.173', 'grad_norm': '5.25', 'learning_rate': '2.213e-06', 'num_tokens': '2.983e+06', 'mean_token_accuracy': '0.7233', 'epoch': '0.7789'}
+{'loss': '1.091', 'grad_norm': '28.12', 'learning_rate': '2.211e-06', 'num_tokens': '2.983e+06', 'mean_token_accuracy': '0.77', 'epoch': '0.7791'}
+{'loss': '1.257', 'grad_norm': '7.281', 'learning_rate': '2.209e-06', 'num_tokens': '2.985e+06', 'mean_token_accuracy': '0.7153', 'epoch': '0.7793'}
+{'loss': '1.72', 'grad_norm': '17.75', 'learning_rate': '2.207e-06', 'num_tokens': '2.986e+06', 'mean_token_accuracy': '0.6441', 'epoch': '0.7794'}
+{'loss': '1.25', 'grad_norm': '13.75', 'learning_rate': '2.206e-06', 'num_tokens': '2.986e+06', 'mean_token_accuracy': '0.6796', 'epoch': '0.7796'}
+{'loss': '1.058', 'grad_norm': '5.25', 'learning_rate': '2.204e-06', 'num_tokens': '2.99e+06', 'mean_token_accuracy': '0.7531', 'epoch': '0.7798'}
+{'loss': '0.739', 'grad_norm': '12.5', 'learning_rate': '2.202e-06', 'num_tokens': '2.991e+06', 'mean_token_accuracy': '0.8217', 'epoch': '0.78'}
+{'loss': '1.579', 'grad_norm': '46.25', 'learning_rate': '2.2e-06', 'num_tokens': '2.991e+06', 'mean_token_accuracy': '0.7077', 'epoch': '0.7801'}
+{'loss': '0.9601', 'grad_norm': '11.81', 'learning_rate': '2.199e-06', 'num_tokens': '2.992e+06', 'mean_token_accuracy': '0.7243', 'epoch': '0.7803'}
+{'loss': '1.986', 'grad_norm': '22.62', 'learning_rate': '2.197e-06', 'num_tokens': '2.992e+06', 'mean_token_accuracy': '0.5922', 'epoch': '0.7805'}
+{'loss': '1.254', 'grad_norm': '20.12', 'learning_rate': '2.195e-06', 'num_tokens': '2.992e+06', 'mean_token_accuracy': '0.7037', 'epoch': '0.7807'}
+
 78%|███████▊  | 4496/5758 [25:34<06:43,  3.12it/s]
 78%|███████▊  | 4497/5758 [25:34<07:03,  2.98it/s]
                                                   

 78%|███████▊  | 4497/5758 [25:34<07:03,  2.98it/s]
 78%|███████▊  | 4498/5758 [25:35<06:57,  3.02it/s]
                                                   

 78%|███████▊  | 4498/5758 [25:35<06:57,  3.02it/s]
 78%|███████▊  | 4499/5758 [25:35<06:45,  3.11it/s]
                                                   

 78%|███████▊  | 4499/5758 [25:35<06:45,  3.11it/s]
 78%|███████▊  | 4500/5758 [25:35<06:44,  3.11it/s]
                                                   

 78%|███████▊  | 4500/5758 [25:35<06:44,  3.11it/s]{'loss': '1.158', 'grad_norm': '13.69', 'learning_rate': '2.193e-06', 'num_tokens': '2.993e+06', 'mean_token_accuracy': '0.6708', 'epoch': '0.7808'}
+{'loss': '1.197', 'grad_norm': '10.44', 'learning_rate': '2.192e-06', 'num_tokens': '2.994e+06', 'mean_token_accuracy': '0.657', 'epoch': '0.781'}
+{'loss': '1.092', 'grad_norm': '27.25', 'learning_rate': '2.19e-06', 'num_tokens': '2.994e+06', 'mean_token_accuracy': '0.7552', 'epoch': '0.7812'}
+{'loss': '0.7494', 'grad_norm': '28.12', 'learning_rate': '2.188e-06', 'num_tokens': '2.994e+06', 'mean_token_accuracy': '0.8239', 'epoch': '0.7813'}
+{'loss': '0.5959', 'grad_norm': '7.406', 'learning_rate': '2.187e-06', 'num_tokens': '2.996e+06', 'mean_token_accuracy': '0.8201', 'epoch': '0.7815'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.59it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.59it/s]
+
 78%|███████▊  | 4501/5758 [25:37<17:47,  1.18it/s]
                                                   

 78%|███████▊  | 4501/5758 [25:37<17:47,  1.18it/s]
 78%|███████▊  | 4502/5758 [25:38<14:18,  1.46it/s]
                                                   

 78%|███████▊  | 4502/5758 [25:38<14:18,  1.46it/s]
 78%|███████▊  | 4503/5758 [25:38<12:28,  1.68it/s]
                                                   

 78%|███████▊  | 4503/5758 [25:38<12:28,  1.68it/s]
 78%|███████▊  | 4504/5758 [25:38<11:48,  1.77it/s]
                                                   

 78%|███████▊  | 4504/5758 [25:38<11:48,  1.77it/s]
 78%|███████▊  | 4505/5758 [25:39<10:23,  2.01it/s]
                                                   

 78%|███████▊  | 4505/5758 [25:39<10:23,  2.01it/s]
 78%|███████▊  | 4506/5758 [25:39<09:46,  2.14it/s]
                                                   

 78%|███████▊  | 4506/5758 [25:39<09:46,  2.14it/s]
 78%|███████▊  | 4507/5758 [25:40<09:43,  2.14it/s]
                                                   

 78%|███████▊  | 4507/5758 [25:40<09:43,  2.14it/s]
 78%|███████▊  | 4508/5758 [25:40<09:22,  2.22it/s]
                                                   

 78%|███████▊  | 4508/5758 [25:40<09:22,  2.22it/s]
 78%|███████▊  | 4509/5758 [25:40<09:30,  2.19it/s]
                                                   

 78%|███████▊  | 4509/5758 [25:40<09:30,  2.19it/s]
 78%|███████▊  | 4510/5758 [25:41<09:42,  2.14it/s]
                                                   

 78%|███████▊  | 4510/5758 [25:41<09:42,  2.14it/s]
 78%|███████▊  | 4511/5758 [25:41<09:50,  2.11it/s]
                                                   

 78%|███████▊  | 4511/5758 [25:41<09:50,  2.11it/s]
 78%|███████▊  | 4512/5758 [25:42<09:10,  2.26it/s]
                                                   

 78%|███████▊  | 4512/5758 [25:42<09:10,  2.26it/s]
 78%|███████▊  | 4513/5758 [25:42<09:00,  2.30it/s]
                                                   

 78%|███████▊  | 4513/5758 [25:42<09:00,  2.30it/s]
 78%|███████▊  | 4514/5758 [25:43<08:38,  2.40it/s]
                                                   

 78%|███████▊  | 4514/5758 [25:43<08:38,  2.40it/s]
 78%|███████▊  | 4515/5758 [25:43<08:55,  2.32it/s]
                                                   

 78%|███████▊  | 4515/5758 [25:43<08:55,  2.32it/s]
 78%|███████▊  | 4516/5758 [25:44<09:05,  2.28it/s]
                                                   

 78%|███████▊  | 4516/5758 [25:44<09:05,  2.28it/s]
 78%|███████▊  | 4517/5758 [25:44<08:48,  2.35it/s]
                                                   

 78%|███████▊  | 4517/5758 [25:44<08:48,  2.35it/s]
 78%|███████▊  | 4518/5758 [25:44<08:16,  2.50it/s]
                                                   

 78%|███████▊  | 4518/5758 [25:44<08:16,  2.50it/s]
 78%|███████▊  | 4519/5758 [25:45<07:55,  2.61it/s]
                                                   

 78%|███████▊  | 4519/5758 [25:45<07:55,  2.61it/s]
 78%|███████▊  | 4520/5758 [25:45<07:45,  2.66it/s]
                                                   

 78%|███████▊  | 4520/5758 [25:45<07:45,  2.66it/s]
 79%|███████▊  | 4521/5758 [25:45<07:39,  2.69it/s]
                                                   

 79%|███████▊  | 4521/5758 [25:45<07:39,  2.69it/s]
 79%|███████▊  | 4522/5758 [25:46<07:40,  2.68it/s]
                                                   

 79%|███████▊  | 4522/5758 [25:46<07:40,  2.68it/s]
 79%|███████▊  | 4523/5758 [25:46<07:24,  2.78it/s]
                                                   

 79%|███████▊  | 4523/5758 [25:46<07:24,  2.78it/s]
 79%|███████▊  | 4524/5758 [25:46<07:05,  2.90it/s]
                                                   

 79%|███████▊  | 4524/5758 [25:46<07:05,  2.90it/s]
 79%|███████▊  | 4525/5758 [25:47<06:51,  2.99it/s]
                                                   

 79%|███████▊  | 4525/5758 [25:47<06:51,  2.99it/s]
 79%|███████▊  | 4526/5758 [25:47<06:51,  2.99it/s]
                                                   

 79%|███████▊  | 4526/5758 [25:47<06:51,  2.99it/s]
 79%|███████▊  | 4527/5758 [25:47<06:40,  3.07it/s]
                                                   

 79%|███████▊  | 4527/5758 [25:47<06:40,  3.07it/s]
 79%|███████▊  | 4528/5758 [25:48<06:43,  3.05it/s]
                                                   

 79%|███████▊  | 4528/5758 [25:48<06:43,  3.05it/s]
 79%|███████▊  | 4529/5758 [25:48<06:49,  3.00it/s]
                                                   

 79%|███████▊  | 4529/5758 [25:48<06:49,  3.00it/s]
 79%|███████▊  | 4530/5758 [25:48<07:24,  2.77it/s]
                                                   

 79%|███████▊  | 4530/5758 [25:48<07:24,  2.77it/s]
 79%|███████▊  | 4531/5758 [25:49<07:04,  2.89it/s]
                                                   

 79%|███████▊  | 4531/5758 [25:49<07:04,  2.89it/s]
 79%|███████▊  | 4532/5758 [25:49<06:45,  3.02it/s]
                                                   

 79%|███████▊  | 4532/5758 [25:49<06:45,  3.02it/s]
 79%|███████▊  | 4533/5758 [25:49<06:53,  2.96it/s]
                                                   

 79%|███████▊  | 4533/5758 [25:49<06:53,  2.96it/s]
 79%|███████▊  | 4534/5758 [25:50<06:58,  2.93it/s]
                                                   

 79%|███████▊  | 4534/5758 [25:50<06:58,  2.93it/s]
 79%|███████▉  | 4535/5758 [25:50<06:41,  3.05it/s]
                                                   

 79%|███████▉  | 4535/5758 [25:50<06:41,  3.05it/s]
 79%|███████▉  | 4536/5758 [25:50<06:28,  3.15it/s]
                                                   

 79%|███████▉  | 4536/5758 [25:50<06:28,  3.15it/s]
 79%|███████▉  | 4537/5758 [25:51<06:25,  3.17it/s]
                                                   

 79%|███████▉  | 4537/5758 [25:51<06:25,  3.17it/s]
 79%|███████▉  | 4538/5758 [25:51<06:17,  3.24it/s]
                                                   

 79%|███████▉  | 4538/5758 [25:51<06:17,  3.24it/s]
 79%|███████▉  | 4539/5758 [25:51<06:18,  3.22it/s]
                                                   

 79%|███████▉  | 4539/5758 [25:51<06:18,  3.22it/s]
 79%|███████▉  | 4540/5758 [25:52<06:17,  3.23it/s]
                                                   

 79%|███████▉  | 4540/5758 [25:52<06:17,  3.23it/s]
 79%|███████▉  | 4541/5758 [25:52<06:45,  3.00it/s]
                                                   

 79%|███████▉  | 4541/5758 [25:52<06:45,  3.00it/s]
 79%|███████▉  | 4542/5758 [25:52<06:37,  3.06it/s]
                                                   

 79%|███████▉  | 4542/5758 [25:52<06:37,  3.06it/s]
 79%|███████▉  | 4543/5758 [25:53<06:39,  3.04it/s]
                                                   

 79%|███████▉  | 4543/5758 [25:53<06:39,  3.04it/s]
 79%|███████▉  | 4544/5758 [25:53<06:31,  3.10it/s]
                                                   

 79%|███████▉  | 4544/5758 [25:53<06:31,  3.10it/s]
 79%|███████▉  | 4545/5758 [25:53<06:09,  3.29it/s]
                                                   

 79%|███████▉  | 4545/5758 [25:53<06:09,  3.29it/s]
 79%|███████▉  | 4546/5758 [25:53<06:07,  3.30it/s]
                                                   

 79%|███████▉  | 4546/5758 [25:53<06:07,  3.30it/s]
 79%|███████▉  | 4547/5758 [25:54<06:31,  3.09it/s]
                                                   

 79%|███████▉  | 4547/5758 [25:54<06:31,  3.09it/s]
 79%|███████▉  | 4548/5758 [25:54<06:48,  2.96it/s]
                                                   

 79%|███████▉  | 4548/5758 [25:54<06:48,  2.96it/s]
 79%|███████▉  | 4549/5758 [25:55<07:05,  2.84it/s]
                                                   

 79%|███████▉  | 4549/5758 [25:55<07:05,  2.84it/s]
 79%|███████▉  | 4550/5758 [25:55<06:50,  2.94it/s]
                                                   

 79%|███████▉  | 4550/5758 [25:55<06:50,  2.94it/s]
 79%|███████▉  | 4551/5758 [25:55<06:37,  3.03it/s]
                                                   

 79%|███████▉  | 4551/5758 [25:55<06:37,  3.03it/s]
 79%|███████▉  | 4552/5758 [25:55<06:33,  3.07it/s]
                                                   

 79%|███████▉  | 4552/5758 [25:56<06:33,  3.07it/s]
 79%|███████▉  | 4553/5758 [25:56<06:41,  3.00it/s]
                                                   

 79%|███████▉  | 4553/5758 [25:56<06:41,  3.00it/s]
 79%|███████▉  | 4554/5758 [25:56<06:29,  3.09it/s]
                                                   

 79%|███████▉  | 4554/5758 [25:56<06:29,  3.09it/s]
 79%|███████▉  | 4555/5758 [25:57<06:51,  2.92it/s]
                                                   

 79%|███████▉  | 4555/5758 [25:57<06:51,  2.92it/s]
 79%|███████▉  | 4556/5758 [25:57<06:37,  3.03it/s]
                                                   
{'loss': '1.83', 'grad_norm': '19', 'learning_rate': '2.185e-06', 'num_tokens': '2.996e+06', 'mean_token_accuracy': '0.6246', 'epoch': '0.7817'}
+{'loss': '1.247', 'grad_norm': '12.88', 'learning_rate': '2.183e-06', 'num_tokens': '2.997e+06', 'mean_token_accuracy': '0.682', 'epoch': '0.7819'}
+{'loss': '1.748', 'grad_norm': '18', 'learning_rate': '2.181e-06', 'num_tokens': '2.997e+06', 'mean_token_accuracy': '0.6232', 'epoch': '0.782'}
+{'loss': '1.12', 'grad_norm': '11.38', 'learning_rate': '2.18e-06', 'num_tokens': '2.998e+06', 'mean_token_accuracy': '0.6775', 'epoch': '0.7822'}
+{'loss': '1.404', 'grad_norm': '14.06', 'learning_rate': '2.178e-06', 'num_tokens': '2.999e+06', 'mean_token_accuracy': '0.6159', 'epoch': '0.7824'}
+{'loss': '1.791', 'grad_norm': '17.5', 'learning_rate': '2.176e-06', 'num_tokens': '2.999e+06', 'mean_token_accuracy': '0.6423', 'epoch': '0.7826'}
+{'loss': '0.5855', 'grad_norm': '30.88', 'learning_rate': '2.174e-06', 'num_tokens': '2.999e+06', 'mean_token_accuracy': '0.8607', 'epoch': '0.7827'}
+{'loss': '1.375', 'grad_norm': '12.94', 'learning_rate': '2.173e-06', 'num_tokens': '3e+06', 'mean_token_accuracy': '0.6364', 'epoch': '0.7829'}
+{'loss': '1.102', 'grad_norm': '12.81', 'learning_rate': '2.171e-06', 'num_tokens': '3.001e+06', 'mean_token_accuracy': '0.7455', 'epoch': '0.7831'}
+{'loss': '0.8532', 'grad_norm': '18.75', 'learning_rate': '2.169e-06', 'num_tokens': '3.001e+06', 'mean_token_accuracy': '0.8288', 'epoch': '0.7833'}
+{'loss': '0.7481', 'grad_norm': '26.5', 'learning_rate': '2.167e-06', 'num_tokens': '3.001e+06', 'mean_token_accuracy': '0.8148', 'epoch': '0.7834'}
+{'loss': '1.079', 'grad_norm': '12.81', 'learning_rate': '2.166e-06', 'num_tokens': '3.002e+06', 'mean_token_accuracy': '0.7035', 'epoch': '0.7836'}
+{'loss': '1.916', 'grad_norm': '10.94', 'learning_rate': '2.164e-06', 'num_tokens': '3.003e+06', 'mean_token_accuracy': '0.6247', 'epoch': '0.7838'}
+{'loss': '1.005', 'grad_norm': '7.188', 'learning_rate': '2.162e-06', 'num_tokens': '3.004e+06', 'mean_token_accuracy': '0.7511', 'epoch': '0.784'}
+{'loss': '1.205', 'grad_norm': '5.281', 'learning_rate': '2.16e-06', 'num_tokens': '3.008e+06', 'mean_token_accuracy': '0.7199', 'epoch': '0.7841'}
+{'loss': '0.837', 'grad_norm': '21.62', 'learning_rate': '2.159e-06', 'num_tokens': '3.008e+06', 'mean_token_accuracy': '0.8066', 'epoch': '0.7843'}
+{'loss': '1.102', 'grad_norm': '12.19', 'learning_rate': '2.157e-06', 'num_tokens': '3.009e+06', 'mean_token_accuracy': '0.6917', 'epoch': '0.7845'}
+{'loss': '0.59', 'grad_norm': '7.219', 'learning_rate': '2.155e-06', 'num_tokens': '3.01e+06', 'mean_token_accuracy': '0.8291', 'epoch': '0.7846'}
+{'loss': '1.328', 'grad_norm': '11.81', 'learning_rate': '2.154e-06', 'num_tokens': '3.011e+06', 'mean_token_accuracy': '0.6373', 'epoch': '0.7848'}
+{'loss': '1.259', 'grad_norm': '10.94', 'learning_rate': '2.152e-06', 'num_tokens': '3.012e+06', 'mean_token_accuracy': '0.6499', 'epoch': '0.785'}
+{'loss': '2.045', 'grad_norm': '39.25', 'learning_rate': '2.15e-06', 'num_tokens': '3.012e+06', 'mean_token_accuracy': '0.5431', 'epoch': '0.7852'}
+{'loss': '1.504', 'grad_norm': '19.12', 'learning_rate': '2.148e-06', 'num_tokens': '3.013e+06', 'mean_token_accuracy': '0.6828', 'epoch': '0.7853'}
+{'loss': '1.969', 'grad_norm': '11.25', 'learning_rate': '2.147e-06', 'num_tokens': '3.014e+06', 'mean_token_accuracy': '0.5934', 'epoch': '0.7855'}
+{'loss': '1.18', 'grad_norm': '14.19', 'learning_rate': '2.145e-06', 'num_tokens': '3.014e+06', 'mean_token_accuracy': '0.7147', 'epoch': '0.7857'}
+{'loss': '0.6837', 'grad_norm': '20.75', 'learning_rate': '2.143e-06', 'num_tokens': '3.014e+06', 'mean_token_accuracy': '0.8187', 'epoch': '0.7859'}
+{'loss': '0.7986', 'grad_norm': '12.69', 'learning_rate': '2.141e-06', 'num_tokens': '3.015e+06', 'mean_token_accuracy': '0.8109', 'epoch': '0.786'}
+{'loss': '1.111', 'grad_norm': '12.19', 'learning_rate': '2.14e-06', 'num_tokens': '3.016e+06', 'mean_token_accuracy': '0.6758', 'epoch': '0.7862'}
+{'loss': '0.6742', 'grad_norm': '6.969', 'learning_rate': '2.138e-06', 'num_tokens': '3.017e+06', 'mean_token_accuracy': '0.8016', 'epoch': '0.7864'}
+{'loss': '0.6728', 'grad_norm': '25', 'learning_rate': '2.136e-06', 'num_tokens': '3.017e+06', 'mean_token_accuracy': '0.8738', 'epoch': '0.7866'}
+{'loss': '1.977', 'grad_norm': '27', 'learning_rate': '2.134e-06', 'num_tokens': '3.018e+06', 'mean_token_accuracy': '0.6286', 'epoch': '0.7867'}
+{'loss': '1.658', 'grad_norm': '26.5', 'learning_rate': '2.133e-06', 'num_tokens': '3.018e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.7869'}
+{'loss': '1.24', 'grad_norm': '9.5', 'learning_rate': '2.131e-06', 'num_tokens': '3.019e+06', 'mean_token_accuracy': '0.6972', 'epoch': '0.7871'}
+{'loss': '1.85', 'grad_norm': '19.25', 'learning_rate': '2.129e-06', 'num_tokens': '3.02e+06', 'mean_token_accuracy': '0.6034', 'epoch': '0.7873'}
+{'loss': '1.946', 'grad_norm': '10.94', 'learning_rate': '2.127e-06', 'num_tokens': '3.02e+06', 'mean_token_accuracy': '0.5994', 'epoch': '0.7874'}
+{'loss': '0.8708', 'grad_norm': '11.94', 'learning_rate': '2.126e-06', 'num_tokens': '3.021e+06', 'mean_token_accuracy': '0.7539', 'epoch': '0.7876'}
+{'loss': '0.6832', 'grad_norm': '16.75', 'learning_rate': '2.124e-06', 'num_tokens': '3.021e+06', 'mean_token_accuracy': '0.8388', 'epoch': '0.7878'}
+{'loss': '1.172', 'grad_norm': '13.06', 'learning_rate': '2.122e-06', 'num_tokens': '3.022e+06', 'mean_token_accuracy': '0.6865', 'epoch': '0.7879'}
+{'loss': '1.136', 'grad_norm': '29.5', 'learning_rate': '2.121e-06', 'num_tokens': '3.022e+06', 'mean_token_accuracy': '0.7362', 'epoch': '0.7881'}
+{'loss': '0.5615', 'grad_norm': '7.531', 'learning_rate': '2.119e-06', 'num_tokens': '3.024e+06', 'mean_token_accuracy': '0.8254', 'epoch': '0.7883'}
+{'loss': '1.417', 'grad_norm': '16.38', 'learning_rate': '2.117e-06', 'num_tokens': '3.024e+06', 'mean_token_accuracy': '0.7083', 'epoch': '0.7885'}
+{'loss': '1.362', 'grad_norm': '17.5', 'learning_rate': '2.115e-06', 'num_tokens': '3.024e+06', 'mean_token_accuracy': '0.6621', 'epoch': '0.7886'}
+{'loss': '2.263', 'grad_norm': '9.25', 'learning_rate': '2.114e-06', 'num_tokens': '3.026e+06', 'mean_token_accuracy': '0.5272', 'epoch': '0.7888'}
+{'loss': '1.767', 'grad_norm': '17.75', 'learning_rate': '2.112e-06', 'num_tokens': '3.026e+06', 'mean_token_accuracy': '0.6199', 'epoch': '0.789'}
+{'loss': '1.043', 'grad_norm': '11.5', 'learning_rate': '2.11e-06', 'num_tokens': '3.027e+06', 'mean_token_accuracy': '0.7099', 'epoch': '0.7892'}
+{'loss': '0.9874', 'grad_norm': '60.75', 'learning_rate': '2.108e-06', 'num_tokens': '3.027e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.7893'}
+{'loss': '0.8605', 'grad_norm': '19.38', 'learning_rate': '2.107e-06', 'num_tokens': '3.027e+06', 'mean_token_accuracy': '0.8277', 'epoch': '0.7895'}
+{'loss': '0.8409', 'grad_norm': '13.56', 'learning_rate': '2.105e-06', 'num_tokens': '3.028e+06', 'mean_token_accuracy': '0.7696', 'epoch': '0.7897'}
+{'loss': '1.843', 'grad_norm': '9.625', 'learning_rate': '2.103e-06', 'num_tokens': '3.029e+06', 'mean_token_accuracy': '0.5978', 'epoch': '0.7899'}
+{'loss': '1.39', 'grad_norm': '25.75', 'learning_rate': '2.101e-06', 'num_tokens': '3.029e+06', 'mean_token_accuracy': '0.7231', 'epoch': '0.79'}
+{'loss': '1.333', 'grad_norm': '16.62', 'learning_rate': '2.1e-06', 'num_tokens': '3.029e+06', 'mean_token_accuracy': '0.723', 'epoch': '0.7902'}
+{'loss': '1.372', 'grad_norm': '19.75', 'learning_rate': '2.098e-06', 'num_tokens': '3.03e+06', 'mean_token_accuracy': '0.6921', 'epoch': '0.7904'}
+{'loss': '1.728', 'grad_norm': '16.12', 'learning_rate': '2.096e-06', 'num_tokens': '3.03e+06', 'mean_token_accuracy': '0.6322', 'epoch': '0.7906'}
+{'loss': '0.931', 'grad_norm': '19.88', 'learning_rate': '2.094e-06', 'num_tokens': '3.03e+06', 'mean_token_accuracy': '0.7787', 'epoch': '0.7907'}
+{'loss': '1.401', 'grad_norm': '14.31', 'learning_rate': '2.093e-06', 'num_tokens': '3.031e+06', 'mean_token_accuracy': '0.6365', 'epoch': '0.7909'}
+{'loss': '1.705', 'grad_norm': '15.12', 'learning_rate': '2.091e-06', 'num_tokens': '3.032e+06', 'mean_token_accuracy': '0.6315', 'epoch': '0.7911'}
+
 79%|███████▉  | 4556/5758 [25:57<06:37,  3.03it/s]
 79%|███████▉  | 4557/5758 [25:57<07:13,  2.77it/s]
                                                   

 79%|███████▉  | 4557/5758 [25:57<07:13,  2.77it/s]
 79%|███████▉  | 4558/5758 [25:58<06:57,  2.87it/s]
                                                   

 79%|███████▉  | 4558/5758 [25:58<06:57,  2.87it/s]
 79%|███████▉  | 4559/5758 [25:58<06:37,  3.02it/s]
                                                   

 79%|███████▉  | 4559/5758 [25:58<06:37,  3.02it/s]
 79%|███████▉  | 4560/5758 [25:58<06:23,  3.13it/s]
                                                   

 79%|███████▉  | 4560/5758 [25:58<06:23,  3.13it/s]
 79%|███████▉  | 4561/5758 [25:58<06:24,  3.12it/s]
                                                   

 79%|███████▉  | 4561/5758 [25:59<06:24,  3.12it/s]
 79%|███████▉  | 4562/5758 [25:59<06:46,  2.94it/s]
                                                   

 79%|███████▉  | 4562/5758 [25:59<06:46,  2.94it/s]
 79%|███████▉  | 4563/5758 [25:59<06:51,  2.90it/s]
                                                   

 79%|███████▉  | 4563/5758 [25:59<06:51,  2.90it/s]
 79%|███████▉  | 4564/5758 [26:00<06:37,  3.01it/s]
                                                   

 79%|███████▉  | 4564/5758 [26:00<06:37,  3.01it/s]
 79%|███████▉  | 4565/5758 [26:00<06:39,  2.99it/s]
                                                   

 79%|███████▉  | 4565/5758 [26:00<06:39,  2.99it/s]
 79%|███████▉  | 4566/5758 [26:00<06:52,  2.89it/s]
                                                   

 79%|███████▉  | 4566/5758 [26:00<06:52,  2.89it/s]
 79%|███████▉  | 4567/5758 [26:01<06:45,  2.94it/s]
                                                   

 79%|███████▉  | 4567/5758 [26:01<06:45,  2.94it/s]
 79%|███████▉  | 4568/5758 [26:01<07:01,  2.83it/s]
                                                   

 79%|███████▉  | 4568/5758 [26:01<07:01,  2.83it/s]
 79%|███████▉  | 4569/5758 [26:01<06:50,  2.90it/s]
                                                   

 79%|███████▉  | 4569/5758 [26:01<06:50,  2.90it/s]
 79%|███████▉  | 4570/5758 [26:02<06:59,  2.83it/s]
                                                   

 79%|███████▉  | 4570/5758 [26:02<06:59,  2.83it/s]
 79%|███████▉  | 4571/5758 [26:02<06:45,  2.92it/s]
                                                   

 79%|███████▉  | 4571/5758 [26:02<06:45,  2.92it/s]
 79%|███████▉  | 4572/5758 [26:02<07:00,  2.82it/s]
                                                   

 79%|███████▉  | 4572/5758 [26:02<07:00,  2.82it/s]
 79%|███████▉  | 4573/5758 [26:03<06:51,  2.88it/s]
                                                   

 79%|███████▉  | 4573/5758 [26:03<06:51,  2.88it/s]
 79%|███████▉  | 4574/5758 [26:03<07:08,  2.76it/s]
                                                   

 79%|███████▉  | 4574/5758 [26:03<07:08,  2.76it/s]
 79%|███████▉  | 4575/5758 [26:03<07:18,  2.70it/s]
                                                   

 79%|███████▉  | 4575/5758 [26:03<07:18,  2.70it/s]
 79%|███████▉  | 4576/5758 [26:04<07:22,  2.67it/s]
                                                   

 79%|███████▉  | 4576/5758 [26:04<07:22,  2.67it/s]
 79%|███████▉  | 4577/5758 [26:04<07:06,  2.77it/s]
                                                   

 79%|███████▉  | 4577/5758 [26:04<07:06,  2.77it/s]
 80%|███████▉  | 4578/5758 [26:05<07:02,  2.79it/s]
                                                   

 80%|███████▉  | 4578/5758 [26:05<07:02,  2.79it/s]
 80%|███████▉  | 4579/5758 [26:05<06:56,  2.83it/s]
                                                   

 80%|███████▉  | 4579/5758 [26:05<06:56,  2.83it/s]
 80%|███████▉  | 4580/5758 [26:05<06:33,  2.99it/s]
                                                   

 80%|███████▉  | 4580/5758 [26:05<06:33,  2.99it/s]
 80%|███████▉  | 4581/5758 [26:05<06:20,  3.09it/s]
                                                   

 80%|███████▉  | 4581/5758 [26:05<06:20,  3.09it/s]
 80%|███████▉  | 4582/5758 [26:06<06:11,  3.17it/s]
                                                   

 80%|███████▉  | 4582/5758 [26:06<06:11,  3.17it/s]
 80%|███████▉  | 4583/5758 [26:06<06:07,  3.20it/s]
                                                   

 80%|███████▉  | 4583/5758 [26:06<06:07,  3.20it/s]
 80%|███████▉  | 4584/5758 [26:06<06:00,  3.26it/s]
                                                   

 80%|███████▉  | 4584/5758 [26:06<06:00,  3.26it/s]
 80%|███████▉  | 4585/5758 [26:07<06:48,  2.87it/s]
                                                   

 80%|███████▉  | 4585/5758 [26:07<06:48,  2.87it/s]
 80%|███████▉  | 4586/5758 [26:07<06:41,  2.92it/s]
                                                   

 80%|███████▉  | 4586/5758 [26:07<06:41,  2.92it/s]
 80%|███████▉  | 4587/5758 [26:07<06:29,  3.01it/s]
                                                   

 80%|███████▉  | 4587/5758 [26:07<06:29,  3.01it/s]
 80%|███████▉  | 4588/5758 [26:08<06:25,  3.04it/s]
                                                   

 80%|███████▉  | 4588/5758 [26:08<06:25,  3.04it/s]
 80%|███████▉  | 4589/5758 [26:08<06:20,  3.07it/s]
                                                   

 80%|███████▉  | 4589/5758 [26:08<06:20,  3.07it/s]
 80%|███████▉  | 4590/5758 [26:08<06:06,  3.18it/s]
                                                   

 80%|███████▉  | 4590/5758 [26:08<06:06,  3.18it/s]
 80%|███████▉  | 4591/5758 [26:09<06:27,  3.01it/s]
                                                   

 80%|███████▉  | 4591/5758 [26:09<06:27,  3.01it/s]
 80%|███████▉  | 4592/5758 [26:09<06:21,  3.05it/s]
                                                   

 80%|███████▉  | 4592/5758 [26:09<06:21,  3.05it/s]
 80%|███████▉  | 4593/5758 [26:09<06:14,  3.11it/s]
                                                   

 80%|███████▉  | 4593/5758 [26:09<06:14,  3.11it/s]
 80%|███████▉  | 4594/5758 [26:10<06:01,  3.22it/s]
                                                   

 80%|███████▉  | 4594/5758 [26:10<06:01,  3.22it/s]
 80%|███████▉  | 4595/5758 [26:10<05:57,  3.26it/s]
                                                   

 80%|███████▉  | 4595/5758 [26:10<05:57,  3.26it/s]
 80%|███████▉  | 4596/5758 [26:10<05:53,  3.29it/s]
                                                   

 80%|███████▉  | 4596/5758 [26:10<05:53,  3.29it/s]
 80%|███████▉  | 4597/5758 [26:11<06:20,  3.05it/s]
                                                   

 80%|███████▉  | 4597/5758 [26:11<06:20,  3.05it/s]
 80%|███████▉  | 4598/5758 [26:11<06:23,  3.02it/s]
                                                   

 80%|███████▉  | 4598/5758 [26:11<06:23,  3.02it/s]
 80%|███████▉  | 4599/5758 [26:11<06:11,  3.12it/s]
                                                   

 80%|███████▉  | 4599/5758 [26:11<06:11,  3.12it/s]
 80%|███████▉  | 4600/5758 [26:12<06:29,  2.97it/s]
                                                   

 80%|███████▉  | 4600/5758 [26:12<06:29,  2.97it/s]
 80%|███████▉  | 4601/5758 [26:12<06:30,  2.96it/s]
                                                   

 80%|███████▉  | 4601/5758 [26:12<06:30,  2.96it/s]
 80%|███████▉  | 4602/5758 [26:12<06:31,  2.95it/s]
                                                   

 80%|███████▉  | 4602/5758 [26:12<06:31,  2.95it/s]
 80%|███████▉  | 4603/5758 [26:13<06:32,  2.95it/s]
                                                   

 80%|███████▉  | 4603/5758 [26:13<06:32,  2.95it/s]
 80%|███████▉  | 4604/5758 [26:13<06:12,  3.10it/s]
                                                   

 80%|███████▉  | 4604/5758 [26:13<06:12,  3.10it/s]
 80%|███████▉  | 4605/5758 [26:13<06:36,  2.91it/s]
                                                   

 80%|███████▉  | 4605/5758 [26:13<06:36,  2.91it/s]
 80%|███████▉  | 4606/5758 [26:14<06:42,  2.87it/s]
                                                   

 80%|███████▉  | 4606/5758 [26:14<06:42,  2.87it/s]
 80%|████████  | 4607/5758 [26:14<06:44,  2.84it/s]
                                                   

 80%|████████  | 4607/5758 [26:14<06:44,  2.84it/s]
 80%|████████  | 4608/5758 [26:14<06:28,  2.96it/s]
                                                   

 80%|████████  | 4608/5758 [26:14<06:28,  2.96it/s]
 80%|████████  | 4609/5758 [26:15<06:14,  3.06it/s]
                                                   

 80%|████████  | 4609/5758 [26:15<06:14,  3.06it/s]
 80%|████████  | 4610/5758 [26:15<06:08,  3.12it/s]
                                                   

 80%|████████  | 4610/5758 [26:15<06:08,  3.12it/s]
 80%|████████  | 4611/5758 [26:15<06:02,  3.17it/s]
                                                   
{'loss': '0.5568', 'grad_norm': '7.438', 'learning_rate': '2.089e-06', 'num_tokens': '3.033e+06', 'mean_token_accuracy': '0.8287', 'epoch': '0.7912'}
+{'loss': '0.5765', 'grad_norm': '6.969', 'learning_rate': '2.088e-06', 'num_tokens': '3.034e+06', 'mean_token_accuracy': '0.8185', 'epoch': '0.7914'}
+{'loss': '1.321', 'grad_norm': '11.88', 'learning_rate': '2.086e-06', 'num_tokens': '3.035e+06', 'mean_token_accuracy': '0.6662', 'epoch': '0.7916'}
+{'loss': '1.39', 'grad_norm': '13.88', 'learning_rate': '2.084e-06', 'num_tokens': '3.036e+06', 'mean_token_accuracy': '0.6388', 'epoch': '0.7918'}
+{'loss': '1.583', 'grad_norm': '13.56', 'learning_rate': '2.082e-06', 'num_tokens': '3.036e+06', 'mean_token_accuracy': '0.6618', 'epoch': '0.7919'}
+{'loss': '1.461', 'grad_norm': '14.25', 'learning_rate': '2.081e-06', 'num_tokens': '3.037e+06', 'mean_token_accuracy': '0.6581', 'epoch': '0.7921'}
+{'loss': '1.25', 'grad_norm': '17.25', 'learning_rate': '2.079e-06', 'num_tokens': '3.037e+06', 'mean_token_accuracy': '0.737', 'epoch': '0.7923'}
+{'loss': '1.067', 'grad_norm': '23', 'learning_rate': '2.077e-06', 'num_tokens': '3.038e+06', 'mean_token_accuracy': '0.7541', 'epoch': '0.7925'}
+{'loss': '1.321', 'grad_norm': '13', 'learning_rate': '2.075e-06', 'num_tokens': '3.038e+06', 'mean_token_accuracy': '0.6159', 'epoch': '0.7926'}
+{'loss': '1.141', 'grad_norm': '44.75', 'learning_rate': '2.074e-06', 'num_tokens': '3.038e+06', 'mean_token_accuracy': '0.732', 'epoch': '0.7928'}
+{'loss': '1.042', 'grad_norm': '14.31', 'learning_rate': '2.072e-06', 'num_tokens': '3.039e+06', 'mean_token_accuracy': '0.7007', 'epoch': '0.793'}
+{'loss': '0.742', 'grad_norm': '14.5', 'learning_rate': '2.07e-06', 'num_tokens': '3.039e+06', 'mean_token_accuracy': '0.8207', 'epoch': '0.7932'}
+{'loss': '1.153', 'grad_norm': '14.38', 'learning_rate': '2.068e-06', 'num_tokens': '3.04e+06', 'mean_token_accuracy': '0.6876', 'epoch': '0.7933'}
+{'loss': '1.185', 'grad_norm': '31.62', 'learning_rate': '2.067e-06', 'num_tokens': '3.04e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.7935'}
+{'loss': '0.8524', 'grad_norm': '13.5', 'learning_rate': '2.065e-06', 'num_tokens': '3.04e+06', 'mean_token_accuracy': '0.7753', 'epoch': '0.7937'}
+{'loss': '1.472', 'grad_norm': '18.75', 'learning_rate': '2.063e-06', 'num_tokens': '3.041e+06', 'mean_token_accuracy': '0.6942', 'epoch': '0.7939'}
+{'loss': '2.192', 'grad_norm': '19', 'learning_rate': '2.061e-06', 'num_tokens': '3.041e+06', 'mean_token_accuracy': '0.6093', 'epoch': '0.794'}
+{'loss': '2.286', 'grad_norm': '21.38', 'learning_rate': '2.06e-06', 'num_tokens': '3.042e+06', 'mean_token_accuracy': '0.5442', 'epoch': '0.7942'}
+{'loss': '0.7928', 'grad_norm': '24.62', 'learning_rate': '2.058e-06', 'num_tokens': '3.042e+06', 'mean_token_accuracy': '0.8456', 'epoch': '0.7944'}
+{'loss': '2.213', 'grad_norm': '13.75', 'learning_rate': '2.056e-06', 'num_tokens': '3.042e+06', 'mean_token_accuracy': '0.5697', 'epoch': '0.7945'}
+{'loss': '1.325', 'grad_norm': '15.31', 'learning_rate': '2.055e-06', 'num_tokens': '3.043e+06', 'mean_token_accuracy': '0.6691', 'epoch': '0.7947'}
+{'loss': '1.13', 'grad_norm': '25.88', 'learning_rate': '2.053e-06', 'num_tokens': '3.043e+06', 'mean_token_accuracy': '0.7379', 'epoch': '0.7949'}
+{'loss': '1.621', 'grad_norm': '18.38', 'learning_rate': '2.051e-06', 'num_tokens': '3.044e+06', 'mean_token_accuracy': '0.6626', 'epoch': '0.7951'}
+{'loss': '1.194', 'grad_norm': '12.19', 'learning_rate': '2.049e-06', 'num_tokens': '3.044e+06', 'mean_token_accuracy': '0.6545', 'epoch': '0.7952'}
+{'loss': '0.968', 'grad_norm': '13.56', 'learning_rate': '2.048e-06', 'num_tokens': '3.045e+06', 'mean_token_accuracy': '0.7225', 'epoch': '0.7954'}
+{'loss': '1.147', 'grad_norm': '12.12', 'learning_rate': '2.046e-06', 'num_tokens': '3.046e+06', 'mean_token_accuracy': '0.7372', 'epoch': '0.7956'}
+{'loss': '1.661', 'grad_norm': '16.62', 'learning_rate': '2.044e-06', 'num_tokens': '3.046e+06', 'mean_token_accuracy': '0.6386', 'epoch': '0.7958'}
+{'loss': '0.5578', 'grad_norm': '7', 'learning_rate': '2.042e-06', 'num_tokens': '3.048e+06', 'mean_token_accuracy': '0.8368', 'epoch': '0.7959'}
+{'loss': '0.5688', 'grad_norm': '7.781', 'learning_rate': '2.041e-06', 'num_tokens': '3.049e+06', 'mean_token_accuracy': '0.8234', 'epoch': '0.7961'}
+{'loss': '1.534', 'grad_norm': '12.12', 'learning_rate': '2.039e-06', 'num_tokens': '3.05e+06', 'mean_token_accuracy': '0.6766', 'epoch': '0.7963'}
+{'loss': '1.446', 'grad_norm': '38.5', 'learning_rate': '2.037e-06', 'num_tokens': '3.05e+06', 'mean_token_accuracy': '0.7253', 'epoch': '0.7965'}
+{'loss': '1.209', 'grad_norm': '13.44', 'learning_rate': '2.035e-06', 'num_tokens': '3.051e+06', 'mean_token_accuracy': '0.6514', 'epoch': '0.7966'}
+{'loss': '1.39', 'grad_norm': '33.75', 'learning_rate': '2.034e-06', 'num_tokens': '3.051e+06', 'mean_token_accuracy': '0.7444', 'epoch': '0.7968'}
+{'loss': '1.055', 'grad_norm': '18', 'learning_rate': '2.032e-06', 'num_tokens': '3.051e+06', 'mean_token_accuracy': '0.7616', 'epoch': '0.797'}
+{'loss': '1.969', 'grad_norm': '12.88', 'learning_rate': '2.03e-06', 'num_tokens': '3.052e+06', 'mean_token_accuracy': '0.608', 'epoch': '0.7972'}
+{'loss': '1.532', 'grad_norm': '39', 'learning_rate': '2.028e-06', 'num_tokens': '3.052e+06', 'mean_token_accuracy': '0.7154', 'epoch': '0.7973'}
+{'loss': '1.81', 'grad_norm': '16.25', 'learning_rate': '2.027e-06', 'num_tokens': '3.052e+06', 'mean_token_accuracy': '0.6369', 'epoch': '0.7975'}
+{'loss': '0.8313', 'grad_norm': '14.44', 'learning_rate': '2.025e-06', 'num_tokens': '3.053e+06', 'mean_token_accuracy': '0.7676', 'epoch': '0.7977'}
+{'loss': '1.872', 'grad_norm': '57.5', 'learning_rate': '2.023e-06', 'num_tokens': '3.053e+06', 'mean_token_accuracy': '0.5882', 'epoch': '0.7978'}
+{'loss': '1.059', 'grad_norm': '24', 'learning_rate': '2.022e-06', 'num_tokens': '3.053e+06', 'mean_token_accuracy': '0.8025', 'epoch': '0.798'}
+{'loss': '1.854', 'grad_norm': '27.5', 'learning_rate': '2.02e-06', 'num_tokens': '3.053e+06', 'mean_token_accuracy': '0.5758', 'epoch': '0.7982'}
+{'loss': '0.6418', 'grad_norm': '8.438', 'learning_rate': '2.018e-06', 'num_tokens': '3.054e+06', 'mean_token_accuracy': '0.8067', 'epoch': '0.7984'}
+{'loss': '1.342', 'grad_norm': '11', 'learning_rate': '2.016e-06', 'num_tokens': '3.055e+06', 'mean_token_accuracy': '0.6571', 'epoch': '0.7985'}
+{'loss': '1.358', 'grad_norm': '13.62', 'learning_rate': '2.015e-06', 'num_tokens': '3.056e+06', 'mean_token_accuracy': '0.648', 'epoch': '0.7987'}
+{'loss': '1.094', 'grad_norm': '13.56', 'learning_rate': '2.013e-06', 'num_tokens': '3.057e+06', 'mean_token_accuracy': '0.6691', 'epoch': '0.7989'}
+{'loss': '0.5198', 'grad_norm': '18.25', 'learning_rate': '2.011e-06', 'num_tokens': '3.057e+06', 'mean_token_accuracy': '0.8739', 'epoch': '0.7991'}
+{'loss': '1.645', 'grad_norm': '38.25', 'learning_rate': '2.009e-06', 'num_tokens': '3.057e+06', 'mean_token_accuracy': '0.6939', 'epoch': '0.7992'}
+{'loss': '1.493', 'grad_norm': '11.56', 'learning_rate': '2.008e-06', 'num_tokens': '3.058e+06', 'mean_token_accuracy': '0.6189', 'epoch': '0.7994'}
+{'loss': '1.202', 'grad_norm': '13.56', 'learning_rate': '2.006e-06', 'num_tokens': '3.058e+06', 'mean_token_accuracy': '0.6789', 'epoch': '0.7996'}
+{'loss': '0.5878', 'grad_norm': '14.62', 'learning_rate': '2.004e-06', 'num_tokens': '3.059e+06', 'mean_token_accuracy': '0.8639', 'epoch': '0.7998'}
+{'loss': '0.9378', 'grad_norm': '19.12', 'learning_rate': '2.002e-06', 'num_tokens': '3.059e+06', 'mean_token_accuracy': '0.7792', 'epoch': '0.7999'}
+{'loss': '1.756', 'grad_norm': '8.25', 'learning_rate': '2.001e-06', 'num_tokens': '3.061e+06', 'mean_token_accuracy': '0.6235', 'epoch': '0.8001'}
+{'loss': '1.219', 'grad_norm': '15.06', 'learning_rate': '1.999e-06', 'num_tokens': '3.061e+06', 'mean_token_accuracy': '0.7456', 'epoch': '0.8003'}
+{'loss': '0.5616', 'grad_norm': '7.438', 'learning_rate': '1.997e-06', 'num_tokens': '3.062e+06', 'mean_token_accuracy': '0.826', 'epoch': '0.8005'}
+{'loss': '1.359', 'grad_norm': '12.69', 'learning_rate': '1.995e-06', 'num_tokens': '3.063e+06', 'mean_token_accuracy': '0.6547', 'epoch': '0.8006'}
+
 80%|████████  | 4611/5758 [26:15<06:02,  3.17it/s]
 80%|████████  | 4612/5758 [26:16<06:08,  3.11it/s]
                                                   

 80%|████████  | 4612/5758 [26:16<06:08,  3.11it/s]
 80%|████████  | 4613/5758 [26:16<07:00,  2.72it/s]
                                                   

 80%|████████  | 4613/5758 [26:16<07:00,  2.72it/s]
 80%|████████  | 4614/5758 [26:16<06:36,  2.89it/s]
                                                   

 80%|████████  | 4614/5758 [26:16<06:36,  2.89it/s]
 80%|████████  | 4615/5758 [26:17<06:42,  2.84it/s]
                                                   

 80%|████████  | 4615/5758 [26:17<06:42,  2.84it/s]
 80%|████████  | 4616/5758 [26:17<06:23,  2.98it/s]
                                                   

 80%|████████  | 4616/5758 [26:17<06:23,  2.98it/s]
 80%|████████  | 4617/5758 [26:17<06:30,  2.92it/s]
                                                   

 80%|████████  | 4617/5758 [26:17<06:30,  2.92it/s]
 80%|████████  | 4618/5758 [26:18<06:29,  2.93it/s]
                                                   

 80%|████████  | 4618/5758 [26:18<06:29,  2.93it/s]
 80%|████████  | 4619/5758 [26:18<06:26,  2.95it/s]
                                                   

 80%|████████  | 4619/5758 [26:18<06:26,  2.95it/s]
 80%|████████  | 4620/5758 [26:18<06:15,  3.03it/s]
                                                   

 80%|████████  | 4620/5758 [26:18<06:15,  3.03it/s]
 80%|████████  | 4621/5758 [26:19<06:12,  3.05it/s]
                                                   

 80%|████████  | 4621/5758 [26:19<06:12,  3.05it/s]
 80%|████████  | 4622/5758 [26:19<06:28,  2.92it/s]
                                                   

 80%|████████  | 4622/5758 [26:19<06:28,  2.92it/s]
 80%|████████  | 4623/5758 [26:19<06:25,  2.94it/s]
                                                   

 80%|████████  | 4623/5758 [26:19<06:25,  2.94it/s]
 80%|████████  | 4624/5758 [26:20<06:38,  2.85it/s]
                                                   

 80%|████████  | 4624/5758 [26:20<06:38,  2.85it/s]
 80%|████████  | 4625/5758 [26:20<06:21,  2.97it/s]
                                                   

 80%|████████  | 4625/5758 [26:20<06:21,  2.97it/s]
 80%|████████  | 4626/5758 [26:20<06:20,  2.98it/s]
                                                   

 80%|████████  | 4626/5758 [26:20<06:20,  2.98it/s]
 80%|████████  | 4627/5758 [26:21<06:19,  2.98it/s]
                                                   

 80%|████████  | 4627/5758 [26:21<06:19,  2.98it/s]
 80%|████████  | 4628/5758 [26:21<06:08,  3.07it/s]
                                                   

 80%|████████  | 4628/5758 [26:21<06:08,  3.07it/s]
 80%|████████  | 4629/5758 [26:21<06:25,  2.93it/s]
                                                   

 80%|████████  | 4629/5758 [26:21<06:25,  2.93it/s]
 80%|████████  | 4630/5758 [26:22<06:39,  2.82it/s]
                                                   

 80%|████████  | 4630/5758 [26:22<06:39,  2.82it/s]
 80%|████████  | 4631/5758 [26:22<06:22,  2.95it/s]
                                                   

 80%|████████  | 4631/5758 [26:22<06:22,  2.95it/s]
 80%|████████  | 4632/5758 [26:22<06:09,  3.05it/s]
                                                   

 80%|████████  | 4632/5758 [26:22<06:09,  3.05it/s]
 80%|████████  | 4633/5758 [26:23<05:57,  3.15it/s]
                                                   

 80%|████████  | 4633/5758 [26:23<05:57,  3.15it/s]
 80%|████████  | 4634/5758 [26:23<06:03,  3.10it/s]
                                                   

 80%|████████  | 4634/5758 [26:23<06:03,  3.10it/s]
 80%|████████  | 4635/5758 [26:23<05:59,  3.13it/s]
                                                   

 80%|████████  | 4635/5758 [26:23<05:59,  3.13it/s]
 81%|████████  | 4636/5758 [26:24<06:00,  3.11it/s]
                                                   

 81%|████████  | 4636/5758 [26:24<06:00,  3.11it/s]
 81%|████████  | 4637/5758 [26:24<05:56,  3.15it/s]
                                                   

 81%|████████  | 4637/5758 [26:24<05:56,  3.15it/s]
 81%|████████  | 4638/5758 [26:24<06:47,  2.75it/s]
                                                   

 81%|████████  | 4638/5758 [26:25<06:47,  2.75it/s]
 81%|████████  | 4639/5758 [26:25<06:32,  2.85it/s]
                                                   

 81%|████████  | 4639/5758 [26:25<06:32,  2.85it/s]
 81%|████████  | 4640/5758 [26:25<06:19,  2.94it/s]
                                                   

 81%|████████  | 4640/5758 [26:25<06:19,  2.94it/s]
 81%|████████  | 4641/5758 [26:26<06:40,  2.79it/s]
                                                   

 81%|████████  | 4641/5758 [26:26<06:40,  2.79it/s]
 81%|████████  | 4642/5758 [26:26<06:36,  2.81it/s]
                                                   

 81%|████████  | 4642/5758 [26:26<06:36,  2.81it/s]
 81%|████████  | 4643/5758 [26:26<06:36,  2.81it/s]
                                                   

 81%|████████  | 4643/5758 [26:26<06:36,  2.81it/s]
 81%|████████  | 4644/5758 [26:27<06:32,  2.84it/s]
                                                   

 81%|████████  | 4644/5758 [26:27<06:32,  2.84it/s]
 81%|████████  | 4645/5758 [26:27<06:17,  2.95it/s]
                                                   

 81%|████████  | 4645/5758 [26:27<06:17,  2.95it/s]
 81%|████████  | 4646/5758 [26:27<06:23,  2.90it/s]
                                                   

 81%|████████  | 4646/5758 [26:27<06:23,  2.90it/s]
 81%|████████  | 4647/5758 [26:28<06:19,  2.93it/s]
                                                   

 81%|████████  | 4647/5758 [26:28<06:19,  2.93it/s]
 81%|████████  | 4648/5758 [26:28<06:14,  2.97it/s]
                                                   

 81%|████████  | 4648/5758 [26:28<06:14,  2.97it/s]
 81%|████████  | 4649/5758 [26:28<06:09,  3.00it/s]
                                                   

 81%|████████  | 4649/5758 [26:28<06:09,  3.00it/s]
 81%|████████  | 4650/5758 [26:29<06:02,  3.05it/s]
                                                   

 81%|████████  | 4650/5758 [26:29<06:02,  3.05it/s]
 81%|████████  | 4651/5758 [26:29<05:46,  3.19it/s]
                                                   

 81%|████████  | 4651/5758 [26:29<05:46,  3.19it/s]
 81%|████████  | 4652/5758 [26:29<06:01,  3.06it/s]
                                                   

 81%|████████  | 4652/5758 [26:29<06:01,  3.06it/s]
 81%|████████  | 4653/5758 [26:30<05:57,  3.09it/s]
                                                   

 81%|████████  | 4653/5758 [26:30<05:57,  3.09it/s]
 81%|████████  | 4654/5758 [26:30<06:00,  3.07it/s]
                                                   

 81%|████████  | 4654/5758 [26:30<06:00,  3.07it/s]
 81%|████████  | 4655/5758 [26:30<05:53,  3.12it/s]
                                                   

 81%|████████  | 4655/5758 [26:30<05:53,  3.12it/s]
 81%|████████  | 4656/5758 [26:30<05:49,  3.15it/s]
                                                   

 81%|████████  | 4656/5758 [26:30<05:49,  3.15it/s]
 81%|████████  | 4657/5758 [26:31<06:10,  2.97it/s]
                                                   

 81%|████████  | 4657/5758 [26:31<06:10,  2.97it/s]
 81%|████████  | 4658/5758 [26:31<06:00,  3.05it/s]
                                                   

 81%|████████  | 4658/5758 [26:31<06:00,  3.05it/s]
 81%|████████  | 4659/5758 [26:31<06:03,  3.02it/s]
                                                   

 81%|████████  | 4659/5758 [26:31<06:03,  3.02it/s]
 81%|████████  | 4660/5758 [26:32<05:54,  3.10it/s]
                                                   

 81%|████████  | 4660/5758 [26:32<05:54,  3.10it/s]
 81%|████████  | 4661/5758 [26:32<05:46,  3.17it/s]
                                                   

 81%|████████  | 4661/5758 [26:32<05:46,  3.17it/s]
 81%|████████  | 4662/5758 [26:32<05:52,  3.11it/s]
                                                   

 81%|████████  | 4662/5758 [26:32<05:52,  3.11it/s]
 81%|████████  | 4663/5758 [26:33<05:35,  3.26it/s]
                                                   

 81%|████████  | 4663/5758 [26:33<05:35,  3.26it/s]
 81%|████████  | 4664/5758 [26:33<06:25,  2.84it/s]
                                                   

 81%|████████  | 4664/5758 [26:33<06:25,  2.84it/s]
 81%|████████  | 4665/5758 [26:33<06:05,  2.99it/s]
                                                   

 81%|████████  | 4665/5758 [26:33<06:05,  2.99it/s]
 81%|████████  | 4666/5758 [26:34<06:05,  2.99it/s]
                                                   
{'loss': '1.136', 'grad_norm': '12.31', 'learning_rate': '1.994e-06', 'num_tokens': '3.064e+06', 'mean_token_accuracy': '0.6831', 'epoch': '0.8008'}
+{'loss': '1.09', 'grad_norm': '6.594', 'learning_rate': '1.992e-06', 'num_tokens': '3.066e+06', 'mean_token_accuracy': '0.7456', 'epoch': '0.801'}
+{'loss': '1.645', 'grad_norm': '14.94', 'learning_rate': '1.99e-06', 'num_tokens': '3.066e+06', 'mean_token_accuracy': '0.6623', 'epoch': '0.8011'}
+{'loss': '1.148', 'grad_norm': '13.62', 'learning_rate': '1.989e-06', 'num_tokens': '3.067e+06', 'mean_token_accuracy': '0.6651', 'epoch': '0.8013'}
+{'loss': '1.574', 'grad_norm': '17.5', 'learning_rate': '1.987e-06', 'num_tokens': '3.067e+06', 'mean_token_accuracy': '0.6673', 'epoch': '0.8015'}
+{'loss': '1.427', 'grad_norm': '14.12', 'learning_rate': '1.985e-06', 'num_tokens': '3.068e+06', 'mean_token_accuracy': '0.6517', 'epoch': '0.8017'}
+{'loss': '1.877', 'grad_norm': '15.5', 'learning_rate': '1.983e-06', 'num_tokens': '3.068e+06', 'mean_token_accuracy': '0.6277', 'epoch': '0.8018'}
+{'loss': '0.6348', 'grad_norm': '19.5', 'learning_rate': '1.982e-06', 'num_tokens': '3.069e+06', 'mean_token_accuracy': '0.8421', 'epoch': '0.802'}
+{'loss': '0.5699', 'grad_norm': '6.312', 'learning_rate': '1.98e-06', 'num_tokens': '3.07e+06', 'mean_token_accuracy': '0.8285', 'epoch': '0.8022'}
+{'loss': '1.407', 'grad_norm': '14.31', 'learning_rate': '1.978e-06', 'num_tokens': '3.071e+06', 'mean_token_accuracy': '0.6231', 'epoch': '0.8024'}
+{'loss': '1.119', 'grad_norm': '7.906', 'learning_rate': '1.976e-06', 'num_tokens': '3.073e+06', 'mean_token_accuracy': '0.7124', 'epoch': '0.8025'}
+{'loss': '2.412', 'grad_norm': '7.469', 'learning_rate': '1.975e-06', 'num_tokens': '3.075e+06', 'mean_token_accuracy': '0.4864', 'epoch': '0.8027'}
+{'loss': '1.332', 'grad_norm': '11.69', 'learning_rate': '1.973e-06', 'num_tokens': '3.076e+06', 'mean_token_accuracy': '0.6379', 'epoch': '0.8029'}
+{'loss': '0.9077', 'grad_norm': '34.5', 'learning_rate': '1.971e-06', 'num_tokens': '3.076e+06', 'mean_token_accuracy': '0.8065', 'epoch': '0.8031'}
+{'loss': '2', 'grad_norm': '12.81', 'learning_rate': '1.969e-06', 'num_tokens': '3.076e+06', 'mean_token_accuracy': '0.591', 'epoch': '0.8032'}
+{'loss': '1.189', 'grad_norm': '6.281', 'learning_rate': '1.968e-06', 'num_tokens': '3.08e+06', 'mean_token_accuracy': '0.7242', 'epoch': '0.8034'}
+{'loss': '1.538', 'grad_norm': '12.62', 'learning_rate': '1.966e-06', 'num_tokens': '3.08e+06', 'mean_token_accuracy': '0.5887', 'epoch': '0.8036'}
+{'loss': '1.427', 'grad_norm': '14.19', 'learning_rate': '1.964e-06', 'num_tokens': '3.081e+06', 'mean_token_accuracy': '0.6104', 'epoch': '0.8038'}
+{'loss': '1.858', 'grad_norm': '9.75', 'learning_rate': '1.962e-06', 'num_tokens': '3.082e+06', 'mean_token_accuracy': '0.6155', 'epoch': '0.8039'}
+{'loss': '1.211', 'grad_norm': '9.062', 'learning_rate': '1.961e-06', 'num_tokens': '3.083e+06', 'mean_token_accuracy': '0.7006', 'epoch': '0.8041'}
+{'loss': '1.44', 'grad_norm': '13.38', 'learning_rate': '1.959e-06', 'num_tokens': '3.084e+06', 'mean_token_accuracy': '0.6423', 'epoch': '0.8043'}
+{'loss': '0.9239', 'grad_norm': '11.88', 'learning_rate': '1.957e-06', 'num_tokens': '3.085e+06', 'mean_token_accuracy': '0.7374', 'epoch': '0.8044'}
+{'loss': '1.035', 'grad_norm': '35.75', 'learning_rate': '1.956e-06', 'num_tokens': '3.085e+06', 'mean_token_accuracy': '0.7206', 'epoch': '0.8046'}
+{'loss': '1.243', 'grad_norm': '7.5', 'learning_rate': '1.954e-06', 'num_tokens': '3.086e+06', 'mean_token_accuracy': '0.7229', 'epoch': '0.8048'}
+{'loss': '1.75', 'grad_norm': '14.81', 'learning_rate': '1.952e-06', 'num_tokens': '3.087e+06', 'mean_token_accuracy': '0.6494', 'epoch': '0.805'}
+{'loss': '1.113', 'grad_norm': '11.12', 'learning_rate': '1.95e-06', 'num_tokens': '3.088e+06', 'mean_token_accuracy': '0.6849', 'epoch': '0.8051'}
+{'loss': '1.182', 'grad_norm': '25.62', 'learning_rate': '1.949e-06', 'num_tokens': '3.088e+06', 'mean_token_accuracy': '0.7572', 'epoch': '0.8053'}
+{'loss': '0.7396', 'grad_norm': '14.38', 'learning_rate': '1.947e-06', 'num_tokens': '3.089e+06', 'mean_token_accuracy': '0.7692', 'epoch': '0.8055'}
+{'loss': '1.785', 'grad_norm': '15.44', 'learning_rate': '1.945e-06', 'num_tokens': '3.089e+06', 'mean_token_accuracy': '0.628', 'epoch': '0.8057'}
+{'loss': '0.9317', 'grad_norm': '18.62', 'learning_rate': '1.943e-06', 'num_tokens': '3.09e+06', 'mean_token_accuracy': '0.8011', 'epoch': '0.8058'}
+{'loss': '0.7051', 'grad_norm': '24.38', 'learning_rate': '1.942e-06', 'num_tokens': '3.09e+06', 'mean_token_accuracy': '0.8531', 'epoch': '0.806'}
+{'loss': '0.8195', 'grad_norm': '27.75', 'learning_rate': '1.94e-06', 'num_tokens': '3.09e+06', 'mean_token_accuracy': '0.8077', 'epoch': '0.8062'}
+{'loss': '1.265', 'grad_norm': '14.12', 'learning_rate': '1.938e-06', 'num_tokens': '3.09e+06', 'mean_token_accuracy': '0.7475', 'epoch': '0.8064'}
+{'loss': '1.2', 'grad_norm': '7.906', 'learning_rate': '1.936e-06', 'num_tokens': '3.092e+06', 'mean_token_accuracy': '0.7059', 'epoch': '0.8065'}
+{'loss': '1.392', 'grad_norm': '43.25', 'learning_rate': '1.935e-06', 'num_tokens': '3.092e+06', 'mean_token_accuracy': '0.6709', 'epoch': '0.8067'}
+{'loss': '1.338', 'grad_norm': '35.75', 'learning_rate': '1.933e-06', 'num_tokens': '3.092e+06', 'mean_token_accuracy': '0.6726', 'epoch': '0.8069'}
+{'loss': '0.9083', 'grad_norm': '10.5', 'learning_rate': '1.931e-06', 'num_tokens': '3.093e+06', 'mean_token_accuracy': '0.7437', 'epoch': '0.8071'}
+{'loss': '2.105', 'grad_norm': '10.75', 'learning_rate': '1.929e-06', 'num_tokens': '3.094e+06', 'mean_token_accuracy': '0.5725', 'epoch': '0.8072'}
+{'loss': '1.758', 'grad_norm': '31.12', 'learning_rate': '1.928e-06', 'num_tokens': '3.094e+06', 'mean_token_accuracy': '0.6142', 'epoch': '0.8074'}
+{'loss': '1.556', 'grad_norm': '13.19', 'learning_rate': '1.926e-06', 'num_tokens': '3.095e+06', 'mean_token_accuracy': '0.6069', 'epoch': '0.8076'}
+{'loss': '0.5876', 'grad_norm': '45.5', 'learning_rate': '1.924e-06', 'num_tokens': '3.095e+06', 'mean_token_accuracy': '0.8525', 'epoch': '0.8077'}
+{'loss': '1.296', 'grad_norm': '13.38', 'learning_rate': '1.923e-06', 'num_tokens': '3.095e+06', 'mean_token_accuracy': '0.7131', 'epoch': '0.8079'}
+{'loss': '1.157', 'grad_norm': '12.5', 'learning_rate': '1.921e-06', 'num_tokens': '3.096e+06', 'mean_token_accuracy': '0.6805', 'epoch': '0.8081'}
+{'loss': '1.626', 'grad_norm': '14', 'learning_rate': '1.919e-06', 'num_tokens': '3.097e+06', 'mean_token_accuracy': '0.603', 'epoch': '0.8083'}
+{'loss': '1.235', 'grad_norm': '12.44', 'learning_rate': '1.917e-06', 'num_tokens': '3.098e+06', 'mean_token_accuracy': '0.6719', 'epoch': '0.8084'}
+{'loss': '1.292', 'grad_norm': '14', 'learning_rate': '1.916e-06', 'num_tokens': '3.098e+06', 'mean_token_accuracy': '0.6595', 'epoch': '0.8086'}
+{'loss': '1.826', 'grad_norm': '18.25', 'learning_rate': '1.914e-06', 'num_tokens': '3.099e+06', 'mean_token_accuracy': '0.6218', 'epoch': '0.8088'}
+{'loss': '1.331', 'grad_norm': '13.88', 'learning_rate': '1.912e-06', 'num_tokens': '3.1e+06', 'mean_token_accuracy': '0.6444', 'epoch': '0.809'}
+{'loss': '0.8051', 'grad_norm': '14.31', 'learning_rate': '1.91e-06', 'num_tokens': '3.1e+06', 'mean_token_accuracy': '0.7868', 'epoch': '0.8091'}
+{'loss': '0.5256', 'grad_norm': '12.88', 'learning_rate': '1.909e-06', 'num_tokens': '3.101e+06', 'mean_token_accuracy': '0.8807', 'epoch': '0.8093'}
+{'loss': '1.385', 'grad_norm': '17.25', 'learning_rate': '1.907e-06', 'num_tokens': '3.101e+06', 'mean_token_accuracy': '0.6994', 'epoch': '0.8095'}
+{'loss': '1.752', 'grad_norm': '16.25', 'learning_rate': '1.905e-06', 'num_tokens': '3.101e+06', 'mean_token_accuracy': '0.5957', 'epoch': '0.8097'}
+{'loss': '2.185', 'grad_norm': '118.5', 'learning_rate': '1.903e-06', 'num_tokens': '3.101e+06', 'mean_token_accuracy': '0.5714', 'epoch': '0.8098'}
+{'loss': '2.203', 'grad_norm': '12', 'learning_rate': '1.902e-06', 'num_tokens': '3.102e+06', 'mean_token_accuracy': '0.5612', 'epoch': '0.81'}
+{'loss': '1.03', 'grad_norm': '12.69', 'learning_rate': '1.9e-06', 'num_tokens': '3.103e+06', 'mean_token_accuracy': '0.7034', 'epoch': '0.8102'}
+
 81%|████████  | 4666/5758 [26:34<06:05,  2.99it/s]
 81%|████████  | 4667/5758 [26:34<06:11,  2.93it/s]
                                                   

 81%|████████  | 4667/5758 [26:34<06:11,  2.93it/s]
 81%|████████  | 4668/5758 [26:34<06:00,  3.02it/s]
                                                   

 81%|████████  | 4668/5758 [26:34<06:00,  3.02it/s]
 81%|████████  | 4669/5758 [26:35<06:12,  2.92it/s]
                                                   

 81%|████████  | 4669/5758 [26:35<06:12,  2.92it/s]
 81%|████████  | 4670/5758 [26:35<06:09,  2.94it/s]
                                                   

 81%|████████  | 4670/5758 [26:35<06:09,  2.94it/s]
 81%|████████  | 4671/5758 [26:36<06:18,  2.87it/s]
                                                   

 81%|████████  | 4671/5758 [26:36<06:18,  2.87it/s]
 81%|████████  | 4672/5758 [26:36<06:06,  2.97it/s]
                                                   

 81%|████████  | 4672/5758 [26:36<06:06,  2.97it/s]
 81%|████████  | 4673/5758 [26:36<06:06,  2.96it/s]
                                                   

 81%|████████  | 4673/5758 [26:36<06:06,  2.96it/s]
 81%|████████  | 4674/5758 [26:36<05:55,  3.05it/s]
                                                   

 81%|████████  | 4674/5758 [26:36<05:55,  3.05it/s]
 81%|████████  | 4675/5758 [26:37<06:13,  2.90it/s]
                                                   

 81%|████████  | 4675/5758 [26:37<06:13,  2.90it/s]
 81%|████████  | 4676/5758 [26:37<06:25,  2.81it/s]
                                                   

 81%|████████  | 4676/5758 [26:37<06:25,  2.81it/s]
 81%|████████  | 4677/5758 [26:38<06:36,  2.73it/s]
                                                   

 81%|████████  | 4677/5758 [26:38<06:36,  2.73it/s]
 81%|████████  | 4678/5758 [26:38<06:14,  2.88it/s]
                                                   

 81%|████████  | 4678/5758 [26:38<06:14,  2.88it/s]
 81%|████████▏ | 4679/5758 [26:38<06:00,  2.99it/s]
                                                   

 81%|████████▏ | 4679/5758 [26:38<06:00,  2.99it/s]
 81%|████████▏ | 4680/5758 [26:39<06:03,  2.96it/s]
                                                   

 81%|████████▏ | 4680/5758 [26:39<06:03,  2.96it/s]
 81%|████████▏ | 4681/5758 [26:39<05:46,  3.11it/s]
                                                   

 81%|████████▏ | 4681/5758 [26:39<05:46,  3.11it/s]
 81%|████████▏ | 4682/5758 [26:39<05:48,  3.09it/s]
                                                   

 81%|████████▏ | 4682/5758 [26:39<05:48,  3.09it/s]
 81%|████████▏ | 4683/5758 [26:39<05:40,  3.16it/s]
                                                   

 81%|████████▏ | 4683/5758 [26:39<05:40,  3.16it/s]
 81%|████████▏ | 4684/5758 [26:40<05:50,  3.07it/s]
                                                   

 81%|████████▏ | 4684/5758 [26:40<05:50,  3.07it/s]
 81%|████████▏ | 4685/5758 [26:40<05:45,  3.10it/s]
                                                   

 81%|████████▏ | 4685/5758 [26:40<05:45,  3.10it/s]
 81%|████████▏ | 4686/5758 [26:40<05:46,  3.10it/s]
                                                   

 81%|████████▏ | 4686/5758 [26:40<05:46,  3.10it/s]
 81%|████████▏ | 4687/5758 [26:41<05:51,  3.05it/s]
                                                   

 81%|████████▏ | 4687/5758 [26:41<05:51,  3.05it/s]
 81%|████████▏ | 4688/5758 [26:41<06:07,  2.91it/s]
                                                   

 81%|████████▏ | 4688/5758 [26:41<06:07,  2.91it/s]
 81%|████████▏ | 4689/5758 [26:42<06:03,  2.94it/s]
                                                   

 81%|████████▏ | 4689/5758 [26:42<06:03,  2.94it/s]
 81%|████████▏ | 4690/5758 [26:42<07:21,  2.42it/s]
                                                   

 81%|████████▏ | 4690/5758 [26:42<07:21,  2.42it/s]
 81%|████████▏ | 4691/5758 [26:43<07:23,  2.41it/s]
                                                   

 81%|████████▏ | 4691/5758 [26:43<07:23,  2.41it/s]
 81%|████████▏ | 4692/5758 [26:43<06:52,  2.58it/s]
                                                   

 81%|████████▏ | 4692/5758 [26:43<06:52,  2.58it/s]
 82%|████████▏ | 4693/5758 [26:43<06:31,  2.72it/s]
                                                   

 82%|████████▏ | 4693/5758 [26:43<06:31,  2.72it/s]
 82%|████████▏ | 4694/5758 [26:44<06:31,  2.72it/s]
                                                   

 82%|████████▏ | 4694/5758 [26:44<06:31,  2.72it/s]
 82%|████████▏ | 4695/5758 [26:44<06:18,  2.81it/s]
                                                   

 82%|████████▏ | 4695/5758 [26:44<06:18,  2.81it/s]
 82%|████████▏ | 4696/5758 [26:44<06:07,  2.89it/s]
                                                   

 82%|████████▏ | 4696/5758 [26:44<06:07,  2.89it/s]
 82%|████████▏ | 4697/5758 [26:45<06:22,  2.78it/s]
                                                   

 82%|████████▏ | 4697/5758 [26:45<06:22,  2.78it/s]
 82%|████████▏ | 4698/5758 [26:45<06:22,  2.77it/s]
                                                   

 82%|████████▏ | 4698/5758 [26:45<06:22,  2.77it/s]
 82%|████████▏ | 4699/5758 [26:45<06:03,  2.91it/s]
                                                   

 82%|████████▏ | 4699/5758 [26:45<06:03,  2.91it/s]
 82%|████████▏ | 4700/5758 [26:46<05:54,  2.98it/s]
                                                   

 82%|████████▏ | 4700/5758 [26:46<05:54,  2.98it/s]
 82%|████████▏ | 4701/5758 [26:46<05:55,  2.98it/s]
                                                   

 82%|████████▏ | 4701/5758 [26:46<05:55,  2.98it/s]
 82%|████████▏ | 4702/5758 [26:46<05:54,  2.98it/s]
                                                   

 82%|████████▏ | 4702/5758 [26:46<05:54,  2.98it/s]
 82%|████████▏ | 4703/5758 [26:47<05:51,  3.00it/s]
                                                   

 82%|████████▏ | 4703/5758 [26:47<05:51,  3.00it/s]
 82%|████████▏ | 4704/5758 [26:47<05:40,  3.10it/s]
                                                   

 82%|████████▏ | 4704/5758 [26:47<05:40,  3.10it/s]
 82%|████████▏ | 4705/5758 [26:47<05:58,  2.94it/s]
                                                   

 82%|████████▏ | 4705/5758 [26:47<05:58,  2.94it/s]
 82%|████████▏ | 4706/5758 [26:48<06:03,  2.89it/s]
                                                   

 82%|████████▏ | 4706/5758 [26:48<06:03,  2.89it/s]
 82%|████████▏ | 4707/5758 [26:48<05:45,  3.04it/s]
                                                   

 82%|████████▏ | 4707/5758 [26:48<05:45,  3.04it/s]
 82%|████████▏ | 4708/5758 [26:48<05:42,  3.07it/s]
                                                   

 82%|████████▏ | 4708/5758 [26:48<05:42,  3.07it/s]
 82%|████████▏ | 4709/5758 [26:49<05:54,  2.96it/s]
                                                   

 82%|████████▏ | 4709/5758 [26:49<05:54,  2.96it/s]
 82%|████████▏ | 4710/5758 [26:49<06:14,  2.80it/s]
                                                   

 82%|████████▏ | 4710/5758 [26:49<06:14,  2.80it/s]
 82%|████████▏ | 4711/5758 [26:49<06:17,  2.77it/s]
                                                   

 82%|████████▏ | 4711/5758 [26:49<06:17,  2.77it/s]
 82%|████████▏ | 4712/5758 [26:50<06:02,  2.88it/s]
                                                   

 82%|████████▏ | 4712/5758 [26:50<06:02,  2.88it/s]
 82%|████████▏ | 4713/5758 [26:50<05:45,  3.03it/s]
                                                   

 82%|████████▏ | 4713/5758 [26:50<05:45,  3.03it/s]
 82%|████████▏ | 4714/5758 [26:50<05:41,  3.05it/s]
                                                   

 82%|████████▏ | 4714/5758 [26:50<05:41,  3.05it/s]
 82%|████████▏ | 4715/5758 [26:51<05:54,  2.94it/s]
                                                   

 82%|████████▏ | 4715/5758 [26:51<05:54,  2.94it/s]
 82%|████████▏ | 4716/5758 [26:51<05:49,  2.99it/s]
                                                   

 82%|████████▏ | 4716/5758 [26:51<05:49,  2.99it/s]
 82%|████████▏ | 4717/5758 [26:51<05:41,  3.05it/s]
                                                   

 82%|████████▏ | 4717/5758 [26:51<05:41,  3.05it/s]
 82%|████████▏ | 4718/5758 [26:52<06:34,  2.63it/s]
                                                   

 82%|████████▏ | 4718/5758 [26:52<06:34,  2.63it/s]
 82%|████████▏ | 4719/5758 [26:52<06:06,  2.84it/s]
                                                   

 82%|████████▏ | 4719/5758 [26:52<06:06,  2.84it/s]
 82%|████████▏ | 4720/5758 [26:52<05:54,  2.93it/s]
                                                   

 82%|████████▏ | 4720/5758 [26:52<05:54,  2.93it/s]
 82%|████████▏ | 4721/5758 [26:53<06:05,  2.84it/s]
                                                   
{'loss': '0.6621', 'grad_norm': '7.656', 'learning_rate': '1.898e-06', 'num_tokens': '3.105e+06', 'mean_token_accuracy': '0.8015', 'epoch': '0.8104'}
+{'loss': '1.219', 'grad_norm': '10.56', 'learning_rate': '1.896e-06', 'num_tokens': '3.106e+06', 'mean_token_accuracy': '0.6436', 'epoch': '0.8105'}
+{'loss': '1.796', 'grad_norm': '15.5', 'learning_rate': '1.895e-06', 'num_tokens': '3.106e+06', 'mean_token_accuracy': '0.5842', 'epoch': '0.8107'}
+{'loss': '2.036', 'grad_norm': '30.5', 'learning_rate': '1.893e-06', 'num_tokens': '3.107e+06', 'mean_token_accuracy': '0.5802', 'epoch': '0.8109'}
+{'loss': '0.5812', 'grad_norm': '6.781', 'learning_rate': '1.891e-06', 'num_tokens': '3.108e+06', 'mean_token_accuracy': '0.8221', 'epoch': '0.811'}
+{'loss': '1.207', 'grad_norm': '15.12', 'learning_rate': '1.89e-06', 'num_tokens': '3.109e+06', 'mean_token_accuracy': '0.6617', 'epoch': '0.8112'}
+{'loss': '1.538', 'grad_norm': '27.62', 'learning_rate': '1.888e-06', 'num_tokens': '3.109e+06', 'mean_token_accuracy': '0.6778', 'epoch': '0.8114'}
+{'loss': '1.152', 'grad_norm': '11.75', 'learning_rate': '1.886e-06', 'num_tokens': '3.11e+06', 'mean_token_accuracy': '0.6686', 'epoch': '0.8116'}
+{'loss': '1.965', 'grad_norm': '11.19', 'learning_rate': '1.884e-06', 'num_tokens': '3.111e+06', 'mean_token_accuracy': '0.5698', 'epoch': '0.8117'}
+{'loss': '1.127', 'grad_norm': '18.38', 'learning_rate': '1.883e-06', 'num_tokens': '3.111e+06', 'mean_token_accuracy': '0.7509', 'epoch': '0.8119'}
+{'loss': '1.817', 'grad_norm': '19.62', 'learning_rate': '1.881e-06', 'num_tokens': '3.111e+06', 'mean_token_accuracy': '0.5898', 'epoch': '0.8121'}
+{'loss': '0.9146', 'grad_norm': '31.88', 'learning_rate': '1.879e-06', 'num_tokens': '3.112e+06', 'mean_token_accuracy': '0.8148', 'epoch': '0.8123'}
+{'loss': '0.4878', 'grad_norm': '7.312', 'learning_rate': '1.877e-06', 'num_tokens': '3.113e+06', 'mean_token_accuracy': '0.8443', 'epoch': '0.8124'}
+{'loss': '1.263', 'grad_norm': '12.81', 'learning_rate': '1.876e-06', 'num_tokens': '3.114e+06', 'mean_token_accuracy': '0.6647', 'epoch': '0.8126'}
+{'loss': '1.894', 'grad_norm': '18.12', 'learning_rate': '1.874e-06', 'num_tokens': '3.114e+06', 'mean_token_accuracy': '0.5882', 'epoch': '0.8128'}
+{'loss': '0.7629', 'grad_norm': '13.38', 'learning_rate': '1.872e-06', 'num_tokens': '3.115e+06', 'mean_token_accuracy': '0.7791', 'epoch': '0.813'}
+{'loss': '0.9258', 'grad_norm': '10.81', 'learning_rate': '1.87e-06', 'num_tokens': '3.115e+06', 'mean_token_accuracy': '0.7271', 'epoch': '0.8131'}
+{'loss': '2.292', 'grad_norm': '22.75', 'learning_rate': '1.869e-06', 'num_tokens': '3.116e+06', 'mean_token_accuracy': '0.5106', 'epoch': '0.8133'}
+{'loss': '1.639', 'grad_norm': '15.88', 'learning_rate': '1.867e-06', 'num_tokens': '3.116e+06', 'mean_token_accuracy': '0.6625', 'epoch': '0.8135'}
+{'loss': '1.948', 'grad_norm': '26.88', 'learning_rate': '1.865e-06', 'num_tokens': '3.116e+06', 'mean_token_accuracy': '0.6203', 'epoch': '0.8137'}
+{'loss': '1.645', 'grad_norm': '23.75', 'learning_rate': '1.863e-06', 'num_tokens': '3.117e+06', 'mean_token_accuracy': '0.6263', 'epoch': '0.8138'}
+{'loss': '1.18', 'grad_norm': '18', 'learning_rate': '1.862e-06', 'num_tokens': '3.117e+06', 'mean_token_accuracy': '0.7377', 'epoch': '0.814'}
+{'loss': '1.953', 'grad_norm': '20.75', 'learning_rate': '1.86e-06', 'num_tokens': '3.117e+06', 'mean_token_accuracy': '0.5679', 'epoch': '0.8142'}
+{'loss': '0.577', 'grad_norm': '22.25', 'learning_rate': '1.858e-06', 'num_tokens': '3.118e+06', 'mean_token_accuracy': '0.8571', 'epoch': '0.8143'}
+{'loss': '1.54', 'grad_norm': '8.312', 'learning_rate': '1.857e-06', 'num_tokens': '3.12e+06', 'mean_token_accuracy': '0.6609', 'epoch': '0.8145'}
+{'loss': '0.9478', 'grad_norm': '21.62', 'learning_rate': '1.855e-06', 'num_tokens': '3.12e+06', 'mean_token_accuracy': '0.8151', 'epoch': '0.8147'}
+{'loss': '1.059', 'grad_norm': '11.88', 'learning_rate': '1.853e-06', 'num_tokens': '3.121e+06', 'mean_token_accuracy': '0.6925', 'epoch': '0.8149'}
+{'loss': '1.277', 'grad_norm': '15', 'learning_rate': '1.851e-06', 'num_tokens': '3.121e+06', 'mean_token_accuracy': '0.6928', 'epoch': '0.815'}
+{'loss': '0.7036', 'grad_norm': '10.75', 'learning_rate': '1.85e-06', 'num_tokens': '3.123e+06', 'mean_token_accuracy': '0.7905', 'epoch': '0.8152'}
+{'loss': '1.048', 'grad_norm': '11.38', 'learning_rate': '1.848e-06', 'num_tokens': '3.123e+06', 'mean_token_accuracy': '0.7494', 'epoch': '0.8154'}
+{'loss': '2.103', 'grad_norm': '15.12', 'learning_rate': '1.846e-06', 'num_tokens': '3.124e+06', 'mean_token_accuracy': '0.5826', 'epoch': '0.8156'}
+{'loss': '1.374', 'grad_norm': '40.5', 'learning_rate': '1.844e-06', 'num_tokens': '3.124e+06', 'mean_token_accuracy': '0.7197', 'epoch': '0.8157'}
+{'loss': '1.06', 'grad_norm': '13.94', 'learning_rate': '1.843e-06', 'num_tokens': '3.125e+06', 'mean_token_accuracy': '0.7606', 'epoch': '0.8159'}
+{'loss': '1.114', 'grad_norm': '8.875', 'learning_rate': '1.841e-06', 'num_tokens': '3.126e+06', 'mean_token_accuracy': '0.7232', 'epoch': '0.8161'}
+{'loss': '1.872', 'grad_norm': '11.44', 'learning_rate': '1.839e-06', 'num_tokens': '3.127e+06', 'mean_token_accuracy': '0.6066', 'epoch': '0.8163'}
+{'loss': '1.283', 'grad_norm': '7.562', 'learning_rate': '1.837e-06', 'num_tokens': '3.129e+06', 'mean_token_accuracy': '0.6796', 'epoch': '0.8164'}
+{'loss': '1.58', 'grad_norm': '24.12', 'learning_rate': '1.836e-06', 'num_tokens': '3.129e+06', 'mean_token_accuracy': '0.6575', 'epoch': '0.8166'}
+{'loss': '0.7119', 'grad_norm': '24.62', 'learning_rate': '1.834e-06', 'num_tokens': '3.129e+06', 'mean_token_accuracy': '0.8146', 'epoch': '0.8168'}
+{'loss': '1.196', 'grad_norm': '16.62', 'learning_rate': '1.832e-06', 'num_tokens': '3.129e+06', 'mean_token_accuracy': '0.727', 'epoch': '0.817'}
+{'loss': '0.9132', 'grad_norm': '13.5', 'learning_rate': '1.83e-06', 'num_tokens': '3.13e+06', 'mean_token_accuracy': '0.7104', 'epoch': '0.8171'}
+{'loss': '1.296', 'grad_norm': '15.25', 'learning_rate': '1.829e-06', 'num_tokens': '3.131e+06', 'mean_token_accuracy': '0.6693', 'epoch': '0.8173'}
+{'loss': '1.729', 'grad_norm': '25.88', 'learning_rate': '1.827e-06', 'num_tokens': '3.131e+06', 'mean_token_accuracy': '0.6194', 'epoch': '0.8175'}
+{'loss': '1.465', 'grad_norm': '8.875', 'learning_rate': '1.825e-06', 'num_tokens': '3.132e+06', 'mean_token_accuracy': '0.6714', 'epoch': '0.8176'}
+{'loss': '1.273', 'grad_norm': '14.94', 'learning_rate': '1.824e-06', 'num_tokens': '3.133e+06', 'mean_token_accuracy': '0.691', 'epoch': '0.8178'}
+{'loss': '0.8714', 'grad_norm': '24.88', 'learning_rate': '1.822e-06', 'num_tokens': '3.133e+06', 'mean_token_accuracy': '0.7708', 'epoch': '0.818'}
+{'loss': '0.7848', 'grad_norm': '35.75', 'learning_rate': '1.82e-06', 'num_tokens': '3.133e+06', 'mean_token_accuracy': '0.8279', 'epoch': '0.8182'}
+{'loss': '0.6062', 'grad_norm': '14.81', 'learning_rate': '1.818e-06', 'num_tokens': '3.133e+06', 'mean_token_accuracy': '0.8442', 'epoch': '0.8183'}
+{'loss': '1.082', 'grad_norm': '14.12', 'learning_rate': '1.817e-06', 'num_tokens': '3.134e+06', 'mean_token_accuracy': '0.6816', 'epoch': '0.8185'}
+{'loss': '1.098', 'grad_norm': '12.5', 'learning_rate': '1.815e-06', 'num_tokens': '3.135e+06', 'mean_token_accuracy': '0.6953', 'epoch': '0.8187'}
+{'loss': '1.711', 'grad_norm': '39', 'learning_rate': '1.813e-06', 'num_tokens': '3.135e+06', 'mean_token_accuracy': '0.632', 'epoch': '0.8189'}
+{'loss': '2.432', 'grad_norm': '11.69', 'learning_rate': '1.811e-06', 'num_tokens': '3.136e+06', 'mean_token_accuracy': '0.539', 'epoch': '0.819'}
+{'loss': '0.8693', 'grad_norm': '24', 'learning_rate': '1.81e-06', 'num_tokens': '3.136e+06', 'mean_token_accuracy': '0.797', 'epoch': '0.8192'}
+{'loss': '1.954', 'grad_norm': '13.88', 'learning_rate': '1.808e-06', 'num_tokens': '3.137e+06', 'mean_token_accuracy': '0.6054', 'epoch': '0.8194'}
+{'loss': '1.103', 'grad_norm': '11.81', 'learning_rate': '1.806e-06', 'num_tokens': '3.137e+06', 'mean_token_accuracy': '0.7217', 'epoch': '0.8196'}
+{'loss': '1.412', 'grad_norm': '13.75', 'learning_rate': '1.804e-06', 'num_tokens': '3.138e+06', 'mean_token_accuracy': '0.6352', 'epoch': '0.8197'}
+
 82%|████████▏ | 4721/5758 [26:53<06:05,  2.84it/s]
 82%|████████▏ | 4722/5758 [26:53<05:54,  2.92it/s]
                                                   

 82%|████████▏ | 4722/5758 [26:53<05:54,  2.92it/s]
 82%|████████▏ | 4723/5758 [26:53<05:54,  2.92it/s]
                                                   

 82%|████████▏ | 4723/5758 [26:53<05:54,  2.92it/s]
 82%|████████▏ | 4724/5758 [26:54<05:50,  2.95it/s]
                                                   

 82%|████████▏ | 4724/5758 [26:54<05:50,  2.95it/s]
 82%|████████▏ | 4725/5758 [26:54<05:51,  2.94it/s]
                                                   

 82%|████████▏ | 4725/5758 [26:54<05:51,  2.94it/s]
 82%|████████▏ | 4726/5758 [26:54<05:41,  3.02it/s]
                                                   

 82%|████████▏ | 4726/5758 [26:54<05:41,  3.02it/s]
 82%|████████▏ | 4727/5758 [26:55<05:41,  3.02it/s]
                                                   

 82%|████████▏ | 4727/5758 [26:55<05:41,  3.02it/s]
 82%|████████▏ | 4728/5758 [26:55<05:54,  2.91it/s]
                                                   

 82%|████████▏ | 4728/5758 [26:55<05:54,  2.91it/s]
 82%|████████▏ | 4729/5758 [26:55<05:49,  2.94it/s]
                                                   

 82%|████████▏ | 4729/5758 [26:55<05:49,  2.94it/s]
 82%|████████▏ | 4730/5758 [26:56<05:38,  3.04it/s]
                                                   

 82%|████████▏ | 4730/5758 [26:56<05:38,  3.04it/s]
 82%|████████▏ | 4731/5758 [26:56<05:46,  2.97it/s]
                                                   

 82%|████████▏ | 4731/5758 [26:56<05:46,  2.97it/s]
 82%|████████▏ | 4732/5758 [26:56<05:41,  3.00it/s]
                                                   

 82%|████████▏ | 4732/5758 [26:56<05:41,  3.00it/s]
 82%|████████▏ | 4733/5758 [26:57<05:53,  2.90it/s]
                                                   

 82%|████████▏ | 4733/5758 [26:57<05:53,  2.90it/s]
 82%|████████▏ | 4734/5758 [26:57<05:51,  2.92it/s]
                                                   

 82%|████████▏ | 4734/5758 [26:57<05:51,  2.92it/s]
 82%|████████▏ | 4735/5758 [26:58<06:04,  2.81it/s]
                                                   

 82%|████████▏ | 4735/5758 [26:58<06:04,  2.81it/s]
 82%|████████▏ | 4736/5758 [26:58<06:11,  2.75it/s]
                                                   

 82%|████████▏ | 4736/5758 [26:58<06:11,  2.75it/s]
 82%|████████▏ | 4737/5758 [26:58<06:06,  2.78it/s]
                                                   

 82%|████████▏ | 4737/5758 [26:58<06:06,  2.78it/s]
 82%|████████▏ | 4738/5758 [26:59<06:01,  2.83it/s]
                                                   

 82%|████████▏ | 4738/5758 [26:59<06:01,  2.83it/s]
 82%|████████▏ | 4739/5758 [26:59<06:12,  2.74it/s]
                                                   

 82%|████████▏ | 4739/5758 [26:59<06:12,  2.74it/s]
 82%|████████▏ | 4740/5758 [26:59<05:48,  2.92it/s]
                                                   

 82%|████████▏ | 4740/5758 [26:59<05:48,  2.92it/s]
 82%|████████▏ | 4741/5758 [27:00<05:59,  2.83it/s]
                                                   

 82%|████████▏ | 4741/5758 [27:00<05:59,  2.83it/s]
 82%|████████▏ | 4742/5758 [27:00<05:51,  2.89it/s]
                                                   

 82%|████████▏ | 4742/5758 [27:00<05:51,  2.89it/s]
 82%|████��███▏ | 4743/5758 [27:00<05:51,  2.89it/s]
                                                   

 82%|████████▏ | 4743/5758 [27:00<05:51,  2.89it/s]
 82%|████████▏ | 4744/5758 [27:01<05:49,  2.90it/s]
                                                   

 82%|████████▏ | 4744/5758 [27:01<05:49,  2.90it/s]
 82%|████████▏ | 4745/5758 [27:01<05:44,  2.94it/s]
                                                   

 82%|████████▏ | 4745/5758 [27:01<05:44,  2.94it/s]
 82%|████████▏ | 4746/5758 [27:01<06:07,  2.76it/s]
                                                   

 82%|████████▏ | 4746/5758 [27:01<06:07,  2.76it/s]
 82%|████████▏ | 4747/5758 [27:02<06:04,  2.77it/s]
                                                   

 82%|████████▏ | 4747/5758 [27:02<06:04,  2.77it/s]
 82%|████████▏ | 4748/5758 [27:02<06:07,  2.75it/s]
                                                   

 82%|████████▏ | 4748/5758 [27:02<06:07,  2.75it/s]
 82%|████████▏ | 4749/5758 [27:02<05:47,  2.90it/s]
                                                   

 82%|████████▏ | 4749/5758 [27:02<05:47,  2.90it/s]
 82%|████████▏ | 4750/5758 [27:03<05:55,  2.84it/s]
                                                   

 82%|████████▏ | 4750/5758 [27:03<05:55,  2.84it/s]
 83%|████████▎ | 4751/5758 [27:03<05:38,  2.97it/s]
                                                   

 83%|████████▎ | 4751/5758 [27:03<05:38,  2.97it/s]
 83%|████████▎ | 4752/5758 [27:03<05:36,  2.99it/s]
                                                   

 83%|████████▎ | 4752/5758 [27:03<05:36,  2.99it/s]
 83%|████████▎ | 4753/5758 [27:04<05:37,  2.98it/s]
                                                   

 83%|████████▎ | 4753/5758 [27:04<05:37,  2.98it/s]
 83%|████████▎ | 4754/5758 [27:04<05:31,  3.03it/s]
                                                   

 83%|████████▎ | 4754/5758 [27:04<05:31,  3.03it/s]
 83%|████████▎ | 4755/5758 [27:04<05:43,  2.92it/s]
                                                   

 83%|████████▎ | 4755/5758 [27:04<05:43,  2.92it/s]
 83%|████████▎ | 4756/5758 [27:05<05:28,  3.05it/s]
                                                   

 83%|████████▎ | 4756/5758 [27:05<05:28,  3.05it/s]
 83%|████████▎ | 4757/5758 [27:05<05:21,  3.12it/s]
                                                   

 83%|████████▎ | 4757/5758 [27:05<05:21,  3.12it/s]
 83%|████████▎ | 4758/5758 [27:05<05:24,  3.08it/s]
                                                   

 83%|████████▎ | 4758/5758 [27:05<05:24,  3.08it/s]
 83%|████████▎ | 4759/5758 [27:06<05:28,  3.04it/s]
                                                   

 83%|████████▎ | 4759/5758 [27:06<05:28,  3.04it/s]
 83%|████████▎ | 4760/5758 [27:06<05:38,  2.95it/s]
                                                   

 83%|████████▎ | 4760/5758 [27:06<05:38,  2.95it/s]
 83%|████████▎ | 4761/5758 [27:06<05:23,  3.08it/s]
                                                   

 83%|████████▎ | 4761/5758 [27:06<05:23,  3.08it/s]
 83%|████████▎ | 4762/5758 [27:07<05:16,  3.14it/s]
                                                   

 83%|████████▎ | 4762/5758 [27:07<05:16,  3.14it/s]
 83%|████████▎ | 4763/5758 [27:07<05:13,  3.17it/s]
                                                   

 83%|████████▎ | 4763/5758 [27:07<05:13,  3.17it/s]
 83%|████████▎ | 4764/5758 [27:07<05:05,  3.25it/s]
                                                   

 83%|████████▎ | 4764/5758 [27:07<05:05,  3.25it/s]
 83%|████████▎ | 4765/5758 [27:08<05:22,  3.08it/s]
                                                   

 83%|████████▎ | 4765/5758 [27:08<05:22,  3.08it/s]
 83%|████████▎ | 4766/5758 [27:08<05:15,  3.14it/s]
                                                   

 83%|████████▎ | 4766/5758 [27:08<05:15,  3.14it/s]
 83%|████████▎ | 4767/5758 [27:08<05:32,  2.98it/s]
                                                   

 83%|████████▎ | 4767/5758 [27:08<05:32,  2.98it/s]
 83%|████████▎ | 4768/5758 [27:09<05:26,  3.03it/s]
                                                   

 83%|████████▎ | 4768/5758 [27:09<05:26,  3.03it/s]
 83%|████████▎ | 4769/5758 [27:09<05:21,  3.08it/s]
                                                   

 83%|████████▎ | 4769/5758 [27:09<05:21,  3.08it/s]
 83%|████████▎ | 4770/5758 [27:09<05:27,  3.02it/s]
                                                   

 83%|████████▎ | 4770/5758 [27:09<05:27,  3.02it/s]
 83%|████████▎ | 4771/5758 [27:10<06:01,  2.73it/s]
                                                   

 83%|████████▎ | 4771/5758 [27:10<06:01,  2.73it/s]
 83%|████████▎ | 4772/5758 [27:10<06:03,  2.71it/s]
                                                   

 83%|████████▎ | 4772/5758 [27:10<06:03,  2.71it/s]
 83%|████████▎ | 4773/5758 [27:11<06:01,  2.73it/s]
                                                   

 83%|████████▎ | 4773/5758 [27:11<06:01,  2.73it/s]
 83%|████████▎ | 4774/5758 [27:11<05:49,  2.81it/s]
                                                   

 83%|████████▎ | 4774/5758 [27:11<05:49,  2.81it/s]
 83%|████████▎ | 4775/5758 [27:11<05:50,  2.81it/s]
                                                   

 83%|████████▎ | 4775/5758 [27:11<05:50,  2.81it/s]
 83%|████████▎ | 4776/5758 [27:12<05:40,  2.89it/s]
                                                   
{'loss': '1.075', 'grad_norm': '6.562', 'learning_rate': '1.803e-06', 'num_tokens': '3.14e+06', 'mean_token_accuracy': '0.7219', 'epoch': '0.8199'}
+{'loss': '0.8753', 'grad_norm': '15', 'learning_rate': '1.801e-06', 'num_tokens': '3.141e+06', 'mean_token_accuracy': '0.8175', 'epoch': '0.8201'}
+{'loss': '2.187', 'grad_norm': '19', 'learning_rate': '1.799e-06', 'num_tokens': '3.141e+06', 'mean_token_accuracy': '0.5841', 'epoch': '0.8203'}
+{'loss': '1.631', 'grad_norm': '10.69', 'learning_rate': '1.797e-06', 'num_tokens': '3.142e+06', 'mean_token_accuracy': '0.6433', 'epoch': '0.8204'}
+{'loss': '1.688', 'grad_norm': '9.75', 'learning_rate': '1.796e-06', 'num_tokens': '3.143e+06', 'mean_token_accuracy': '0.6375', 'epoch': '0.8206'}
+{'loss': '2.047', 'grad_norm': '17.75', 'learning_rate': '1.794e-06', 'num_tokens': '3.143e+06', 'mean_token_accuracy': '0.5957', 'epoch': '0.8208'}
+{'loss': '1.432', 'grad_norm': '14.31', 'learning_rate': '1.792e-06', 'num_tokens': '3.144e+06', 'mean_token_accuracy': '0.6842', 'epoch': '0.8209'}
+{'loss': '0.5157', 'grad_norm': '23.5', 'learning_rate': '1.791e-06', 'num_tokens': '3.144e+06', 'mean_token_accuracy': '0.8413', 'epoch': '0.8211'}
+{'loss': '0.513', 'grad_norm': '6.438', 'learning_rate': '1.789e-06', 'num_tokens': '3.146e+06', 'mean_token_accuracy': '0.8396', 'epoch': '0.8213'}
+{'loss': '0.6289', 'grad_norm': '7.656', 'learning_rate': '1.787e-06', 'num_tokens': '3.147e+06', 'mean_token_accuracy': '0.8052', 'epoch': '0.8215'}
+{'loss': '0.7792', 'grad_norm': '17', 'learning_rate': '1.785e-06', 'num_tokens': '3.147e+06', 'mean_token_accuracy': '0.824', 'epoch': '0.8216'}
+{'loss': '1.415', 'grad_norm': '17.88', 'learning_rate': '1.784e-06', 'num_tokens': '3.148e+06', 'mean_token_accuracy': '0.6927', 'epoch': '0.8218'}
+{'loss': '1.646', 'grad_norm': '15.19', 'learning_rate': '1.782e-06', 'num_tokens': '3.148e+06', 'mean_token_accuracy': '0.6402', 'epoch': '0.822'}
+{'loss': '1.575', 'grad_norm': '18.12', 'learning_rate': '1.78e-06', 'num_tokens': '3.149e+06', 'mean_token_accuracy': '0.672', 'epoch': '0.8222'}
+{'loss': '1.073', 'grad_norm': '20.12', 'learning_rate': '1.778e-06', 'num_tokens': '3.149e+06', 'mean_token_accuracy': '0.7759', 'epoch': '0.8223'}
+{'loss': '0.9452', 'grad_norm': '12.94', 'learning_rate': '1.777e-06', 'num_tokens': '3.149e+06', 'mean_token_accuracy': '0.7389', 'epoch': '0.8225'}
+{'loss': '0.8666', 'grad_norm': '17.75', 'learning_rate': '1.775e-06', 'num_tokens': '3.15e+06', 'mean_token_accuracy': '0.824', 'epoch': '0.8227'}
+{'loss': '1.327', 'grad_norm': '12.06', 'learning_rate': '1.773e-06', 'num_tokens': '3.151e+06', 'mean_token_accuracy': '0.614', 'epoch': '0.8229'}
+{'loss': '1.234', 'grad_norm': '13.88', 'learning_rate': '1.771e-06', 'num_tokens': '3.151e+06', 'mean_token_accuracy': '0.6644', 'epoch': '0.823'}
+{'loss': '1.966', 'grad_norm': '12.69', 'learning_rate': '1.77e-06', 'num_tokens': '3.152e+06', 'mean_token_accuracy': '0.608', 'epoch': '0.8232'}
+{'loss': '1.726', 'grad_norm': '18.12', 'learning_rate': '1.768e-06', 'num_tokens': '3.152e+06', 'mean_token_accuracy': '0.6084', 'epoch': '0.8234'}
+{'loss': '1.223', 'grad_norm': '20.88', 'learning_rate': '1.766e-06', 'num_tokens': '3.153e+06', 'mean_token_accuracy': '0.734', 'epoch': '0.8235'}
+{'loss': '1.782', 'grad_norm': '39.75', 'learning_rate': '1.765e-06', 'num_tokens': '3.153e+06', 'mean_token_accuracy': '0.6182', 'epoch': '0.8237'}
+{'loss': '0.9103', 'grad_norm': '9.438', 'learning_rate': '1.763e-06', 'num_tokens': '3.154e+06', 'mean_token_accuracy': '0.7808', 'epoch': '0.8239'}
+{'loss': '1.649', 'grad_norm': '21.38', 'learning_rate': '1.761e-06', 'num_tokens': '3.154e+06', 'mean_token_accuracy': '0.6927', 'epoch': '0.8241'}
+{'loss': '0.5684', 'grad_norm': '7.656', 'learning_rate': '1.759e-06', 'num_tokens': '3.155e+06', 'mean_token_accuracy': '0.8332', 'epoch': '0.8242'}
+{'loss': '0.7774', 'grad_norm': '10.38', 'learning_rate': '1.758e-06', 'num_tokens': '3.156e+06', 'mean_token_accuracy': '0.7589', 'epoch': '0.8244'}
+{'loss': '1.447', 'grad_norm': '9.25', 'learning_rate': '1.756e-06', 'num_tokens': '3.157e+06', 'mean_token_accuracy': '0.6611', 'epoch': '0.8246'}
+{'loss': '1.84', 'grad_norm': '14.88', 'learning_rate': '1.754e-06', 'num_tokens': '3.158e+06', 'mean_token_accuracy': '0.6265', 'epoch': '0.8248'}
+{'loss': '1.1', 'grad_norm': '9.188', 'learning_rate': '1.752e-06', 'num_tokens': '3.159e+06', 'mean_token_accuracy': '0.7437', 'epoch': '0.8249'}
+{'loss': '1.031', 'grad_norm': '13.5', 'learning_rate': '1.751e-06', 'num_tokens': '3.159e+06', 'mean_token_accuracy': '0.7094', 'epoch': '0.8251'}
+{'loss': '0.4186', 'grad_norm': '25.38', 'learning_rate': '1.749e-06', 'num_tokens': '3.16e+06', 'mean_token_accuracy': '0.9289', 'epoch': '0.8253'}
+{'loss': '1.2', 'grad_norm': '10.94', 'learning_rate': '1.747e-06', 'num_tokens': '3.16e+06', 'mean_token_accuracy': '0.6708', 'epoch': '0.8255'}
+{'loss': '1.226', 'grad_norm': '14', 'learning_rate': '1.745e-06', 'num_tokens': '3.161e+06', 'mean_token_accuracy': '0.6559', 'epoch': '0.8256'}
+{'loss': '1.091', 'grad_norm': '27', 'learning_rate': '1.744e-06', 'num_tokens': '3.161e+06', 'mean_token_accuracy': '0.7724', 'epoch': '0.8258'}
+{'loss': '1.077', 'grad_norm': '14.56', 'learning_rate': '1.742e-06', 'num_tokens': '3.162e+06', 'mean_token_accuracy': '0.6931', 'epoch': '0.826'}
+{'loss': '1.086', 'grad_norm': '13.06', 'learning_rate': '1.74e-06', 'num_tokens': '3.163e+06', 'mean_token_accuracy': '0.6926', 'epoch': '0.8262'}
+{'loss': '1.151', 'grad_norm': '15.94', 'learning_rate': '1.738e-06', 'num_tokens': '3.163e+06', 'mean_token_accuracy': '0.7466', 'epoch': '0.8263'}
+{'loss': '1.63', 'grad_norm': '23', 'learning_rate': '1.737e-06', 'num_tokens': '3.163e+06', 'mean_token_accuracy': '0.6784', 'epoch': '0.8265'}
+{'loss': '1.034', 'grad_norm': '18.62', 'learning_rate': '1.735e-06', 'num_tokens': '3.164e+06', 'mean_token_accuracy': '0.7452', 'epoch': '0.8267'}
+{'loss': '2.45', 'grad_norm': '22.88', 'learning_rate': '1.733e-06', 'num_tokens': '3.164e+06', 'mean_token_accuracy': '0.5217', 'epoch': '0.8268'}
+{'loss': '1.176', 'grad_norm': '13.88', 'learning_rate': '1.732e-06', 'num_tokens': '3.165e+06', 'mean_token_accuracy': '0.6741', 'epoch': '0.827'}
+{'loss': '1.247', 'grad_norm': '11.88', 'learning_rate': '1.73e-06', 'num_tokens': '3.165e+06', 'mean_token_accuracy': '0.6584', 'epoch': '0.8272'}
+{'loss': '1.405', 'grad_norm': '25.12', 'learning_rate': '1.728e-06', 'num_tokens': '3.166e+06', 'mean_token_accuracy': '0.6966', 'epoch': '0.8274'}
+{'loss': '0.8953', 'grad_norm': '13', 'learning_rate': '1.726e-06', 'num_tokens': '3.166e+06', 'mean_token_accuracy': '0.7922', 'epoch': '0.8275'}
+{'loss': '3.073', 'grad_norm': '15.06', 'learning_rate': '1.725e-06', 'num_tokens': '3.168e+06', 'mean_token_accuracy': '0.4663', 'epoch': '0.8277'}
+{'loss': '2.166', 'grad_norm': '17.62', 'learning_rate': '1.723e-06', 'num_tokens': '3.168e+06', 'mean_token_accuracy': '0.5671', 'epoch': '0.8279'}
+{'loss': '1.603', 'grad_norm': '19.62', 'learning_rate': '1.721e-06', 'num_tokens': '3.169e+06', 'mean_token_accuracy': '0.6115', 'epoch': '0.8281'}
+{'loss': '1.262', 'grad_norm': '14.38', 'learning_rate': '1.719e-06', 'num_tokens': '3.169e+06', 'mean_token_accuracy': '0.6387', 'epoch': '0.8282'}
+{'loss': '0.8157', 'grad_norm': '20.12', 'learning_rate': '1.718e-06', 'num_tokens': '3.17e+06', 'mean_token_accuracy': '0.8289', 'epoch': '0.8284'}
+{'loss': '1.353', 'grad_norm': '12.75', 'learning_rate': '1.716e-06', 'num_tokens': '3.17e+06', 'mean_token_accuracy': '0.651', 'epoch': '0.8286'}
+{'loss': '1.192', 'grad_norm': '37.5', 'learning_rate': '1.714e-06', 'num_tokens': '3.171e+06', 'mean_token_accuracy': '0.8103', 'epoch': '0.8288'}
+{'loss': '1.678', 'grad_norm': '20.88', 'learning_rate': '1.712e-06', 'num_tokens': '3.171e+06', 'mean_token_accuracy': '0.6746', 'epoch': '0.8289'}
+{'loss': '1.066', 'grad_norm': '5.25', 'learning_rate': '1.711e-06', 'num_tokens': '3.174e+06', 'mean_token_accuracy': '0.736', 'epoch': '0.8291'}
+{'loss': '0.952', 'grad_norm': '21.75', 'learning_rate': '1.709e-06', 'num_tokens': '3.174e+06', 'mean_token_accuracy': '0.7895', 'epoch': '0.8293'}
+
 83%|████████▎ | 4776/5758 [27:12<05:40,  2.89it/s]
 83%|████████▎ | 4777/5758 [27:12<05:32,  2.95it/s]
                                                   

 83%|████████▎ | 4777/5758 [27:12<05:32,  2.95it/s]
 83%|████████▎ | 4778/5758 [27:12<05:22,  3.04it/s]
                                                   

 83%|████████▎ | 4778/5758 [27:12<05:22,  3.04it/s]
 83%|████████▎ | 4779/5758 [27:12<05:21,  3.05it/s]
                                                   

 83%|████████▎ | 4779/5758 [27:12<05:21,  3.05it/s]
 83%|████████▎ | 4780/5758 [27:13<05:15,  3.10it/s]
                                                   

 83%|████████▎ | 4780/5758 [27:13<05:15,  3.10it/s]
 83%|████████▎ | 4781/5758 [27:13<05:30,  2.96it/s]
                                                   

 83%|████████▎ | 4781/5758 [27:13<05:30,  2.96it/s]
 83%|████████▎ | 4782/5758 [27:13<05:23,  3.02it/s]
                                                   

 83%|████████▎ | 4782/5758 [27:13<05:23,  3.02it/s]
 83%|████████▎ | 4783/5758 [27:14<05:38,  2.88it/s]
                                                   

 83%|████████▎ | 4783/5758 [27:14<05:38,  2.88it/s]
 83%|████████▎ | 4784/5758 [27:14<05:27,  2.97it/s]
                                                   

 83%|████████▎ | 4784/5758 [27:14<05:27,  2.97it/s]
 83%|████████▎ | 4785/5758 [27:14<05:15,  3.09it/s]
                                                   

 83%|████████▎ | 4785/5758 [27:14<05:15,  3.09it/s]
 83%|████████▎ | 4786/5758 [27:15<05:23,  3.00it/s]
                                                   

 83%|████████▎ | 4786/5758 [27:15<05:23,  3.00it/s]
 83%|████████▎ | 4787/5758 [27:15<05:16,  3.07it/s]
                                                   

 83%|████████▎ | 4787/5758 [27:15<05:16,  3.07it/s]
 83%|████████▎ | 4788/5758 [27:15<05:21,  3.02it/s]
                                                   

 83%|████████▎ | 4788/5758 [27:15<05:21,  3.02it/s]
 83%|████████▎ | 4789/5758 [27:16<05:09,  3.13it/s]
                                                   

 83%|████████▎ | 4789/5758 [27:16<05:09,  3.13it/s]
 83%|████████▎ | 4790/5758 [27:16<05:07,  3.15it/s]
                                                   

 83%|████████▎ | 4790/5758 [27:16<05:07,  3.15it/s]
 83%|████████▎ | 4791/5758 [27:16<05:09,  3.13it/s]
                                                   

 83%|████████▎ | 4791/5758 [27:16<05:09,  3.13it/s]
 83%|████████▎ | 4792/5758 [27:17<05:05,  3.16it/s]
                                                   

 83%|████████▎ | 4792/5758 [27:17<05:05,  3.16it/s]
 83%|████████▎ | 4793/5758 [27:17<05:02,  3.19it/s]
                                                   

 83%|████████▎ | 4793/5758 [27:17<05:02,  3.19it/s]
 83%|████████▎ | 4794/5758 [27:17<04:59,  3.22it/s]
                                                   

 83%|████████▎ | 4794/5758 [27:17<04:59,  3.22it/s]
 83%|████████▎ | 4795/5758 [27:18<04:58,  3.22it/s]
                                                   

 83%|████████▎ | 4795/5758 [27:18<04:58,  3.22it/s]
 83%|████████▎ | 4796/5758 [27:18<05:02,  3.18it/s]
                                                   

 83%|████████▎ | 4796/5758 [27:18<05:02,  3.18it/s]
 83%|████████▎ | 4797/5758 [27:18<05:56,  2.70it/s]
                                                   

 83%|████████▎ | 4797/5758 [27:18<05:56,  2.70it/s]
 83%|████████▎ | 4798/5758 [27:19<05:35,  2.86it/s]
                                                   

 83%|████████▎ | 4798/5758 [27:19<05:35,  2.86it/s]
 83%|████████▎ | 4799/5758 [27:19<05:22,  2.98it/s]
                                                   

 83%|████████▎ | 4799/5758 [27:19<05:22,  2.98it/s]
 83%|████████▎ | 4800/5758 [27:19<05:25,  2.94it/s]
                                                   

 83%|████████▎ | 4800/5758 [27:19<05:25,  2.94it/s]
 83%|████████▎ | 4801/5758 [27:20<05:44,  2.78it/s]
                                                   

 83%|████████▎ | 4801/5758 [27:20<05:44,  2.78it/s]
 83%|████████▎ | 4802/5758 [27:20<05:37,  2.83it/s]
                                                   

 83%|████████▎ | 4802/5758 [27:20<05:37,  2.83it/s]
 83%|████████▎ | 4803/5758 [27:21<05:41,  2.79it/s]
                                                   

 83%|████████▎ | 4803/5758 [27:21<05:41,  2.79it/s]
 83%|████████▎ | 4804/5758 [27:21<05:29,  2.89it/s]
                                                   

 83%|████████▎ | 4804/5758 [27:21<05:29,  2.89it/s]
 83%|████████▎ | 4805/5758 [27:21<05:23,  2.94it/s]
                                                   

 83%|████████▎ | 4805/5758 [27:21<05:23,  2.94it/s]
 83%|████████▎ | 4806/5758 [27:21<05:08,  3.09it/s]
                                                   

 83%|████████▎ | 4806/5758 [27:21<05:08,  3.09it/s]
 83%|████████▎ | 4807/5758 [27:22<05:05,  3.11it/s]
                                                   

 83%|████████▎ | 4807/5758 [27:22<05:05,  3.11it/s]
 84%|████████▎ | 4808/5758 [27:22<05:24,  2.93it/s]
                                                   

 84%|████████▎ | 4808/5758 [27:22<05:24,  2.93it/s]
 84%|████████▎ | 4809/5758 [27:22<05:13,  3.03it/s]
                                                   

 84%|████████▎ | 4809/5758 [27:22<05:13,  3.03it/s]
 84%|████████▎ | 4810/5758 [27:23<05:12,  3.04it/s]
                                                   

 84%|████████▎ | 4810/5758 [27:23<05:12,  3.04it/s]
 84%|████████▎ | 4811/5758 [27:23<05:05,  3.10it/s]
                                                   

 84%|████████▎ | 4811/5758 [27:23<05:05,  3.10it/s]
 84%|████████▎ | 4812/5758 [27:23<04:59,  3.16it/s]
                                                   

 84%|████████▎ | 4812/5758 [27:23<04:59,  3.16it/s]
 84%|████████▎ | 4813/5758 [27:24<05:07,  3.07it/s]
                                                   

 84%|████████▎ | 4813/5758 [27:24<05:07,  3.07it/s]
 84%|████████▎ | 4814/5758 [27:24<05:10,  3.04it/s]
                                                   

 84%|████████▎ | 4814/5758 [27:24<05:10,  3.04it/s]
 84%|████████▎ | 4815/5758 [27:24<05:12,  3.02it/s]
                                                   

 84%|████████▎ | 4815/5758 [27:24<05:12,  3.02it/s]
 84%|████████▎ | 4816/5758 [27:25<05:10,  3.03it/s]
                                                   

 84%|████████▎ | 4816/5758 [27:25<05:10,  3.03it/s]
 84%|████████▎ | 4817/5758 [27:25<05:07,  3.06it/s]
                                                   

 84%|████████▎ | 4817/5758 [27:25<05:07,  3.06it/s]
 84%|████████▎ | 4818/5758 [27:25<05:12,  3.00it/s]
                                                   

 84%|████████▎ | 4818/5758 [27:25<05:12,  3.00it/s]
 84%|████████▎ | 4819/5758 [27:26<05:16,  2.97it/s]
                                                   

 84%|████████▎ | 4819/5758 [27:26<05:16,  2.97it/s]
 84%|████████▎ | 4820/5758 [27:26<05:09,  3.03it/s]
                                                   

 84%|████████▎ | 4820/5758 [27:26<05:09,  3.03it/s]
 84%|████████▎ | 4821/5758 [27:26<05:14,  2.98it/s]
                                                   

 84%|████████▎ | 4821/5758 [27:26<05:14,  2.98it/s]
 84%|████████▎ | 4822/5758 [27:27<05:19,  2.93it/s]
                                                   

 84%|████████▎ | 4822/5758 [27:27<05:19,  2.93it/s]
 84%|████████▍ | 4823/5758 [27:27<05:53,  2.65it/s]
                                                   

 84%|████████▍ | 4823/5758 [27:27<05:53,  2.65it/s]
 84%|████████▍ | 4824/5758 [27:28<05:38,  2.76it/s]
                                                   

 84%|████████▍ | 4824/5758 [27:28<05:38,  2.76it/s]
 84%|████████▍ | 4825/5758 [27:28<05:19,  2.92it/s]
                                                   

 84%|████████▍ | 4825/5758 [27:28<05:19,  2.92it/s]
 84%|████████▍ | 4826/5758 [27:28<05:33,  2.79it/s]
                                                   

 84%|████████▍ | 4826/5758 [27:28<05:33,  2.79it/s]
 84%|████████▍ | 4827/5758 [27:29<05:29,  2.82it/s]
                                                   

 84%|████████▍ | 4827/5758 [27:29<05:29,  2.82it/s]
 84%|████████▍ | 4828/5758 [27:29<05:30,  2.81it/s]
                                                   

 84%|████████▍ | 4828/5758 [27:29<05:30,  2.81it/s]
 84%|████████▍ | 4829/5758 [27:29<05:20,  2.90it/s]
                                                   

 84%|████████▍ | 4829/5758 [27:29<05:20,  2.90it/s]
 84%|████████▍ | 4830/5758 [27:30<05:26,  2.84it/s]
                                                   

 84%|████████▍ | 4830/5758 [27:30<05:26,  2.84it/s]
 84%|████████▍ | 4831/5758 [27:30<05:19,  2.90it/s]
                                                   
{'loss': '1.196', 'grad_norm': '13.25', 'learning_rate': '1.707e-06', 'num_tokens': '3.175e+06', 'mean_token_accuracy': '0.6692', 'epoch': '0.8295'}
+{'loss': '0.8427', 'grad_norm': '13.94', 'learning_rate': '1.705e-06', 'num_tokens': '3.175e+06', 'mean_token_accuracy': '0.8191', 'epoch': '0.8296'}
+{'loss': '1.093', 'grad_norm': '5.812', 'learning_rate': '1.704e-06', 'num_tokens': '3.178e+06', 'mean_token_accuracy': '0.7408', 'epoch': '0.8298'}
+{'loss': '1.072', 'grad_norm': '11.62', 'learning_rate': '1.702e-06', 'num_tokens': '3.179e+06', 'mean_token_accuracy': '0.6955', 'epoch': '0.83'}
+{'loss': '2.2', 'grad_norm': '12.5', 'learning_rate': '1.7e-06', 'num_tokens': '3.18e+06', 'mean_token_accuracy': '0.5579', 'epoch': '0.8301'}
+{'loss': '0.7721', 'grad_norm': '13.31', 'learning_rate': '1.699e-06', 'num_tokens': '3.18e+06', 'mean_token_accuracy': '0.7545', 'epoch': '0.8303'}
+{'loss': '0.3752', 'grad_norm': '16.75', 'learning_rate': '1.697e-06', 'num_tokens': '3.18e+06', 'mean_token_accuracy': '0.9189', 'epoch': '0.8305'}
+{'loss': '0.6376', 'grad_norm': '9.312', 'learning_rate': '1.695e-06', 'num_tokens': '3.182e+06', 'mean_token_accuracy': '0.8104', 'epoch': '0.8307'}
+{'loss': '1.93', 'grad_norm': '15.75', 'learning_rate': '1.693e-06', 'num_tokens': '3.182e+06', 'mean_token_accuracy': '0.5992', 'epoch': '0.8308'}
+{'loss': '1.131', 'grad_norm': '13.75', 'learning_rate': '1.692e-06', 'num_tokens': '3.183e+06', 'mean_token_accuracy': '0.6606', 'epoch': '0.831'}
+{'loss': '1.921', 'grad_norm': '23', 'learning_rate': '1.69e-06', 'num_tokens': '3.183e+06', 'mean_token_accuracy': '0.556', 'epoch': '0.8312'}
+{'loss': '1.398', 'grad_norm': '14.31', 'learning_rate': '1.688e-06', 'num_tokens': '3.184e+06', 'mean_token_accuracy': '0.6223', 'epoch': '0.8314'}
+{'loss': '0.9393', 'grad_norm': '26.62', 'learning_rate': '1.686e-06', 'num_tokens': '3.184e+06', 'mean_token_accuracy': '0.7843', 'epoch': '0.8315'}
+{'loss': '1.367', 'grad_norm': '15.38', 'learning_rate': '1.685e-06', 'num_tokens': '3.185e+06', 'mean_token_accuracy': '0.6408', 'epoch': '0.8317'}
+{'loss': '2.054', 'grad_norm': '11.69', 'learning_rate': '1.683e-06', 'num_tokens': '3.185e+06', 'mean_token_accuracy': '0.5744', 'epoch': '0.8319'}
+{'loss': '1.936', 'grad_norm': '24.75', 'learning_rate': '1.681e-06', 'num_tokens': '3.185e+06', 'mean_token_accuracy': '0.5825', 'epoch': '0.8321'}
+{'loss': '1.164', 'grad_norm': '16.62', 'learning_rate': '1.679e-06', 'num_tokens': '3.186e+06', 'mean_token_accuracy': '0.739', 'epoch': '0.8322'}
+{'loss': '1.57', 'grad_norm': '12.75', 'learning_rate': '1.678e-06', 'num_tokens': '3.187e+06', 'mean_token_accuracy': '0.6653', 'epoch': '0.8324'}
+{'loss': '1.02', 'grad_norm': '11.81', 'learning_rate': '1.676e-06', 'num_tokens': '3.187e+06', 'mean_token_accuracy': '0.7145', 'epoch': '0.8326'}
+{'loss': '2.068', 'grad_norm': '11.88', 'learning_rate': '1.674e-06', 'num_tokens': '3.188e+06', 'mean_token_accuracy': '0.5565', 'epoch': '0.8328'}
+{'loss': '1.649', 'grad_norm': '13.25', 'learning_rate': '1.672e-06', 'num_tokens': '3.189e+06', 'mean_token_accuracy': '0.6655', 'epoch': '0.8329'}
+{'loss': '0.6718', 'grad_norm': '6.062', 'learning_rate': '1.671e-06', 'num_tokens': '3.191e+06', 'mean_token_accuracy': '0.7949', 'epoch': '0.8331'}
+{'loss': '1.814', 'grad_norm': '9', 'learning_rate': '1.669e-06', 'num_tokens': '3.192e+06', 'mean_token_accuracy': '0.6258', 'epoch': '0.8333'}
+{'loss': '1.08', 'grad_norm': '11.88', 'learning_rate': '1.667e-06', 'num_tokens': '3.193e+06', 'mean_token_accuracy': '0.7147', 'epoch': '0.8334'}
+{'loss': '1.597', 'grad_norm': '10.19', 'learning_rate': '1.666e-06', 'num_tokens': '3.194e+06', 'mean_token_accuracy': '0.6418', 'epoch': '0.8336'}
+{'loss': '0.9357', 'grad_norm': '21.25', 'learning_rate': '1.664e-06', 'num_tokens': '3.194e+06', 'mean_token_accuracy': '0.8082', 'epoch': '0.8338'}
+{'loss': '0.5258', 'grad_norm': '16.38', 'learning_rate': '1.662e-06', 'num_tokens': '3.194e+06', 'mean_token_accuracy': '0.9', 'epoch': '0.834'}
+{'loss': '0.9548', 'grad_norm': '13.25', 'learning_rate': '1.66e-06', 'num_tokens': '3.195e+06', 'mean_token_accuracy': '0.7519', 'epoch': '0.8341'}
+{'loss': '1.157', 'grad_norm': '11.69', 'learning_rate': '1.659e-06', 'num_tokens': '3.196e+06', 'mean_token_accuracy': '0.7051', 'epoch': '0.8343'}
+{'loss': '1.427', 'grad_norm': '14.06', 'learning_rate': '1.657e-06', 'num_tokens': '3.196e+06', 'mean_token_accuracy': '0.6899', 'epoch': '0.8345'}
+{'loss': '0.8951', 'grad_norm': '12.5', 'learning_rate': '1.655e-06', 'num_tokens': '3.197e+06', 'mean_token_accuracy': '0.7525', 'epoch': '0.8347'}
+{'loss': '1.441', 'grad_norm': '30.75', 'learning_rate': '1.653e-06', 'num_tokens': '3.197e+06', 'mean_token_accuracy': '0.7674', 'epoch': '0.8348'}
+{'loss': '1.194', 'grad_norm': '13.12', 'learning_rate': '1.652e-06', 'num_tokens': '3.197e+06', 'mean_token_accuracy': '0.7182', 'epoch': '0.835'}
+{'loss': '1.319', 'grad_norm': '14.75', 'learning_rate': '1.65e-06', 'num_tokens': '3.198e+06', 'mean_token_accuracy': '0.6517', 'epoch': '0.8352'}
+{'loss': '1.112', 'grad_norm': '19.5', 'learning_rate': '1.648e-06', 'num_tokens': '3.198e+06', 'mean_token_accuracy': '0.7488', 'epoch': '0.8354'}
+{'loss': '2.055', 'grad_norm': '12.31', 'learning_rate': '1.646e-06', 'num_tokens': '3.199e+06', 'mean_token_accuracy': '0.5471', 'epoch': '0.8355'}
+{'loss': '0.6717', 'grad_norm': '40.5', 'learning_rate': '1.645e-06', 'num_tokens': '3.199e+06', 'mean_token_accuracy': '0.8553', 'epoch': '0.8357'}
+{'loss': '1.275', 'grad_norm': '11.56', 'learning_rate': '1.643e-06', 'num_tokens': '3.2e+06', 'mean_token_accuracy': '0.6866', 'epoch': '0.8359'}
+{'loss': '1.986', 'grad_norm': '8', 'learning_rate': '1.641e-06', 'num_tokens': '3.202e+06', 'mean_token_accuracy': '0.5934', 'epoch': '0.8361'}
+{'loss': '1.76', 'grad_norm': '16.5', 'learning_rate': '1.639e-06', 'num_tokens': '3.202e+06', 'mean_token_accuracy': '0.5788', 'epoch': '0.8362'}
+{'loss': '1.035', 'grad_norm': '11.81', 'learning_rate': '1.638e-06', 'num_tokens': '3.203e+06', 'mean_token_accuracy': '0.7162', 'epoch': '0.8364'}
+{'loss': '1.792', 'grad_norm': '17.38', 'learning_rate': '1.636e-06', 'num_tokens': '3.204e+06', 'mean_token_accuracy': '0.6276', 'epoch': '0.8366'}
+{'loss': '2.243', 'grad_norm': '26.38', 'learning_rate': '1.634e-06', 'num_tokens': '3.204e+06', 'mean_token_accuracy': '0.5415', 'epoch': '0.8367'}
+{'loss': '1.102', 'grad_norm': '12.44', 'learning_rate': '1.633e-06', 'num_tokens': '3.205e+06', 'mean_token_accuracy': '0.6949', 'epoch': '0.8369'}
+{'loss': '1.871', 'grad_norm': '26.25', 'learning_rate': '1.631e-06', 'num_tokens': '3.205e+06', 'mean_token_accuracy': '0.6129', 'epoch': '0.8371'}
+{'loss': '1.347', 'grad_norm': '11', 'learning_rate': '1.629e-06', 'num_tokens': '3.206e+06', 'mean_token_accuracy': '0.6418', 'epoch': '0.8373'}
+{'loss': '1.215', 'grad_norm': '10.88', 'learning_rate': '1.627e-06', 'num_tokens': '3.207e+06', 'mean_token_accuracy': '0.6405', 'epoch': '0.8374'}
+{'loss': '1.487', 'grad_norm': '9.438', 'learning_rate': '1.626e-06', 'num_tokens': '3.209e+06', 'mean_token_accuracy': '0.6723', 'epoch': '0.8376'}
+{'loss': '1.268', 'grad_norm': '12.12', 'learning_rate': '1.624e-06', 'num_tokens': '3.209e+06', 'mean_token_accuracy': '0.6485', 'epoch': '0.8378'}
+{'loss': '1.222', 'grad_norm': '6.594', 'learning_rate': '1.622e-06', 'num_tokens': '3.212e+06', 'mean_token_accuracy': '0.7323', 'epoch': '0.838'}
+{'loss': '1.803', 'grad_norm': '31.25', 'learning_rate': '1.62e-06', 'num_tokens': '3.212e+06', 'mean_token_accuracy': '0.6029', 'epoch': '0.8381'}
+{'loss': '1.339', 'grad_norm': '15.88', 'learning_rate': '1.619e-06', 'num_tokens': '3.212e+06', 'mean_token_accuracy': '0.6364', 'epoch': '0.8383'}
+{'loss': '1.778', 'grad_norm': '19.12', 'learning_rate': '1.617e-06', 'num_tokens': '3.213e+06', 'mean_token_accuracy': '0.6012', 'epoch': '0.8385'}
+{'loss': '2.224', 'grad_norm': '8.812', 'learning_rate': '1.615e-06', 'num_tokens': '3.214e+06', 'mean_token_accuracy': '0.5274', 'epoch': '0.8387'}
+{'loss': '0.7663', 'grad_norm': '31.62', 'learning_rate': '1.613e-06', 'num_tokens': '3.214e+06', 'mean_token_accuracy': '0.8136', 'epoch': '0.8388'}
+
 84%|████████▍ | 4831/5758 [27:30<05:19,  2.90it/s]
 84%|████████▍ | 4832/5758 [27:30<05:10,  2.99it/s]
                                                   

 84%|████████▍ | 4832/5758 [27:30<05:10,  2.99it/s]
 84%|████████▍ | 4833/5758 [27:31<05:05,  3.02it/s]
                                                   

 84%|████████▍ | 4833/5758 [27:31<05:05,  3.02it/s]
 84%|████████▍ | 4834/5758 [27:31<05:19,  2.89it/s]
                                                   

 84%|████████▍ | 4834/5758 [27:31<05:19,  2.89it/s]
 84%|████████▍ | 4835/5758 [27:31<05:14,  2.93it/s]
                                                   

 84%|████████▍ | 4835/5758 [27:31<05:14,  2.93it/s]
 84%|████████▍ | 4836/5758 [27:32<05:03,  3.04it/s]
                                                   

 84%|████████▍ | 4836/5758 [27:32<05:03,  3.04it/s]
 84%|████████▍ | 4837/5758 [27:32<05:05,  3.02it/s]
                                                   

 84%|████████▍ | 4837/5758 [27:32<05:05,  3.02it/s]
 84%|████████▍ | 4838/5758 [27:32<05:07,  2.99it/s]
                                                   

 84%|████████▍ | 4838/5758 [27:32<05:07,  2.99it/s]
 84%|████████▍ | 4839/5758 [27:33<05:26,  2.81it/s]
                                                   

 84%|████████▍ | 4839/5758 [27:33<05:26,  2.81it/s]
 84%|████████▍ | 4840/5758 [27:33<05:31,  2.77it/s]
                                                   

 84%|████████▍ | 4840/5758 [27:33<05:31,  2.77it/s]
 84%|████████▍ | 4841/5758 [27:33<05:19,  2.87it/s]
                                                   

 84%|████████▍ | 4841/5758 [27:33<05:19,  2.87it/s]
 84%|████████▍ | 4842/5758 [27:34<05:16,  2.89it/s]
                                                   

 84%|████████▍ | 4842/5758 [27:34<05:16,  2.89it/s]
 84%|████████▍ | 4843/5758 [27:34<05:26,  2.81it/s]
                                                   

 84%|███��████▍ | 4843/5758 [27:34<05:26,  2.81it/s]
 84%|████████▍ | 4844/5758 [27:34<05:06,  2.98it/s]
                                                   

 84%|████████▍ | 4844/5758 [27:34<05:06,  2.98it/s]
 84%|████████▍ | 4845/5758 [27:35<05:05,  2.99it/s]
                                                   

 84%|████████▍ | 4845/5758 [27:35<05:05,  2.99it/s]
 84%|████████▍ | 4846/5758 [27:35<04:58,  3.05it/s]
                                                   

 84%|████████▍ | 4846/5758 [27:35<04:58,  3.05it/s]
 84%|████████▍ | 4847/5758 [27:35<05:12,  2.91it/s]
                                                   

 84%|████████▍ | 4847/5758 [27:35<05:12,  2.91it/s]
 84%|████████▍ | 4848/5758 [27:36<05:13,  2.91it/s]
                                                   

 84%|████████▍ | 4848/5758 [27:36<05:13,  2.91it/s]
 84%|████████▍ | 4849/5758 [27:36<05:50,  2.59it/s]
                                                   

 84%|████████▍ | 4849/5758 [27:36<05:50,  2.59it/s]
 84%|████████▍ | 4850/5758 [27:37<05:26,  2.78it/s]
                                                   

 84%|████████▍ | 4850/5758 [27:37<05:26,  2.78it/s]
 84%|████████▍ | 4851/5758 [27:37<05:29,  2.75it/s]
                                                   

 84%|████████▍ | 4851/5758 [27:37<05:29,  2.75it/s]
 84%|████████▍ | 4852/5758 [27:37<05:10,  2.91it/s]
                                                   

 84%|████████▍ | 4852/5758 [27:37<05:10,  2.91it/s]
 84%|████████▍ | 4853/5758 [27:38<05:05,  2.96it/s]
                                                   

 84%|████████▍ | 4853/5758 [27:38<05:05,  2.96it/s]
 84%|████████▍ | 4854/5758 [27:38<05:10,  2.91it/s]
                                                   

 84%|████████▍ | 4854/5758 [27:38<05:10,  2.91it/s]
 84%|████████▍ | 4855/5758 [27:38<04:59,  3.01it/s]
                                                   

 84%|████████▍ | 4855/5758 [27:38<04:59,  3.01it/s]
 84%|████████▍ | 4856/5758 [27:38<04:50,  3.11it/s]
                                                   

 84%|████████▍ | 4856/5758 [27:39<04:50,  3.11it/s]
 84%|████████▍ | 4857/5758 [27:39<04:52,  3.08it/s]
                                                   

 84%|████████▍ | 4857/5758 [27:39<04:52,  3.08it/s]
 84%|████████▍ | 4858/5758 [27:39<04:47,  3.13it/s]
                                                   

 84%|████████▍ | 4858/5758 [27:39<04:47,  3.13it/s]
 84%|████████▍ | 4859/5758 [27:39<04:45,  3.14it/s]
                                                   

 84%|████████▍ | 4859/5758 [27:39<04:45,  3.14it/s]
 84%|████████▍ | 4860/5758 [27:40<04:51,  3.08it/s]
                                                   

 84%|████████▍ | 4860/5758 [27:40<04:51,  3.08it/s]
 84%|████████▍ | 4861/5758 [27:40<04:58,  3.01it/s]
                                                   

 84%|████████▍ | 4861/5758 [27:40<04:58,  3.01it/s]
 84%|████████▍ | 4862/5758 [27:40<04:51,  3.07it/s]
                                                   

 84%|████████▍ | 4862/5758 [27:40<04:51,  3.07it/s]
 84%|████████▍ | 4863/5758 [27:41<04:53,  3.05it/s]
                                                   

 84%|████████▍ | 4863/5758 [27:41<04:53,  3.05it/s]
 84%|████████▍ | 4864/5758 [27:41<05:06,  2.92it/s]
                                                   

 84%|████████▍ | 4864/5758 [27:41<05:06,  2.92it/s]
 84%|████████▍ | 4865/5758 [27:42<05:17,  2.81it/s]
                                                   

 84%|████████▍ | 4865/5758 [27:42<05:17,  2.81it/s]
 85%|████████▍ | 4866/5758 [27:42<05:00,  2.97it/s]
                                                   

 85%|████████▍ | 4866/5758 [27:42<05:00,  2.97it/s]
 85%|████████▍ | 4867/5758 [27:42<04:55,  3.02it/s]
                                                   

 85%|████████▍ | 4867/5758 [27:42<04:55,  3.02it/s]
 85%|████████▍ | 4868/5758 [27:42<04:52,  3.04it/s]
                                                   

 85%|████████▍ | 4868/5758 [27:42<04:52,  3.04it/s]
 85%|████████▍ | 4869/5758 [27:43<05:02,  2.93it/s]
                                                   

 85%|████████▍ | 4869/5758 [27:43<05:02,  2.93it/s]
 85%|████████▍ | 4870/5758 [27:43<04:50,  3.05it/s]
                                                   

 85%|████████▍ | 4870/5758 [27:43<04:50,  3.05it/s]
 85%|████████▍ | 4871/5758 [27:43<04:46,  3.09it/s]
                                                   

 85%|████████▍ | 4871/5758 [27:43<04:46,  3.09it/s]
 85%|████████▍ | 4872/5758 [27:44<04:44,  3.11it/s]
                                                   

 85%|████████▍ | 4872/5758 [27:44<04:44,  3.11it/s]
 85%|████████▍ | 4873/5758 [27:44<04:43,  3.12it/s]
                                                   

 85%|████████▍ | 4873/5758 [27:44<04:43,  3.12it/s]
 85%|████████▍ | 4874/5758 [27:44<04:43,  3.12it/s]
                                                   

 85%|████████▍ | 4874/5758 [27:44<04:43,  3.12it/s]
 85%|████████▍ | 4875/5758 [27:45<05:08,  2.87it/s]
                                                   

 85%|████████▍ | 4875/5758 [27:45<05:08,  2.87it/s]
 85%|████████▍ | 4876/5758 [27:45<05:05,  2.89it/s]
                                                   

 85%|████████▍ | 4876/5758 [27:45<05:05,  2.89it/s]
 85%|████████▍ | 4877/5758 [27:46<05:00,  2.93it/s]
                                                   

 85%|████████▍ | 4877/5758 [27:46<05:00,  2.93it/s]
 85%|████████▍ | 4878/5758 [27:46<04:49,  3.04it/s]
                                                   

 85%|████████▍ | 4878/5758 [27:46<04:49,  3.04it/s]
 85%|████████▍ | 4879/5758 [27:46<04:41,  3.12it/s]
                                                   

 85%|████████▍ | 4879/5758 [27:46<04:41,  3.12it/s]
 85%|████████▍ | 4880/5758 [27:46<04:39,  3.14it/s]
                                                   

 85%|████████▍ | 4880/5758 [27:46<04:39,  3.14it/s]
 85%|████████▍ | 4881/5758 [27:47<04:57,  2.95it/s]
                                                   

 85%|████████▍ | 4881/5758 [27:47<04:57,  2.95it/s]
 85%|████████▍ | 4882/5758 [27:47<04:48,  3.04it/s]
                                                   

 85%|████████▍ | 4882/5758 [27:47<04:48,  3.04it/s]
 85%|████████▍ | 4883/5758 [27:47<04:44,  3.08it/s]
                                                   

 85%|████████▍ | 4883/5758 [27:47<04:44,  3.08it/s]
 85%|████████▍ | 4884/5758 [27:48<04:35,  3.17it/s]
                                                   

 85%|████████▍ | 4884/5758 [27:48<04:35,  3.17it/s]
 85%|████████▍ | 4885/5758 [27:48<04:29,  3.24it/s]
                                                   

 85%|████████▍ | 4885/5758 [27:48<04:29,  3.24it/s]
 85%|████████▍ | 4886/5758 [27:48<04:34,  3.18it/s]
                                                   
{'loss': '0.6468', 'grad_norm': '16.25', 'learning_rate': '1.612e-06', 'num_tokens': '3.215e+06', 'mean_token_accuracy': '0.8439', 'epoch': '0.839'}
+{'loss': '1.802', 'grad_norm': '20', 'learning_rate': '1.61e-06', 'num_tokens': '3.215e+06', 'mean_token_accuracy': '0.6492', 'epoch': '0.8392'}
+{'loss': '1.13', 'grad_norm': '12.06', 'learning_rate': '1.608e-06', 'num_tokens': '3.216e+06', 'mean_token_accuracy': '0.6623', 'epoch': '0.8394'}
+{'loss': '1.05', 'grad_norm': '15.94', 'learning_rate': '1.606e-06', 'num_tokens': '3.216e+06', 'mean_token_accuracy': '0.7432', 'epoch': '0.8395'}
+{'loss': '1.333', 'grad_norm': '12.62', 'learning_rate': '1.605e-06', 'num_tokens': '3.217e+06', 'mean_token_accuracy': '0.6428', 'epoch': '0.8397'}
+{'loss': '1.697', 'grad_norm': '13.12', 'learning_rate': '1.603e-06', 'num_tokens': '3.217e+06', 'mean_token_accuracy': '0.6399', 'epoch': '0.8399'}
+{'loss': '0.6718', 'grad_norm': '7.875', 'learning_rate': '1.601e-06', 'num_tokens': '3.219e+06', 'mean_token_accuracy': '0.7963', 'epoch': '0.84'}
+{'loss': '1.238', 'grad_norm': '11.06', 'learning_rate': '1.6e-06', 'num_tokens': '3.22e+06', 'mean_token_accuracy': '0.6622', 'epoch': '0.8402'}
+{'loss': '0.8383', 'grad_norm': '29', 'learning_rate': '1.598e-06', 'num_tokens': '3.22e+06', 'mean_token_accuracy': '0.8264', 'epoch': '0.8404'}
+{'loss': '1.059', 'grad_norm': '16.25', 'learning_rate': '1.596e-06', 'num_tokens': '3.221e+06', 'mean_token_accuracy': '0.6899', 'epoch': '0.8406'}
+{'loss': '1.367', 'grad_norm': '12.31', 'learning_rate': '1.594e-06', 'num_tokens': '3.222e+06', 'mean_token_accuracy': '0.6472', 'epoch': '0.8407'}
+{'loss': '1.328', 'grad_norm': '17.25', 'learning_rate': '1.593e-06', 'num_tokens': '3.222e+06', 'mean_token_accuracy': '0.7198', 'epoch': '0.8409'}
+{'loss': '1.271', 'grad_norm': '16.5', 'learning_rate': '1.591e-06', 'num_tokens': '3.223e+06', 'mean_token_accuracy': '0.6613', 'epoch': '0.8411'}
+{'loss': '0.8743', 'grad_norm': '36.75', 'learning_rate': '1.589e-06', 'num_tokens': '3.223e+06', 'mean_token_accuracy': '0.8154', 'epoch': '0.8413'}
+{'loss': '2.085', 'grad_norm': '11.38', 'learning_rate': '1.587e-06', 'num_tokens': '3.223e+06', 'mean_token_accuracy': '0.5434', 'epoch': '0.8414'}
+{'loss': '1.169', 'grad_norm': '12', 'learning_rate': '1.586e-06', 'num_tokens': '3.224e+06', 'mean_token_accuracy': '0.678', 'epoch': '0.8416'}
+{'loss': '1.524', 'grad_norm': '34.75', 'learning_rate': '1.584e-06', 'num_tokens': '3.224e+06', 'mean_token_accuracy': '0.6692', 'epoch': '0.8418'}
+{'loss': '1.183', 'grad_norm': '37', 'learning_rate': '1.582e-06', 'num_tokens': '3.224e+06', 'mean_token_accuracy': '0.7255', 'epoch': '0.842'}
+{'loss': '1.199', 'grad_norm': '9.688', 'learning_rate': '1.58e-06', 'num_tokens': '3.225e+06', 'mean_token_accuracy': '0.7105', 'epoch': '0.8421'}
+{'loss': '1.346', 'grad_norm': '13.81', 'learning_rate': '1.579e-06', 'num_tokens': '3.226e+06', 'mean_token_accuracy': '0.6422', 'epoch': '0.8423'}
+{'loss': '2.545', 'grad_norm': '16', 'learning_rate': '1.577e-06', 'num_tokens': '3.227e+06', 'mean_token_accuracy': '0.5206', 'epoch': '0.8425'}
+{'loss': '1.117', 'grad_norm': '12.62', 'learning_rate': '1.575e-06', 'num_tokens': '3.227e+06', 'mean_token_accuracy': '0.7024', 'epoch': '0.8427'}
+{'loss': '1.23', 'grad_norm': '11.62', 'learning_rate': '1.573e-06', 'num_tokens': '3.228e+06', 'mean_token_accuracy': '0.6509', 'epoch': '0.8428'}
+{'loss': '1.15', 'grad_norm': '9.375', 'learning_rate': '1.572e-06', 'num_tokens': '3.229e+06', 'mean_token_accuracy': '0.7133', 'epoch': '0.843'}
+{'loss': '0.8189', 'grad_norm': '16.38', 'learning_rate': '1.57e-06', 'num_tokens': '3.229e+06', 'mean_token_accuracy': '0.7919', 'epoch': '0.8432'}
+{'loss': '1.198', 'grad_norm': '12.44', 'learning_rate': '1.568e-06', 'num_tokens': '3.23e+06', 'mean_token_accuracy': '0.6897', 'epoch': '0.8433'}
+{'loss': '1.696', 'grad_norm': '10.19', 'learning_rate': '1.567e-06', 'num_tokens': '3.231e+06', 'mean_token_accuracy': '0.6447', 'epoch': '0.8435'}
+{'loss': '1.187', 'grad_norm': '11.31', 'learning_rate': '1.565e-06', 'num_tokens': '3.231e+06', 'mean_token_accuracy': '0.7197', 'epoch': '0.8437'}
+{'loss': '1.856', 'grad_norm': '38.75', 'learning_rate': '1.563e-06', 'num_tokens': '3.231e+06', 'mean_token_accuracy': '0.5833', 'epoch': '0.8439'}
+{'loss': '1.261', 'grad_norm': '7.719', 'learning_rate': '1.561e-06', 'num_tokens': '3.233e+06', 'mean_token_accuracy': '0.6986', 'epoch': '0.844'}
+{'loss': '1.177', 'grad_norm': '9.5', 'learning_rate': '1.56e-06', 'num_tokens': '3.234e+06', 'mean_token_accuracy': '0.7181', 'epoch': '0.8442'}
+{'loss': '1.791', 'grad_norm': '11.44', 'learning_rate': '1.558e-06', 'num_tokens': '3.235e+06', 'mean_token_accuracy': '0.6552', 'epoch': '0.8444'}
+{'loss': '1.78', 'grad_norm': '14', 'learning_rate': '1.556e-06', 'num_tokens': '3.236e+06', 'mean_token_accuracy': '0.638', 'epoch': '0.8446'}
+{'loss': '0.7148', 'grad_norm': '18.38', 'learning_rate': '1.554e-06', 'num_tokens': '3.236e+06', 'mean_token_accuracy': '0.8123', 'epoch': '0.8447'}
+{'loss': '1.344', 'grad_norm': '19', 'learning_rate': '1.553e-06', 'num_tokens': '3.236e+06', 'mean_token_accuracy': '0.6958', 'epoch': '0.8449'}
+{'loss': '0.7514', 'grad_norm': '11.56', 'learning_rate': '1.551e-06', 'num_tokens': '3.237e+06', 'mean_token_accuracy': '0.7907', 'epoch': '0.8451'}
+{'loss': '1.38', 'grad_norm': '22.38', 'learning_rate': '1.549e-06', 'num_tokens': '3.237e+06', 'mean_token_accuracy': '0.6702', 'epoch': '0.8453'}
+{'loss': '0.6464', 'grad_norm': '39', 'learning_rate': '1.547e-06', 'num_tokens': '3.237e+06', 'mean_token_accuracy': '0.8791', 'epoch': '0.8454'}
+{'loss': '1.404', 'grad_norm': '10.56', 'learning_rate': '1.546e-06', 'num_tokens': '3.238e+06', 'mean_token_accuracy': '0.6418', 'epoch': '0.8456'}
+{'loss': '1.039', 'grad_norm': '13.19', 'learning_rate': '1.544e-06', 'num_tokens': '3.239e+06', 'mean_token_accuracy': '0.7096', 'epoch': '0.8458'}
+{'loss': '1.293', 'grad_norm': '14.44', 'learning_rate': '1.542e-06', 'num_tokens': '3.24e+06', 'mean_token_accuracy': '0.645', 'epoch': '0.846'}
+{'loss': '0.5987', 'grad_norm': '7.562', 'learning_rate': '1.54e-06', 'num_tokens': '3.241e+06', 'mean_token_accuracy': '0.8123', 'epoch': '0.8461'}
+{'loss': '2.008', 'grad_norm': '10.44', 'learning_rate': '1.539e-06', 'num_tokens': '3.242e+06', 'mean_token_accuracy': '0.5866', 'epoch': '0.8463'}
+{'loss': '1.049', 'grad_norm': '11.88', 'learning_rate': '1.537e-06', 'num_tokens': '3.243e+06', 'mean_token_accuracy': '0.689', 'epoch': '0.8465'}
+{'loss': '1.178', 'grad_norm': '12.75', 'learning_rate': '1.535e-06', 'num_tokens': '3.243e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.8466'}
+{'loss': '1.489', 'grad_norm': '13.31', 'learning_rate': '1.534e-06', 'num_tokens': '3.244e+06', 'mean_token_accuracy': '0.6109', 'epoch': '0.8468'}
+{'loss': '0.5961', 'grad_norm': '7', 'learning_rate': '1.532e-06', 'num_tokens': '3.246e+06', 'mean_token_accuracy': '0.8234', 'epoch': '0.847'}
+{'loss': '1.151', 'grad_norm': '14.75', 'learning_rate': '1.53e-06', 'num_tokens': '3.247e+06', 'mean_token_accuracy': '0.6682', 'epoch': '0.8472'}
+{'loss': '1.316', 'grad_norm': '13.06', 'learning_rate': '1.528e-06', 'num_tokens': '3.247e+06', 'mean_token_accuracy': '0.6534', 'epoch': '0.8473'}
+{'loss': '0.5558', 'grad_norm': '7.312', 'learning_rate': '1.527e-06', 'num_tokens': '3.249e+06', 'mean_token_accuracy': '0.831', 'epoch': '0.8475'}
+{'loss': '1.033', 'grad_norm': '13.75', 'learning_rate': '1.525e-06', 'num_tokens': '3.249e+06', 'mean_token_accuracy': '0.721', 'epoch': '0.8477'}
+{'loss': '1.286', 'grad_norm': '12.75', 'learning_rate': '1.523e-06', 'num_tokens': '3.25e+06', 'mean_token_accuracy': '0.6692', 'epoch': '0.8479'}
+{'loss': '2.001', 'grad_norm': '49', 'learning_rate': '1.521e-06', 'num_tokens': '3.25e+06', 'mean_token_accuracy': '0.5833', 'epoch': '0.848'}
+{'loss': '1.096', 'grad_norm': '13.94', 'learning_rate': '1.52e-06', 'num_tokens': '3.251e+06', 'mean_token_accuracy': '0.6816', 'epoch': '0.8482'}
+{'loss': '0.9944', 'grad_norm': '16', 'learning_rate': '1.518e-06', 'num_tokens': '3.251e+06', 'mean_token_accuracy': '0.703', 'epoch': '0.8484'}
+
 85%|████████▍ | 4886/5758 [27:48<04:34,  3.18it/s]
 85%|████████▍ | 4887/5758 [27:49<04:37,  3.14it/s]
                                                   

 85%|████████▍ | 4887/5758 [27:49<04:37,  3.14it/s]
 85%|████████▍ | 4888/5758 [27:49<04:32,  3.20it/s]
                                                   

 85%|████████▍ | 4888/5758 [27:49<04:32,  3.20it/s]
 85%|████████▍ | 4889/5758 [27:49<04:30,  3.21it/s]
                                                   

 85%|████████▍ | 4889/5758 [27:49<04:30,  3.21it/s]
 85%|████████▍ | 4890/5758 [27:50<04:37,  3.13it/s]
                                                   

 85%|████████▍ | 4890/5758 [27:50<04:37,  3.13it/s]
 85%|████████▍ | 4891/5758 [27:50<04:30,  3.21it/s]
                                                   

 85%|████████▍ | 4891/5758 [27:50<04:30,  3.21it/s]
 85%|████████▍ | 4892/5758 [27:50<04:28,  3.23it/s]
                                                   

 85%|████████▍ | 4892/5758 [27:50<04:28,  3.23it/s]
 85%|████████▍ | 4893/5758 [27:51<04:35,  3.14it/s]
                                                   

 85%|████████▍ | 4893/5758 [27:51<04:35,  3.14it/s]
 85%|████████▍ | 4894/5758 [27:51<04:47,  3.01it/s]
                                                   

 85%|████████▍ | 4894/5758 [27:51<04:47,  3.01it/s]
 85%|████████▌ | 4895/5758 [27:51<04:54,  2.93it/s]
                                                   

 85%|████████▌ | 4895/5758 [27:51<04:54,  2.93it/s]
 85%|████████▌ | 4896/5758 [27:52<04:44,  3.03it/s]
                                                   

 85%|████████▌ | 4896/5758 [27:52<04:44,  3.03it/s]
 85%|████████▌ | 4897/5758 [27:52<04:38,  3.09it/s]
                                                   

 85%|████████▌ | 4897/5758 [27:52<04:38,  3.09it/s]
 85%|████████▌ | 4898/5758 [27:52<04:38,  3.09it/s]
                                                   

 85%|████████▌ | 4898/5758 [27:52<04:38,  3.09it/s]
 85%|████████▌ | 4899/5758 [27:53<04:30,  3.18it/s]
                                                   

 85%|████████▌ | 4899/5758 [27:53<04:30,  3.18it/s]
 85%|████████▌ | 4900/5758 [27:53<04:33,  3.14it/s]
                                                   

 85%|████████▌ | 4900/5758 [27:53<04:33,  3.14it/s]
 85%|████████▌ | 4901/5758 [27:53<04:26,  3.22it/s]
                                                   

 85%|████████▌ | 4901/5758 [27:53<04:26,  3.22it/s]
 85%|████████▌ | 4902/5758 [27:54<05:14,  2.72it/s]
                                                   

 85%|████████▌ | 4902/5758 [27:54<05:14,  2.72it/s]
 85%|████████▌ | 4903/5758 [27:54<05:01,  2.83it/s]
                                                   

 85%|████████▌ | 4903/5758 [27:54<05:01,  2.83it/s]
 85%|████████▌ | 4904/5758 [27:54<04:51,  2.93it/s]
                                                   

 85%|████████▌ | 4904/5758 [27:54<04:51,  2.93it/s]
 85%|████████▌ | 4905/5758 [27:55<04:38,  3.06it/s]
                                                   

 85%|████████▌ | 4905/5758 [27:55<04:38,  3.06it/s]
 85%|████████▌ | 4906/5758 [27:55<04:40,  3.04it/s]
                                                   

 85%|████████▌ | 4906/5758 [27:55<04:40,  3.04it/s]
 85%|████████▌ | 4907/5758 [27:55<04:33,  3.11it/s]
                                                   

 85%|████████▌ | 4907/5758 [27:55<04:33,  3.11it/s]
 85%|████████▌ | 4908/5758 [27:55<04:26,  3.19it/s]
                                                   

 85%|████████▌ | 4908/5758 [27:55<04:26,  3.19it/s]
 85%|████████▌ | 4909/5758 [27:56<04:27,  3.18it/s]
                                                   

 85%|████████▌ | 4909/5758 [27:56<04:27,  3.18it/s]
 85%|████████▌ | 4910/5758 [27:56<04:26,  3.18it/s]
                                                   

 85%|████████▌ | 4910/5758 [27:56<04:26,  3.18it/s]
 85%|████████▌ | 4911/5758 [27:56<04:42,  3.00it/s]
                                                   

 85%|████████▌ | 4911/5758 [27:57<04:42,  3.00it/s]
 85%|████████▌ | 4912/5758 [27:57<05:00,  2.82it/s]
                                                   

 85%|████████▌ | 4912/5758 [27:57<05:00,  2.82it/s]
 85%|████████▌ | 4913/5758 [27:57<04:57,  2.84it/s]
                                                   

 85%|████████▌ | 4913/5758 [27:57<04:57,  2.84it/s]
 85%|████████▌ | 4914/5758 [27:58<05:06,  2.76it/s]
                                                   

 85%|████████▌ | 4914/5758 [27:58<05:06,  2.76it/s]
 85%|████████▌ | 4915/5758 [27:58<05:11,  2.71it/s]
                                                   

 85%|████████▌ | 4915/5758 [27:58<05:11,  2.71it/s]
 85%|████████▌ | 4916/5758 [27:58<04:53,  2.87it/s]
                                                   

 85%|████████▌ | 4916/5758 [27:58<04:53,  2.87it/s]
 85%|████████▌ | 4917/5758 [27:59<04:43,  2.96it/s]
                                                   

 85%|████████▌ | 4917/5758 [27:59<04:43,  2.96it/s]
 85%|████████▌ | 4918/5758 [27:59<04:44,  2.95it/s]
                                                   

 85%|████████▌ | 4918/5758 [27:59<04:44,  2.95it/s]
 85%|████████▌ | 4919/5758 [27:59<04:40,  3.00it/s]
                                                   

 85%|████████▌ | 4919/5758 [27:59<04:40,  3.00it/s]
 85%|████████▌ | 4920/5758 [28:00<04:28,  3.12it/s]
                                                   

 85%|████████▌ | 4920/5758 [28:00<04:28,  3.12it/s]
 85%|████████▌ | 4921/5758 [28:00<04:30,  3.10it/s]
                                                   

 85%|████████▌ | 4921/5758 [28:00<04:30,  3.10it/s]
 85%|████████▌ | 4922/5758 [28:00<04:37,  3.01it/s]
                                                   

 85%|████████▌ | 4922/5758 [28:00<04:37,  3.01it/s]
 85%|████████▌ | 4923/5758 [28:01<04:31,  3.07it/s]
                                                   

 85%|████████▌ | 4923/5758 [28:01<04:31,  3.07it/s]
 86%|████████▌ | 4924/5758 [28:01<04:42,  2.95it/s]
                                                   

 86%|████████▌ | 4924/5758 [28:01<04:42,  2.95it/s]
 86%|████████▌ | 4925/5758 [28:01<04:32,  3.06it/s]
                                                   

 86%|████████▌ | 4925/5758 [28:01<04:32,  3.06it/s]
 86%|████████▌ | 4926/5758 [28:02<04:25,  3.14it/s]
                                                   

 86%|████████▌ | 4926/5758 [28:02<04:25,  3.14it/s]
 86%|████████▌ | 4927/5758 [28:02<04:19,  3.20it/s]
                                                   

 86%|████████▌ | 4927/5758 [28:02<04:19,  3.20it/s]
 86%|████████▌ | 4928/5758 [28:02<04:19,  3.20it/s]
                                                   

 86%|████████▌ | 4928/5758 [28:02<04:19,  3.20it/s]
 86%|████████▌ | 4929/5758 [28:03<04:29,  3.08it/s]
                                                   

 86%|████████▌ | 4929/5758 [28:03<04:29,  3.08it/s]
 86%|████████▌ | 4930/5758 [28:03<05:07,  2.70it/s]
                                                   

 86%|████████▌ | 4930/5758 [28:03<05:07,  2.70it/s]
 86%|████████▌ | 4931/5758 [28:03<04:48,  2.87it/s]
                                                   

 86%|████████▌ | 4931/5758 [28:03<04:48,  2.87it/s]
 86%|████████▌ | 4932/5758 [28:04<04:59,  2.76it/s]
                                                   

 86%|████████▌ | 4932/5758 [28:04<04:59,  2.76it/s]
 86%|████████▌ | 4933/5758 [28:04<04:59,  2.75it/s]
                                                   

 86%|████████▌ | 4933/5758 [28:04<04:59,  2.75it/s]
 86%|████████▌ | 4934/5758 [28:04<04:55,  2.79it/s]
                                                   

 86%|████████▌ | 4934/5758 [28:04<04:55,  2.79it/s]
 86%|████████▌ | 4935/5758 [28:05<04:42,  2.92it/s]
                                                   

 86%|████████▌ | 4935/5758 [28:05<04:42,  2.92it/s]
 86%|████████▌ | 4936/5758 [28:05<04:31,  3.02it/s]
                                                   

 86%|████████▌ | 4936/5758 [28:05<04:31,  3.02it/s]
 86%|████████▌ | 4937/5758 [28:05<04:25,  3.09it/s]
                                                   

 86%|████████▌ | 4937/5758 [28:05<04:25,  3.09it/s]
 86%|████████▌ | 4938/5758 [28:06<04:39,  2.94it/s]
                                                   

 86%|████████▌ | 4938/5758 [28:06<04:39,  2.94it/s]
 86%|████████▌ | 4939/5758 [28:06<04:49,  2.82it/s]
                                                   

 86%|████████▌ | 4939/5758 [28:06<04:49,  2.82it/s]
 86%|████████▌ | 4940/5758 [28:06<04:39,  2.93it/s]
                                                   

 86%|████████▌ | 4940/5758 [28:06<04:39,  2.93it/s]
 86%|████████▌ | 4941/5758 [28:07<04:26,  3.07it/s]
                                                   
{'loss': '1.13', 'grad_norm': '11.44', 'learning_rate': '1.516e-06', 'num_tokens': '3.252e+06', 'mean_token_accuracy': '0.6829', 'epoch': '0.8486'}
+{'loss': '1.368', 'grad_norm': '12.19', 'learning_rate': '1.514e-06', 'num_tokens': '3.253e+06', 'mean_token_accuracy': '0.6311', 'epoch': '0.8487'}
+{'loss': '1.213', 'grad_norm': '12.62', 'learning_rate': '1.513e-06', 'num_tokens': '3.254e+06', 'mean_token_accuracy': '0.6962', 'epoch': '0.8489'}
+{'loss': '1.099', 'grad_norm': '13.38', 'learning_rate': '1.511e-06', 'num_tokens': '3.254e+06', 'mean_token_accuracy': '0.685', 'epoch': '0.8491'}
+{'loss': '0.5247', 'grad_norm': '9.5', 'learning_rate': '1.509e-06', 'num_tokens': '3.255e+06', 'mean_token_accuracy': '0.8356', 'epoch': '0.8493'}
+{'loss': '1.305', 'grad_norm': '12.81', 'learning_rate': '1.507e-06', 'num_tokens': '3.256e+06', 'mean_token_accuracy': '0.6628', 'epoch': '0.8494'}
+{'loss': '1.376', 'grad_norm': '14.25', 'learning_rate': '1.506e-06', 'num_tokens': '3.256e+06', 'mean_token_accuracy': '0.6276', 'epoch': '0.8496'}
+{'loss': '1.141', 'grad_norm': '11.81', 'learning_rate': '1.504e-06', 'num_tokens': '3.257e+06', 'mean_token_accuracy': '0.6735', 'epoch': '0.8498'}
+{'loss': '1.322', 'grad_norm': '11.12', 'learning_rate': '1.502e-06', 'num_tokens': '3.258e+06', 'mean_token_accuracy': '0.6377', 'epoch': '0.8499'}
+{'loss': '1.342', 'grad_norm': '36.5', 'learning_rate': '1.501e-06', 'num_tokens': '3.258e+06', 'mean_token_accuracy': '0.7387', 'epoch': '0.8501'}
+{'loss': '0.7016', 'grad_norm': '21.5', 'learning_rate': '1.499e-06', 'num_tokens': '3.259e+06', 'mean_token_accuracy': '0.8204', 'epoch': '0.8503'}
+{'loss': '1.294', 'grad_norm': '13.38', 'learning_rate': '1.497e-06', 'num_tokens': '3.259e+06', 'mean_token_accuracy': '0.6509', 'epoch': '0.8505'}
+{'loss': '1.083', 'grad_norm': '30.88', 'learning_rate': '1.495e-06', 'num_tokens': '3.259e+06', 'mean_token_accuracy': '0.7889', 'epoch': '0.8506'}
+{'loss': '1.75', 'grad_norm': '25.25', 'learning_rate': '1.494e-06', 'num_tokens': '3.26e+06', 'mean_token_accuracy': '0.63', 'epoch': '0.8508'}
+{'loss': '1.617', 'grad_norm': '38', 'learning_rate': '1.492e-06', 'num_tokens': '3.26e+06', 'mean_token_accuracy': '0.6957', 'epoch': '0.851'}
+{'loss': '1.183', 'grad_norm': '17.25', 'learning_rate': '1.49e-06', 'num_tokens': '3.26e+06', 'mean_token_accuracy': '0.7318', 'epoch': '0.8512'}
+{'loss': '1.075', 'grad_norm': '19', 'learning_rate': '1.488e-06', 'num_tokens': '3.26e+06', 'mean_token_accuracy': '0.7296', 'epoch': '0.8513'}
+{'loss': '0.8301', 'grad_norm': '22.5', 'learning_rate': '1.487e-06', 'num_tokens': '3.261e+06', 'mean_token_accuracy': '0.8108', 'epoch': '0.8515'}
+{'loss': '1.305', 'grad_norm': '12.62', 'learning_rate': '1.485e-06', 'num_tokens': '3.261e+06', 'mean_token_accuracy': '0.6453', 'epoch': '0.8517'}
+{'loss': '2.141', 'grad_norm': '17.38', 'learning_rate': '1.483e-06', 'num_tokens': '3.263e+06', 'mean_token_accuracy': '0.5823', 'epoch': '0.8519'}
+{'loss': '0.413', 'grad_norm': '12.31', 'learning_rate': '1.481e-06', 'num_tokens': '3.263e+06', 'mean_token_accuracy': '0.9054', 'epoch': '0.852'}
+{'loss': '1.729', 'grad_norm': '13.75', 'learning_rate': '1.48e-06', 'num_tokens': '3.264e+06', 'mean_token_accuracy': '0.6183', 'epoch': '0.8522'}
+{'loss': '1.254', 'grad_norm': '13', 'learning_rate': '1.478e-06', 'num_tokens': '3.264e+06', 'mean_token_accuracy': '0.6599', 'epoch': '0.8524'}
+{'loss': '0.512', 'grad_norm': '6.875', 'learning_rate': '1.476e-06', 'num_tokens': '3.266e+06', 'mean_token_accuracy': '0.8483', 'epoch': '0.8526'}
+{'loss': '0.5806', 'grad_norm': '7.562', 'learning_rate': '1.474e-06', 'num_tokens': '3.267e+06', 'mean_token_accuracy': '0.8161', 'epoch': '0.8527'}
+{'loss': '1.102', 'grad_norm': '19.75', 'learning_rate': '1.473e-06', 'num_tokens': '3.268e+06', 'mean_token_accuracy': '0.7628', 'epoch': '0.8529'}
+{'loss': '1.2', 'grad_norm': '27.25', 'learning_rate': '1.471e-06', 'num_tokens': '3.268e+06', 'mean_token_accuracy': '0.7569', 'epoch': '0.8531'}
+{'loss': '1.037', 'grad_norm': '20.62', 'learning_rate': '1.469e-06', 'num_tokens': '3.268e+06', 'mean_token_accuracy': '0.7689', 'epoch': '0.8532'}
+{'loss': '0.7298', 'grad_norm': '9.75', 'learning_rate': '1.468e-06', 'num_tokens': '3.269e+06', 'mean_token_accuracy': '0.7753', 'epoch': '0.8534'}
+{'loss': '1.387', 'grad_norm': '15.75', 'learning_rate': '1.466e-06', 'num_tokens': '3.27e+06', 'mean_token_accuracy': '0.6769', 'epoch': '0.8536'}
+{'loss': '2.108', 'grad_norm': '18', 'learning_rate': '1.464e-06', 'num_tokens': '3.27e+06', 'mean_token_accuracy': '0.5821', 'epoch': '0.8538'}
+{'loss': '0.6193', 'grad_norm': '7.812', 'learning_rate': '1.462e-06', 'num_tokens': '3.272e+06', 'mean_token_accuracy': '0.805', 'epoch': '0.8539'}
+{'loss': '1.04', 'grad_norm': '19.5', 'learning_rate': '1.461e-06', 'num_tokens': '3.272e+06', 'mean_token_accuracy': '0.7835', 'epoch': '0.8541'}
+{'loss': '1.582', 'grad_norm': '16.5', 'learning_rate': '1.459e-06', 'num_tokens': '3.272e+06', 'mean_token_accuracy': '0.6759', 'epoch': '0.8543'}
+{'loss': '0.71', 'grad_norm': '11.19', 'learning_rate': '1.457e-06', 'num_tokens': '3.273e+06', 'mean_token_accuracy': '0.8239', 'epoch': '0.8545'}
+{'loss': '0.6185', 'grad_norm': '7.25', 'learning_rate': '1.455e-06', 'num_tokens': '3.275e+06', 'mean_token_accuracy': '0.8114', 'epoch': '0.8546'}
+{'loss': '0.8543', 'grad_norm': '29.5', 'learning_rate': '1.454e-06', 'num_tokens': '3.275e+06', 'mean_token_accuracy': '0.8', 'epoch': '0.8548'}
+{'loss': '1.62', 'grad_norm': '17.12', 'learning_rate': '1.452e-06', 'num_tokens': '3.275e+06', 'mean_token_accuracy': '0.6629', 'epoch': '0.855'}
+{'loss': '1.629', 'grad_norm': '23.12', 'learning_rate': '1.45e-06', 'num_tokens': '3.275e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.8552'}
+{'loss': '1.259', 'grad_norm': '13.19', 'learning_rate': '1.448e-06', 'num_tokens': '3.276e+06', 'mean_token_accuracy': '0.6813', 'epoch': '0.8553'}
+{'loss': '1.813', 'grad_norm': '16', 'learning_rate': '1.447e-06', 'num_tokens': '3.277e+06', 'mean_token_accuracy': '0.6033', 'epoch': '0.8555'}
+{'loss': '1.676', 'grad_norm': '29', 'learning_rate': '1.445e-06', 'num_tokens': '3.277e+06', 'mean_token_accuracy': '0.6287', 'epoch': '0.8557'}
+{'loss': '0.6667', 'grad_norm': '7.656', 'learning_rate': '1.443e-06', 'num_tokens': '3.278e+06', 'mean_token_accuracy': '0.8043', 'epoch': '0.8559'}
+{'loss': '0.6986', 'grad_norm': '20.12', 'learning_rate': '1.441e-06', 'num_tokens': '3.279e+06', 'mean_token_accuracy': '0.8214', 'epoch': '0.856'}
+{'loss': '1.366', 'grad_norm': '11.25', 'learning_rate': '1.44e-06', 'num_tokens': '3.28e+06', 'mean_token_accuracy': '0.6215', 'epoch': '0.8562'}
+{'loss': '1.563', 'grad_norm': '19.25', 'learning_rate': '1.438e-06', 'num_tokens': '3.28e+06', 'mean_token_accuracy': '0.6468', 'epoch': '0.8564'}
+{'loss': '0.9069', 'grad_norm': '17.38', 'learning_rate': '1.436e-06', 'num_tokens': '3.281e+06', 'mean_token_accuracy': '0.8102', 'epoch': '0.8565'}
+{'loss': '1.259', 'grad_norm': '10.88', 'learning_rate': '1.435e-06', 'num_tokens': '3.282e+06', 'mean_token_accuracy': '0.6489', 'epoch': '0.8567'}
+{'loss': '1.494', 'grad_norm': '20.62', 'learning_rate': '1.433e-06', 'num_tokens': '3.282e+06', 'mean_token_accuracy': '0.6769', 'epoch': '0.8569'}
+{'loss': '1.308', 'grad_norm': '45', 'learning_rate': '1.431e-06', 'num_tokens': '3.282e+06', 'mean_token_accuracy': '0.68', 'epoch': '0.8571'}
+{'loss': '1.791', 'grad_norm': '7.875', 'learning_rate': '1.429e-06', 'num_tokens': '3.283e+06', 'mean_token_accuracy': '0.6147', 'epoch': '0.8572'}
+{'loss': '1.943', 'grad_norm': '18', 'learning_rate': '1.428e-06', 'num_tokens': '3.284e+06', 'mean_token_accuracy': '0.5856', 'epoch': '0.8574'}
+{'loss': '0.629', 'grad_norm': '9.062', 'learning_rate': '1.426e-06', 'num_tokens': '3.285e+06', 'mean_token_accuracy': '0.8096', 'epoch': '0.8576'}
+{'loss': '0.624', 'grad_norm': '8.312', 'learning_rate': '1.424e-06', 'num_tokens': '3.286e+06', 'mean_token_accuracy': '0.812', 'epoch': '0.8578'}
+{'loss': '1.077', 'grad_norm': '7.469', 'learning_rate': '1.422e-06', 'num_tokens': '3.287e+06', 'mean_token_accuracy': '0.7128', 'epoch': '0.8579'}
+
 86%|████████▌ | 4941/5758 [28:07<04:26,  3.07it/s]
 86%|████████▌ | 4942/5758 [28:07<04:35,  2.96it/s]
                                                   

 86%|████████▌ | 4942/5758 [28:07<04:35,  2.96it/s]
 86%|████████▌ | 4943/5758 [28:07<04:32,  2.99it/s]
                                                   

 86%|████████▌ | 4943/5758 [28:07<04:32,  2.99it/s]
 86%|████████▌ | 4944/5758 [28:08<04:40,  2.91it/s]
                                                   

 86%|████████▌ | 4944/5758 [28:08<04:40,  2.91it/s]
 86%|████████▌ | 4945/5758 [28:08<04:25,  3.06it/s]
                                                   

 86%|████████▌ | 4945/5758 [28:08<04:25,  3.06it/s]
 86%|████████▌ | 4946/5758 [28:08<04:18,  3.14it/s]
                                                   

 86%|████████▌ | 4946/5758 [28:08<04:18,  3.14it/s]
 86%|████████▌ | 4947/5758 [28:09<04:20,  3.11it/s]
                                                   

 86%|████████▌ | 4947/5758 [28:09<04:20,  3.11it/s]
 86%|████████▌ | 4948/5758 [28:09<04:32,  2.98it/s]
                                                   

 86%|████████▌ | 4948/5758 [28:09<04:32,  2.98it/s]
 86%|████████▌ | 4949/5758 [28:09<04:22,  3.08it/s]
                                                   

 86%|████████▌ | 4949/5758 [28:09<04:22,  3.08it/s]
 86%|████████▌ | 4950/5758 [28:10<04:26,  3.03it/s]
                                                   

 86%|████████▌ | 4950/5758 [28:10<04:26,  3.03it/s]
 86%|████████▌ | 4951/5758 [28:10<04:20,  3.10it/s]
                                                   

 86%|████████▌ | 4951/5758 [28:10<04:20,  3.10it/s]
 86%|████████▌ | 4952/5758 [28:10<04:27,  3.01it/s]
                                                   

 86%|████████▌ | 4952/5758 [28:10<04:27,  3.01it/s]
 86%|████████▌ | 4953/5758 [28:11<04:18,  3.11it/s]
                                                   

 86%|████████▌ | 4953/5758 [28:11<04:18,  3.11it/s]
 86%|████████▌ | 4954/5758 [28:11<04:33,  2.94it/s]
                                                   

 86%|████████▌ | 4954/5758 [28:11<04:33,  2.94it/s]
 86%|████████▌ | 4955/5758 [28:11<04:31,  2.96it/s]
                                                   

 86%|████████▌ | 4955/5758 [28:11<04:31,  2.96it/s]
 86%|████████▌ | 4956/5758 [28:12<04:24,  3.03it/s]
                                                   

 86%|████████▌ | 4956/5758 [28:12<04:24,  3.03it/s]
 86%|████████▌ | 4957/5758 [28:12<04:10,  3.20it/s]
                                                   

 86%|████████▌ | 4957/5758 [28:12<04:10,  3.20it/s]
 86%|████████▌ | 4958/5758 [28:12<04:53,  2.73it/s]
                                                   

 86%|████████▌ | 4958/5758 [28:12<04:53,  2.73it/s]
 86%|████████▌ | 4959/5758 [28:13<04:46,  2.79it/s]
                                                   

 86%|████████▌ | 4959/5758 [28:13<04:46,  2.79it/s]
 86%|████████▌ | 4960/5758 [28:13<04:43,  2.81it/s]
                                                   

 86%|████████▌ | 4960/5758 [28:13<04:43,  2.81it/s]
 86%|████████▌ | 4961/5758 [28:13<04:50,  2.74it/s]
                                                   

 86%|████████▌ | 4961/5758 [28:13<04:50,  2.74it/s]
 86%|████████▌ | 4962/5758 [28:14<04:53,  2.71it/s]
                                                   

 86%|████████▌ | 4962/5758 [28:14<04:53,  2.71it/s]
 86%|████████▌ | 4963/5758 [28:14<04:51,  2.72it/s]
                                                   

 86%|████████▌ | 4963/5758 [28:14<04:51,  2.72it/s]
 86%|████████▌ | 4964/5758 [28:15<04:32,  2.91it/s]
                                                   

 86%|████████▌ | 4964/5758 [28:15<04:32,  2.91it/s]
 86%|████████▌ | 4965/5758 [28:15<04:36,  2.87it/s]
                                                   

 86%|████████▌ | 4965/5758 [28:15<04:36,  2.87it/s]
 86%|████████▌ | 4966/5758 [28:15<04:39,  2.83it/s]
                                                   

 86%|████████▌ | 4966/5758 [28:15<04:39,  2.83it/s]
 86%|████████▋ | 4967/5758 [28:16<04:39,  2.83it/s]
                                                   

 86%|████████▋ | 4967/5758 [28:16<04:39,  2.83it/s]
 86%|████████▋ | 4968/5758 [28:16<04:28,  2.95it/s]
                                                   

 86%|████████▋ | 4968/5758 [28:16<04:28,  2.95it/s]
 86%|████████▋ | 4969/5758 [28:16<04:26,  2.96it/s]
                                                   

 86%|████████▋ | 4969/5758 [28:16<04:26,  2.96it/s]
 86%|████████▋ | 4970/5758 [28:17<04:27,  2.94it/s]
                                                   

 86%|████████▋ | 4970/5758 [28:17<04:27,  2.94it/s]
 86%|██��█████▋ | 4971/5758 [28:17<04:15,  3.08it/s]
                                                   

 86%|████████▋ | 4971/5758 [28:17<04:15,  3.08it/s]
 86%|████████▋ | 4972/5758 [28:17<04:29,  2.91it/s]
                                                   

 86%|████████▋ | 4972/5758 [28:17<04:29,  2.91it/s]
 86%|████████▋ | 4973/5758 [28:18<04:23,  2.98it/s]
                                                   

 86%|████████▋ | 4973/5758 [28:18<04:23,  2.98it/s]
 86%|████████▋ | 4974/5758 [28:18<04:30,  2.90it/s]
                                                   

 86%|████████▋ | 4974/5758 [28:18<04:30,  2.90it/s]
 86%|████████▋ | 4975/5758 [28:18<04:21,  2.99it/s]
                                                   

 86%|████████▋ | 4975/5758 [28:18<04:21,  2.99it/s]
 86%|████████▋ | 4976/5758 [28:19<04:31,  2.88it/s]
                                                   

 86%|████████▋ | 4976/5758 [28:19<04:31,  2.88it/s]
 86%|████████▋ | 4977/5758 [28:19<04:27,  2.91it/s]
                                                   

 86%|████████▋ | 4977/5758 [28:19<04:27,  2.91it/s]
 86%|████████▋ | 4978/5758 [28:19<04:25,  2.94it/s]
                                                   

 86%|████████▋ | 4978/5758 [28:19<04:25,  2.94it/s]
 86%|████████▋ | 4979/5758 [28:20<04:14,  3.07it/s]
                                                   

 86%|████████▋ | 4979/5758 [28:20<04:14,  3.07it/s]
 86%|████████▋ | 4980/5758 [28:20<04:16,  3.03it/s]
                                                   

 86%|████████▋ | 4980/5758 [28:20<04:16,  3.03it/s]
 87%|████████▋ | 4981/5758 [28:20<04:27,  2.90it/s]
                                                   

 87%|████████▋ | 4981/5758 [28:20<04:27,  2.90it/s]
 87%|████████▋ | 4982/5758 [28:21<04:22,  2.95it/s]
                                                   

 87%|████████▋ | 4982/5758 [28:21<04:22,  2.95it/s]
 87%|████████▋ | 4983/5758 [28:21<04:38,  2.78it/s]
                                                   

 87%|████████▋ | 4983/5758 [28:21<04:38,  2.78it/s]
 87%|████████▋ | 4984/5758 [28:21<04:25,  2.92it/s]
                                                   

 87%|████████▋ | 4984/5758 [28:21<04:25,  2.92it/s]
 87%|████████▋ | 4985/5758 [28:22<04:39,  2.77it/s]
                                                   

 87%|████████▋ | 4985/5758 [28:22<04:39,  2.77it/s]
 87%|████████▋ | 4986/5758 [28:22<04:44,  2.71it/s]
                                                   

 87%|████████▋ | 4986/5758 [28:22<04:44,  2.71it/s]
 87%|████████▋ | 4987/5758 [28:23<04:46,  2.69it/s]
                                                   

 87%|████████▋ | 4987/5758 [28:23<04:46,  2.69it/s]
 87%|████████▋ | 4988/5758 [28:23<04:29,  2.86it/s]
                                                   

 87%|████████▋ | 4988/5758 [28:23<04:29,  2.86it/s]
 87%|████████▋ | 4989/5758 [28:23<04:26,  2.89it/s]
                                                   

 87%|████████▋ | 4989/5758 [28:23<04:26,  2.89it/s]
 87%|████████▋ | 4990/5758 [28:23<04:17,  2.98it/s]
                                                   

 87%|████████▋ | 4990/5758 [28:23<04:17,  2.98it/s]
 87%|████████▋ | 4991/5758 [28:24<04:11,  3.04it/s]
                                                   

 87%|████████▋ | 4991/5758 [28:24<04:11,  3.04it/s]
 87%|████████▋ | 4992/5758 [28:24<04:13,  3.03it/s]
                                                   

 87%|████████▋ | 4992/5758 [28:24<04:13,  3.03it/s]
 87%|████████▋ | 4993/5758 [28:24<04:06,  3.10it/s]
                                                   

 87%|████████▋ | 4993/5758 [28:24<04:06,  3.10it/s]
 87%|████████▋ | 4994/5758 [28:25<04:01,  3.17it/s]
                                                   

 87%|████████▋ | 4994/5758 [28:25<04:01,  3.17it/s]
 87%|████████▋ | 4995/5758 [28:25<04:00,  3.17it/s]
                                                   

 87%|████████▋ | 4995/5758 [28:25<04:00,  3.17it/s]
 87%|████████▋ | 4996/5758 [28:25<04:08,  3.07it/s]
                                                   
{'loss': '1.128', 'grad_norm': '13.69', 'learning_rate': '1.421e-06', 'num_tokens': '3.288e+06', 'mean_token_accuracy': '0.6743', 'epoch': '0.8581'}
+{'loss': '0.9109', 'grad_norm': '13.31', 'learning_rate': '1.419e-06', 'num_tokens': '3.289e+06', 'mean_token_accuracy': '0.7981', 'epoch': '0.8583'}
+{'loss': '1.335', 'grad_norm': '11.81', 'learning_rate': '1.417e-06', 'num_tokens': '3.289e+06', 'mean_token_accuracy': '0.6363', 'epoch': '0.8585'}
+{'loss': '1.244', 'grad_norm': '11.06', 'learning_rate': '1.415e-06', 'num_tokens': '3.29e+06', 'mean_token_accuracy': '0.6563', 'epoch': '0.8586'}
+{'loss': '1.583', 'grad_norm': '14.94', 'learning_rate': '1.414e-06', 'num_tokens': '3.291e+06', 'mean_token_accuracy': '0.6539', 'epoch': '0.8588'}
+{'loss': '0.6756', 'grad_norm': '15.12', 'learning_rate': '1.412e-06', 'num_tokens': '3.292e+06', 'mean_token_accuracy': '0.8645', 'epoch': '0.859'}
+{'loss': '0.5001', 'grad_norm': '6.031', 'learning_rate': '1.41e-06', 'num_tokens': '3.293e+06', 'mean_token_accuracy': '0.8574', 'epoch': '0.8592'}
+{'loss': '0.6866', 'grad_norm': '9.625', 'learning_rate': '1.408e-06', 'num_tokens': '3.294e+06', 'mean_token_accuracy': '0.7794', 'epoch': '0.8593'}
+{'loss': '1.259', 'grad_norm': '14.88', 'learning_rate': '1.407e-06', 'num_tokens': '3.295e+06', 'mean_token_accuracy': '0.6334', 'epoch': '0.8595'}
+{'loss': '1.587', 'grad_norm': '23.88', 'learning_rate': '1.405e-06', 'num_tokens': '3.295e+06', 'mean_token_accuracy': '0.6568', 'epoch': '0.8597'}
+{'loss': '1.654', 'grad_norm': '12.19', 'learning_rate': '1.403e-06', 'num_tokens': '3.296e+06', 'mean_token_accuracy': '0.6471', 'epoch': '0.8598'}
+{'loss': '2.042', 'grad_norm': '19.75', 'learning_rate': '1.402e-06', 'num_tokens': '3.297e+06', 'mean_token_accuracy': '0.5651', 'epoch': '0.86'}
+{'loss': '1.248', 'grad_norm': '12.88', 'learning_rate': '1.4e-06', 'num_tokens': '3.297e+06', 'mean_token_accuracy': '0.6557', 'epoch': '0.8602'}
+{'loss': '0.7563', 'grad_norm': '23.62', 'learning_rate': '1.398e-06', 'num_tokens': '3.297e+06', 'mean_token_accuracy': '0.8163', 'epoch': '0.8604'}
+{'loss': '0.503', 'grad_norm': '6.656', 'learning_rate': '1.396e-06', 'num_tokens': '3.299e+06', 'mean_token_accuracy': '0.849', 'epoch': '0.8605'}
+{'loss': '1.146', 'grad_norm': '12.75', 'learning_rate': '1.395e-06', 'num_tokens': '3.3e+06', 'mean_token_accuracy': '0.668', 'epoch': '0.8607'}
+{'loss': '0.8636', 'grad_norm': '37.75', 'learning_rate': '1.393e-06', 'num_tokens': '3.3e+06', 'mean_token_accuracy': '0.8136', 'epoch': '0.8609'}
+{'loss': '1.438', 'grad_norm': '19.38', 'learning_rate': '1.391e-06', 'num_tokens': '3.3e+06', 'mean_token_accuracy': '0.7283', 'epoch': '0.8611'}
+{'loss': '1.19', 'grad_norm': '38', 'learning_rate': '1.389e-06', 'num_tokens': '3.3e+06', 'mean_token_accuracy': '0.7451', 'epoch': '0.8612'}
+{'loss': '1.763', 'grad_norm': '10.19', 'learning_rate': '1.388e-06', 'num_tokens': '3.301e+06', 'mean_token_accuracy': '0.6381', 'epoch': '0.8614'}
+{'loss': '1.514', 'grad_norm': '18.5', 'learning_rate': '1.386e-06', 'num_tokens': '3.301e+06', 'mean_token_accuracy': '0.6542', 'epoch': '0.8616'}
+{'loss': '0.7827', 'grad_norm': '12.75', 'learning_rate': '1.384e-06', 'num_tokens': '3.302e+06', 'mean_token_accuracy': '0.8141', 'epoch': '0.8618'}
+{'loss': '0.6583', 'grad_norm': '28.5', 'learning_rate': '1.382e-06', 'num_tokens': '3.302e+06', 'mean_token_accuracy': '0.8403', 'epoch': '0.8619'}
+{'loss': '0.5853', 'grad_norm': '7.531', 'learning_rate': '1.381e-06', 'num_tokens': '3.303e+06', 'mean_token_accuracy': '0.832', 'epoch': '0.8621'}
+{'loss': '2.299', 'grad_norm': '21', 'learning_rate': '1.379e-06', 'num_tokens': '3.304e+06', 'mean_token_accuracy': '0.5506', 'epoch': '0.8623'}
+{'loss': '1.097', 'grad_norm': '17.5', 'learning_rate': '1.377e-06', 'num_tokens': '3.304e+06', 'mean_token_accuracy': '0.7778', 'epoch': '0.8625'}
+{'loss': '1.357', 'grad_norm': '11.44', 'learning_rate': '1.375e-06', 'num_tokens': '3.305e+06', 'mean_token_accuracy': '0.6514', 'epoch': '0.8626'}
+{'loss': '1.277', 'grad_norm': '13.12', 'learning_rate': '1.374e-06', 'num_tokens': '3.306e+06', 'mean_token_accuracy': '0.6534', 'epoch': '0.8628'}
+{'loss': '2.478', 'grad_norm': '51.5', 'learning_rate': '1.372e-06', 'num_tokens': '3.306e+06', 'mean_token_accuracy': '0.551', 'epoch': '0.863'}
+{'loss': '1.72', 'grad_norm': '13.75', 'learning_rate': '1.37e-06', 'num_tokens': '3.306e+06', 'mean_token_accuracy': '0.6383', 'epoch': '0.8631'}
+{'loss': '0.8262', 'grad_norm': '38.25', 'learning_rate': '1.369e-06', 'num_tokens': '3.307e+06', 'mean_token_accuracy': '0.8133', 'epoch': '0.8633'}
+{'loss': '2.087', 'grad_norm': '17.25', 'learning_rate': '1.367e-06', 'num_tokens': '3.307e+06', 'mean_token_accuracy': '0.5826', 'epoch': '0.8635'}
+{'loss': '1.84', 'grad_norm': '17.5', 'learning_rate': '1.365e-06', 'num_tokens': '3.308e+06', 'mean_token_accuracy': '0.6208', 'epoch': '0.8637'}
+{'loss': '1.902', 'grad_norm': '23.5', 'learning_rate': '1.363e-06', 'num_tokens': '3.308e+06', 'mean_token_accuracy': '0.6022', 'epoch': '0.8638'}
+{'loss': '1.252', 'grad_norm': '12.12', 'learning_rate': '1.362e-06', 'num_tokens': '3.309e+06', 'mean_token_accuracy': '0.6662', 'epoch': '0.864'}
+{'loss': '0.8151', 'grad_norm': '29.5', 'learning_rate': '1.36e-06', 'num_tokens': '3.309e+06', 'mean_token_accuracy': '0.8387', 'epoch': '0.8642'}
+{'loss': '1.8', 'grad_norm': '15', 'learning_rate': '1.358e-06', 'num_tokens': '3.309e+06', 'mean_token_accuracy': '0.6393', 'epoch': '0.8644'}
+{'loss': '1.305', 'grad_norm': '10.88', 'learning_rate': '1.356e-06', 'num_tokens': '3.31e+06', 'mean_token_accuracy': '0.7009', 'epoch': '0.8645'}
+{'loss': '1.804', 'grad_norm': '16.88', 'learning_rate': '1.355e-06', 'num_tokens': '3.311e+06', 'mean_token_accuracy': '0.6086', 'epoch': '0.8647'}
+{'loss': '2.147', 'grad_norm': '21', 'learning_rate': '1.353e-06', 'num_tokens': '3.311e+06', 'mean_token_accuracy': '0.5824', 'epoch': '0.8649'}
+{'loss': '1.294', 'grad_norm': '14.19', 'learning_rate': '1.351e-06', 'num_tokens': '3.312e+06', 'mean_token_accuracy': '0.6524', 'epoch': '0.8651'}
+{'loss': '2.303', 'grad_norm': '12.81', 'learning_rate': '1.349e-06', 'num_tokens': '3.312e+06', 'mean_token_accuracy': '0.5251', 'epoch': '0.8652'}
+{'loss': '1.047', 'grad_norm': '12.44', 'learning_rate': '1.348e-06', 'num_tokens': '3.313e+06', 'mean_token_accuracy': '0.703', 'epoch': '0.8654'}
+{'loss': '1.034', 'grad_norm': '8.438', 'learning_rate': '1.346e-06', 'num_tokens': '3.314e+06', 'mean_token_accuracy': '0.7825', 'epoch': '0.8656'}
+{'loss': '0.7192', 'grad_norm': '24.88', 'learning_rate': '1.344e-06', 'num_tokens': '3.314e+06', 'mean_token_accuracy': '0.8', 'epoch': '0.8658'}
+{'loss': '1.43', 'grad_norm': '34', 'learning_rate': '1.342e-06', 'num_tokens': '3.315e+06', 'mean_token_accuracy': '0.6977', 'epoch': '0.8659'}
+{'loss': '1.216', 'grad_norm': '24', 'learning_rate': '1.341e-06', 'num_tokens': '3.315e+06', 'mean_token_accuracy': '0.7256', 'epoch': '0.8661'}
+{'loss': '1.936', 'grad_norm': '13.69', 'learning_rate': '1.339e-06', 'num_tokens': '3.315e+06', 'mean_token_accuracy': '0.5816', 'epoch': '0.8663'}
+{'loss': '1.123', 'grad_norm': '21.62', 'learning_rate': '1.337e-06', 'num_tokens': '3.316e+06', 'mean_token_accuracy': '0.7736', 'epoch': '0.8664'}
+{'loss': '1.085', 'grad_norm': '11.69', 'learning_rate': '1.336e-06', 'num_tokens': '3.316e+06', 'mean_token_accuracy': '0.7035', 'epoch': '0.8666'}
+{'loss': '0.4427', 'grad_norm': '6.062', 'learning_rate': '1.334e-06', 'num_tokens': '3.318e+06', 'mean_token_accuracy': '0.8659', 'epoch': '0.8668'}
+{'loss': '1.243', 'grad_norm': '24.88', 'learning_rate': '1.332e-06', 'num_tokens': '3.318e+06', 'mean_token_accuracy': '0.7581', 'epoch': '0.867'}
+{'loss': '1.083', 'grad_norm': '12.94', 'learning_rate': '1.33e-06', 'num_tokens': '3.319e+06', 'mean_token_accuracy': '0.6824', 'epoch': '0.8671'}
+{'loss': '1.66', 'grad_norm': '16.62', 'learning_rate': '1.329e-06', 'num_tokens': '3.319e+06', 'mean_token_accuracy': '0.6662', 'epoch': '0.8673'}
+{'loss': '1.297', 'grad_norm': '12.69', 'learning_rate': '1.327e-06', 'num_tokens': '3.32e+06', 'mean_token_accuracy': '0.6568', 'epoch': '0.8675'}
+
 87%|████████▋ | 4996/5758 [28:25<04:08,  3.07it/s]
 87%|████████▋ | 4997/5758 [28:26<03:58,  3.19it/s]
                                                   

 87%|████████▋ | 4997/5758 [28:26<03:58,  3.19it/s]
 87%|████████▋ | 4998/5758 [28:26<03:54,  3.23it/s]
                                                   

 87%|████████▋ | 4998/5758 [28:26<03:54,  3.23it/s]
 87%|████████▋ | 4999/5758 [28:26<04:01,  3.15it/s]
                                                   

 87%|████████▋ | 4999/5758 [28:26<04:01,  3.15it/s]
 87%|████████▋ | 5000/5758 [28:27<04:02,  3.12it/s]
                                                   

 87%|████████▋ | 5000/5758 [28:27<04:02,  3.12it/s]{'loss': '1.05', 'grad_norm': '23.88', 'learning_rate': '1.325e-06', 'num_tokens': '3.32e+06', 'mean_token_accuracy': '0.7724', 'epoch': '0.8677'}
+{'loss': '1.732', 'grad_norm': '15.31', 'learning_rate': '1.323e-06', 'num_tokens': '3.321e+06', 'mean_token_accuracy': '0.5812', 'epoch': '0.8678'}
+{'loss': '1.193', 'grad_norm': '13.88', 'learning_rate': '1.322e-06', 'num_tokens': '3.322e+06', 'mean_token_accuracy': '0.6735', 'epoch': '0.868'}
+{'loss': '0.736', 'grad_norm': '34.25', 'learning_rate': '1.32e-06', 'num_tokens': '3.322e+06', 'mean_token_accuracy': '0.8641', 'epoch': '0.8682'}
+{'loss': '1.467', 'grad_norm': '11.81', 'learning_rate': '1.318e-06', 'num_tokens': '3.323e+06', 'mean_token_accuracy': '0.6332', 'epoch': '0.8684'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.57it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.56it/s]
+
 87%|████████▋ | 5001/5758 [28:29<10:52,  1.16it/s]
                                                   

 87%|████████▋ | 5001/5758 [28:29<10:52,  1.16it/s]
 87%|████████▋ | 5002/5758 [28:29<08:45,  1.44it/s]
                                                   

 87%|████████▋ | 5002/5758 [28:29<08:45,  1.44it/s]
 87%|████████▋ | 5003/5758 [28:29<07:27,  1.69it/s]
                                                   

 87%|████████▋ | 5003/5758 [28:29<07:27,  1.69it/s]
 87%|████████▋ | 5004/5758 [28:30<06:29,  1.94it/s]
                                                   

 87%|████████▋ | 5004/5758 [28:30<06:29,  1.94it/s]
 87%|████████▋ | 5005/5758 [28:30<05:42,  2.20it/s]
                                                   

 87%|████████▋ | 5005/5758 [28:30<05:42,  2.20it/s]
 87%|████████▋ | 5006/5758 [28:30<05:05,  2.47it/s]
                                                   

 87%|████████▋ | 5006/5758 [28:30<05:05,  2.47it/s]
 87%|████████▋ | 5007/5758 [28:31<04:58,  2.52it/s]
                                                   

 87%|████████▋ | 5007/5758 [28:31<04:58,  2.52it/s]
 87%|████████▋ | 5008/5758 [28:31<05:06,  2.45it/s]
                                                   

 87%|████████▋ | 5008/5758 [28:31<05:06,  2.45it/s]
 87%|████████▋ | 5009/5758 [28:31<04:42,  2.66it/s]
                                                   

 87%|████████▋ | 5009/5758 [28:31<04:42,  2.66it/s]
 87%|████████▋ | 5010/5758 [28:32<04:44,  2.63it/s]
                                                   

 87%|████████▋ | 5010/5758 [28:32<04:44,  2.63it/s]
 87%|████████▋ | 5011/5758 [28:32<04:23,  2.84it/s]
                                                   

 87%|████████▋ | 5011/5758 [28:32<04:23,  2.84it/s]
 87%|████████▋ | 5012/5758 [28:32<04:15,  2.92it/s]
                                                   

 87%|████████▋ | 5012/5758 [28:32<04:15,  2.92it/s]
 87%|████████▋ | 5013/5758 [28:33<04:11,  2.97it/s]
                                                   

 87%|████████▋ | 5013/5758 [28:33<04:11,  2.97it/s]
 87%|████████▋ | 5014/5758 [28:33<04:00,  3.09it/s]
                                                   

 87%|████████▋ | 5014/5758 [28:33<04:00,  3.09it/s]
 87%|████████▋ | 5015/5758 [28:33<04:13,  2.93it/s]
                                                   

 87%|████████▋ | 5015/5758 [28:33<04:13,  2.93it/s]
 87%|████████▋ | 5016/5758 [28:34<04:02,  3.06it/s]
                                                   

 87%|████████▋ | 5016/5758 [28:34<04:02,  3.06it/s]
 87%|████████▋ | 5017/5758 [28:34<04:17,  2.88it/s]
                                                   

 87%|████████▋ | 5017/5758 [28:34<04:17,  2.88it/s]
 87%|████████▋ | 5018/5758 [28:34<04:14,  2.91it/s]
                                                   

 87%|████████▋ | 5018/5758 [28:34<04:14,  2.91it/s]
 87%|████████▋ | 5019/5758 [28:35<04:07,  2.99it/s]
                                                   

 87%|████████▋ | 5019/5758 [28:35<04:07,  2.99it/s]
 87%|████████▋ | 5020/5758 [28:35<04:19,  2.85it/s]
                                                   

 87%|████████▋ | 5020/5758 [28:35<04:19,  2.85it/s]
 87%|████████▋ | 5021/5758 [28:36<04:13,  2.90it/s]
                                                   

 87%|████████▋ | 5021/5758 [28:36<04:13,  2.90it/s]
 87%|████████▋ | 5022/5758 [28:36<04:03,  3.02it/s]
                                                   

 87%|████████▋ | 5022/5758 [28:36<04:03,  3.02it/s]
 87%|████████▋ | 5023/5758 [28:36<04:08,  2.96it/s]
                                                   

 87%|████████▋ | 5023/5758 [28:36<04:08,  2.96it/s]
 87%|████████▋ | 5024/5758 [28:36<04:00,  3.05it/s]
                                                   

 87%|████████▋ | 5024/5758 [28:36<04:00,  3.05it/s]
 87%|████████▋ | 5025/5758 [28:37<03:56,  3.09it/s]
                                                   

 87%|████████▋ | 5025/5758 [28:37<03:56,  3.09it/s]
 87%|████████▋ | 5026/5758 [28:37<04:04,  2.99it/s]
                                                   

 87%|████████▋ | 5026/5758 [28:37<04:04,  2.99it/s]
 87%|████████▋ | 5027/5758 [28:37<04:06,  2.96it/s]
                                                   

 87%|████████▋ | 5027/5758 [28:37<04:06,  2.96it/s]
 87%|████████▋ | 5028/5758 [28:38<04:11,  2.91it/s]
                                                   

 87%|████████▋ | 5028/5758 [28:38<04:11,  2.91it/s]
 87%|████████▋ | 5029/5758 [28:38<04:03,  2.99it/s]
                                                   

 87%|████████▋ | 5029/5758 [28:38<04:03,  2.99it/s]
 87%|████████▋ | 5030/5758 [28:38<03:55,  3.09it/s]
                                                   

 87%|████████▋ | 5030/5758 [28:38<03:55,  3.09it/s]
 87%|████████▋ | 5031/5758 [28:39<03:51,  3.14it/s]
                                                   

 87%|████████▋ | 5031/5758 [28:39<03:51,  3.14it/s]
 87%|████████▋ | 5032/5758 [28:39<03:45,  3.21it/s]
                                                   

 87%|████████▋ | 5032/5758 [28:39<03:45,  3.21it/s]
 87%|████████▋ | 5033/5758 [28:39<03:40,  3.29it/s]
                                                   

 87%|████████▋ | 5033/5758 [28:39<03:40,  3.29it/s]
 87%|████████▋ | 5034/5758 [28:40<03:37,  3.34it/s]
                                                   

 87%|████████▋ | 5034/5758 [28:40<03:37,  3.34it/s]
 87%|████████▋ | 5035/5758 [28:40<03:39,  3.30it/s]
                                                   

 87%|████████▋ | 5035/5758 [28:40<03:39,  3.30it/s]
 87%|████████▋ | 5036/5758 [28:40<03:41,  3.26it/s]
                                                   

 87%|████████▋ | 5036/5758 [28:40<03:41,  3.26it/s]
 87%|████████▋ | 5037/5758 [28:41<04:08,  2.90it/s]
                                                   

 87%|████████▋ | 5037/5758 [28:41<04:08,  2.90it/s]
 87%|████████▋ | 5038/5758 [28:41<04:11,  2.86it/s]
                                                   

 87%|████████▋ | 5038/5758 [28:41<04:11,  2.86it/s]
 88%|████████▊ | 5039/5758 [28:41<04:06,  2.92it/s]
                                                   

 88%|████████▊ | 5039/5758 [28:41<04:06,  2.92it/s]
 88%|████████▊ | 5040/5758 [28:42<04:12,  2.84it/s]
                                                   

 88%|████████▊ | 5040/5758 [28:42<04:12,  2.84it/s]
 88%|████████▊ | 5041/5758 [28:42<04:01,  2.97it/s]
                                                   

 88%|████████▊ | 5041/5758 [28:42<04:01,  2.97it/s]
 88%|████████▊ | 5042/5758 [28:42<04:00,  2.98it/s]
                                                   

 88%|████████▊ | 5042/5758 [28:42<04:00,  2.98it/s]
 88%|████████▊ | 5043/5758 [28:43<03:50,  3.11it/s]
                                                   

 88%|████████▊ | 5043/5758 [28:43<03:50,  3.11it/s]
 88%|████████▊ | 5044/5758 [28:43<03:42,  3.21it/s]
                                                   

 88%|████████▊ | 5044/5758 [28:43<03:42,  3.21it/s]
 88%|████████▊ | 5045/5758 [28:43<03:56,  3.02it/s]
                                                   

 88%|████████▊ | 5045/5758 [28:43<03:56,  3.02it/s]
 88%|████████▊ | 5046/5758 [28:44<03:49,  3.10it/s]
                                                   

 88%|████████▊ | 5046/5758 [28:44<03:49,  3.10it/s]
 88%|████████▊ | 5047/5758 [28:44<03:49,  3.10it/s]
                                                   

 88%|████████▊ | 5047/5758 [28:44<03:49,  3.10it/s]
 88%|████████▊ | 5048/5758 [28:44<03:42,  3.19it/s]
                                                   

 88%|████████▊ | 5048/5758 [28:44<03:42,  3.19it/s]
 88%|████████▊ | 5049/5758 [28:45<03:54,  3.03it/s]
                                                   

 88%|████████▊ | 5049/5758 [28:45<03:54,  3.03it/s]
 88%|████████▊ | 5050/5758 [28:45<03:45,  3.14it/s]
                                                   

 88%|████████▊ | 5050/5758 [28:45<03:45,  3.14it/s]
 88%|████████▊ | 5051/5758 [28:45<03:54,  3.02it/s]
                                                   

 88%|████████▊ | 5051/5758 [28:45<03:54,  3.02it/s]
 88%|████████▊ | 5052/5758 [28:46<04:01,  2.93it/s]
                                                   

 88%|████████▊ | 5052/5758 [28:46<04:01,  2.93it/s]
 88%|████████▊ | 5053/5758 [28:46<04:02,  2.90it/s]
                                                   

 88%|████████▊ | 5053/5758 [28:46<04:02,  2.90it/s]
 88%|████████▊ | 5054/5758 [28:46<04:04,  2.87it/s]
                                                   

 88%|████████▊ | 5054/5758 [28:46<04:04,  2.87it/s]
 88%|████████▊ | 5055/5758 [28:47<03:57,  2.96it/s]
                                                   

 88%|████████▊ | 5055/5758 [28:47<03:57,  2.96it/s]
 88%|████████▊ | 5056/5758 [28:47<03:52,  3.02it/s]
                                                   
{'loss': '2.052', 'grad_norm': '16.25', 'learning_rate': '1.316e-06', 'num_tokens': '3.323e+06', 'mean_token_accuracy': '0.6026', 'epoch': '0.8685'}
+{'loss': '0.8852', 'grad_norm': '21', 'learning_rate': '1.315e-06', 'num_tokens': '3.324e+06', 'mean_token_accuracy': '0.8421', 'epoch': '0.8687'}
+{'loss': '1.506', 'grad_norm': '33.25', 'learning_rate': '1.313e-06', 'num_tokens': '3.324e+06', 'mean_token_accuracy': '0.6637', 'epoch': '0.8689'}
+{'loss': '1.083', 'grad_norm': '10.06', 'learning_rate': '1.311e-06', 'num_tokens': '3.325e+06', 'mean_token_accuracy': '0.7401', 'epoch': '0.8691'}
+{'loss': '1.468', 'grad_norm': '13.81', 'learning_rate': '1.309e-06', 'num_tokens': '3.325e+06', 'mean_token_accuracy': '0.6054', 'epoch': '0.8692'}
+{'loss': '0.9448', 'grad_norm': '33.25', 'learning_rate': '1.308e-06', 'num_tokens': '3.326e+06', 'mean_token_accuracy': '0.7707', 'epoch': '0.8694'}
+{'loss': '1.333', 'grad_norm': '9.312', 'learning_rate': '1.306e-06', 'num_tokens': '3.327e+06', 'mean_token_accuracy': '0.7024', 'epoch': '0.8696'}
+{'loss': '1.234', 'grad_norm': '12.25', 'learning_rate': '1.304e-06', 'num_tokens': '3.328e+06', 'mean_token_accuracy': '0.7034', 'epoch': '0.8697'}
+{'loss': '0.8515', 'grad_norm': '12.88', 'learning_rate': '1.303e-06', 'num_tokens': '3.328e+06', 'mean_token_accuracy': '0.7837', 'epoch': '0.8699'}
+{'loss': '1.156', 'grad_norm': '12.88', 'learning_rate': '1.301e-06', 'num_tokens': '3.329e+06', 'mean_token_accuracy': '0.6973', 'epoch': '0.8701'}
+{'loss': '1.132', 'grad_norm': '10', 'learning_rate': '1.299e-06', 'num_tokens': '3.33e+06', 'mean_token_accuracy': '0.6716', 'epoch': '0.8703'}
+{'loss': '1.142', 'grad_norm': '11.19', 'learning_rate': '1.297e-06', 'num_tokens': '3.331e+06', 'mean_token_accuracy': '0.683', 'epoch': '0.8704'}
+{'loss': '1.061', 'grad_norm': '11.06', 'learning_rate': '1.296e-06', 'num_tokens': '3.332e+06', 'mean_token_accuracy': '0.6892', 'epoch': '0.8706'}
+{'loss': '1.716', 'grad_norm': '15.5', 'learning_rate': '1.294e-06', 'num_tokens': '3.332e+06', 'mean_token_accuracy': '0.6352', 'epoch': '0.8708'}
+{'loss': '1.524', 'grad_norm': '16.38', 'learning_rate': '1.292e-06', 'num_tokens': '3.333e+06', 'mean_token_accuracy': '0.6823', 'epoch': '0.871'}
+{'loss': '1.869', 'grad_norm': '6.344', 'learning_rate': '1.29e-06', 'num_tokens': '3.335e+06', 'mean_token_accuracy': '0.6004', 'epoch': '0.8711'}
+{'loss': '1.647', 'grad_norm': '19.5', 'learning_rate': '1.289e-06', 'num_tokens': '3.336e+06', 'mean_token_accuracy': '0.653', 'epoch': '0.8713'}
+{'loss': '0.9847', 'grad_norm': '23.62', 'learning_rate': '1.287e-06', 'num_tokens': '3.336e+06', 'mean_token_accuracy': '0.8138', 'epoch': '0.8715'}
+{'loss': '1.851', 'grad_norm': '16.38', 'learning_rate': '1.285e-06', 'num_tokens': '3.336e+06', 'mean_token_accuracy': '0.6053', 'epoch': '0.8717'}
+{'loss': '0.8238', 'grad_norm': '12.69', 'learning_rate': '1.283e-06', 'num_tokens': '3.337e+06', 'mean_token_accuracy': '0.7956', 'epoch': '0.8718'}
+{'loss': '1.335', 'grad_norm': '17.25', 'learning_rate': '1.282e-06', 'num_tokens': '3.337e+06', 'mean_token_accuracy': '0.7218', 'epoch': '0.872'}
+{'loss': '2.107', 'grad_norm': '10.56', 'learning_rate': '1.28e-06', 'num_tokens': '3.338e+06', 'mean_token_accuracy': '0.5853', 'epoch': '0.8722'}
+{'loss': '1.498', 'grad_norm': '12.31', 'learning_rate': '1.278e-06', 'num_tokens': '3.339e+06', 'mean_token_accuracy': '0.5936', 'epoch': '0.8724'}
+{'loss': '1.096', 'grad_norm': '13', 'learning_rate': '1.276e-06', 'num_tokens': '3.34e+06', 'mean_token_accuracy': '0.6873', 'epoch': '0.8725'}
+{'loss': '1.417', 'grad_norm': '18.62', 'learning_rate': '1.275e-06', 'num_tokens': '3.34e+06', 'mean_token_accuracy': '0.7186', 'epoch': '0.8727'}
+{'loss': '1.585', 'grad_norm': '14.25', 'learning_rate': '1.273e-06', 'num_tokens': '3.341e+06', 'mean_token_accuracy': '0.6445', 'epoch': '0.8729'}
+{'loss': '1.935', 'grad_norm': '11.19', 'learning_rate': '1.271e-06', 'num_tokens': '3.342e+06', 'mean_token_accuracy': '0.5962', 'epoch': '0.873'}
+{'loss': '1.036', 'grad_norm': '9.062', 'learning_rate': '1.27e-06', 'num_tokens': '3.343e+06', 'mean_token_accuracy': '0.7574', 'epoch': '0.8732'}
+{'loss': '1.847', 'grad_norm': '11.69', 'learning_rate': '1.268e-06', 'num_tokens': '3.344e+06', 'mean_token_accuracy': '0.6096', 'epoch': '0.8734'}
+{'loss': '1.113', 'grad_norm': '12.88', 'learning_rate': '1.266e-06', 'num_tokens': '3.344e+06', 'mean_token_accuracy': '0.6891', 'epoch': '0.8736'}
+{'loss': '2.051', 'grad_norm': '9.438', 'learning_rate': '1.264e-06', 'num_tokens': '3.346e+06', 'mean_token_accuracy': '0.5962', 'epoch': '0.8737'}
+{'loss': '1.913', 'grad_norm': '18.62', 'learning_rate': '1.263e-06', 'num_tokens': '3.346e+06', 'mean_token_accuracy': '0.6036', 'epoch': '0.8739'}
+{'loss': '1.523', 'grad_norm': '51.5', 'learning_rate': '1.261e-06', 'num_tokens': '3.346e+06', 'mean_token_accuracy': '0.7857', 'epoch': '0.8741'}
+{'loss': '1.923', 'grad_norm': '13.81', 'learning_rate': '1.259e-06', 'num_tokens': '3.347e+06', 'mean_token_accuracy': '0.6034', 'epoch': '0.8743'}
+{'loss': '0.7254', 'grad_norm': '19.12', 'learning_rate': '1.257e-06', 'num_tokens': '3.347e+06', 'mean_token_accuracy': '0.8058', 'epoch': '0.8744'}
+{'loss': '2.179', 'grad_norm': '13', 'learning_rate': '1.256e-06', 'num_tokens': '3.348e+06', 'mean_token_accuracy': '0.5352', 'epoch': '0.8746'}
+{'loss': '1.715', 'grad_norm': '32.75', 'learning_rate': '1.254e-06', 'num_tokens': '3.348e+06', 'mean_token_accuracy': '0.5924', 'epoch': '0.8748'}
+{'loss': '0.6482', 'grad_norm': '9.125', 'learning_rate': '1.252e-06', 'num_tokens': '3.349e+06', 'mean_token_accuracy': '0.8094', 'epoch': '0.875'}
+{'loss': '0.8226', 'grad_norm': '22.12', 'learning_rate': '1.25e-06', 'num_tokens': '3.35e+06', 'mean_token_accuracy': '0.774', 'epoch': '0.8751'}
+{'loss': '1.462', 'grad_norm': '19.5', 'learning_rate': '1.249e-06', 'num_tokens': '3.35e+06', 'mean_token_accuracy': '0.6702', 'epoch': '0.8753'}
+{'loss': '1.146', 'grad_norm': '12.88', 'learning_rate': '1.247e-06', 'num_tokens': '3.351e+06', 'mean_token_accuracy': '0.6562', 'epoch': '0.8755'}
+{'loss': '1.798', 'grad_norm': '7.312', 'learning_rate': '1.245e-06', 'num_tokens': '3.353e+06', 'mean_token_accuracy': '0.6231', 'epoch': '0.8757'}
+{'loss': '1.904', 'grad_norm': '19.5', 'learning_rate': '1.243e-06', 'num_tokens': '3.353e+06', 'mean_token_accuracy': '0.5803', 'epoch': '0.8758'}
+{'loss': '1.226', 'grad_norm': '15.5', 'learning_rate': '1.242e-06', 'num_tokens': '3.354e+06', 'mean_token_accuracy': '0.6661', 'epoch': '0.876'}
+{'loss': '1.646', 'grad_norm': '33', 'learning_rate': '1.24e-06', 'num_tokens': '3.354e+06', 'mean_token_accuracy': '0.6094', 'epoch': '0.8762'}
+{'loss': '2.048', 'grad_norm': '21.25', 'learning_rate': '1.238e-06', 'num_tokens': '3.354e+06', 'mean_token_accuracy': '0.5859', 'epoch': '0.8763'}
+{'loss': '0.4788', 'grad_norm': '6.406', 'learning_rate': '1.237e-06', 'num_tokens': '3.356e+06', 'mean_token_accuracy': '0.8555', 'epoch': '0.8765'}
+{'loss': '1.24', 'grad_norm': '13.88', 'learning_rate': '1.235e-06', 'num_tokens': '3.356e+06', 'mean_token_accuracy': '0.6579', 'epoch': '0.8767'}
+{'loss': '1.092', 'grad_norm': '18.5', 'learning_rate': '1.233e-06', 'num_tokens': '3.357e+06', 'mean_token_accuracy': '0.7311', 'epoch': '0.8769'}
+{'loss': '0.5205', 'grad_norm': '17.25', 'learning_rate': '1.231e-06', 'num_tokens': '3.357e+06', 'mean_token_accuracy': '0.8758', 'epoch': '0.877'}
+{'loss': '1.969', 'grad_norm': '16.12', 'learning_rate': '1.23e-06', 'num_tokens': '3.357e+06', 'mean_token_accuracy': '0.5893', 'epoch': '0.8772'}
+{'loss': '2.262', 'grad_norm': '9.688', 'learning_rate': '1.228e-06', 'num_tokens': '3.358e+06', 'mean_token_accuracy': '0.5531', 'epoch': '0.8774'}
+{'loss': '1.601', 'grad_norm': '14.75', 'learning_rate': '1.226e-06', 'num_tokens': '3.359e+06', 'mean_token_accuracy': '0.6481', 'epoch': '0.8776'}
+{'loss': '1.532', 'grad_norm': '19.5', 'learning_rate': '1.224e-06', 'num_tokens': '3.359e+06', 'mean_token_accuracy': '0.6639', 'epoch': '0.8777'}
+{'loss': '1.085', 'grad_norm': '12', 'learning_rate': '1.223e-06', 'num_tokens': '3.36e+06', 'mean_token_accuracy': '0.7115', 'epoch': '0.8779'}
+
 88%|████████▊ | 5056/5758 [28:47<03:52,  3.02it/s]
 88%|████████▊ | 5057/5758 [28:47<03:45,  3.11it/s]
                                                   

 88%|████████▊ | 5057/5758 [28:47<03:45,  3.11it/s]
 88%|████████▊ | 5058/5758 [28:48<03:56,  2.96it/s]
                                                   

 88%|████████▊ | 5058/5758 [28:48<03:56,  2.96it/s]
 88%|████████▊ | 5059/5758 [28:48<04:01,  2.90it/s]
                                                   

 88%|████████▊ | 5059/5758 [28:48<04:01,  2.90it/s]
 88%|████████▊ | 5060/5758 [28:48<03:51,  3.01it/s]
                                                   

 88%|████████▊ | 5060/5758 [28:48<03:51,  3.01it/s]
 88%|████████▊ | 5061/5758 [28:49<03:46,  3.07it/s]
                                                   

 88%|████████▊ | 5061/5758 [28:49<03:46,  3.07it/s]
 88%|████████▊ | 5062/5758 [28:49<04:15,  2.72it/s]
                                                   

 88%|████████▊ | 5062/5758 [28:49<04:15,  2.72it/s]
 88%|████████▊ | 5063/5758 [28:49<04:02,  2.87it/s]
                                                   

 88%|████████▊ | 5063/5758 [28:49<04:02,  2.87it/s]
 88%|████████▊ | 5064/5758 [28:50<04:05,  2.83it/s]
                                                   

 88%|████████▊ | 5064/5758 [28:50<04:05,  2.83it/s]
 88%|████████▊ | 5065/5758 [28:50<04:03,  2.85it/s]
                                                   

 88%|████████▊ | 5065/5758 [28:50<04:03,  2.85it/s]
 88%|████████▊ | 5066/5758 [28:50<03:59,  2.89it/s]
                                                   

 88%|████████▊ | 5066/5758 [28:50<03:59,  2.89it/s]
 88%|████████▊ | 5067/5758 [28:51<04:10,  2.76it/s]
                                                   

 88%|████████▊ | 5067/5758 [28:51<04:10,  2.76it/s]
 88%|████████▊ | 5068/5758 [28:51<03:54,  2.94it/s]
                                                   

 88%|████████▊ | 5068/5758 [28:51<03:54,  2.94it/s]
 88%|████████▊ | 5069/5758 [28:51<03:47,  3.02it/s]
                                                   

 88%|████████▊ | 5069/5758 [28:51<03:47,  3.02it/s]
 88%|████████▊ | 5070/5758 [28:52<03:45,  3.05it/s]
                                                   

 88%|████████▊ | 5070/5758 [28:52<03:45,  3.05it/s]
 88%|████████▊ | 5071/5758 [28:52<03:39,  3.12it/s]
                                                   

 88%|████████▊ | 5071/5758 [28:52<03:39,  3.12it/s]
 88%|████████▊ | 5072/5758 [28:52<03:40,  3.11it/s]
                                                   

 88%|████████▊ | 5072/5758 [28:52<03:40,  3.11it/s]
 88%|████████▊ | 5073/5758 [28:53<03:35,  3.18it/s]
                                                   

 88%|████████▊ | 5073/5758 [28:53<03:35,  3.18it/s]
 88%|████████▊ | 5074/5758 [28:53<03:30,  3.25it/s]
                                                   

 88%|████████▊ | 5074/5758 [28:53<03:30,  3.25it/s]
 88%|████████▊ | 5075/5758 [28:53<03:32,  3.21it/s]
                                                   

 88%|████████▊ | 5075/5758 [28:53<03:32,  3.21it/s]
 88%|████████▊ | 5076/5758 [28:54<03:31,  3.22it/s]
                                                   

 88%|████████▊ | 5076/5758 [28:54<03:31,  3.22it/s]
 88%|████████▊ | 5077/5758 [28:54<03:48,  2.98it/s]
                                                   

 88%|████████▊ | 5077/5758 [28:54<03:48,  2.98it/s]
 88%|████████▊ | 5078/5758 [28:54<03:43,  3.05it/s]
                                                   

 88%|████████▊ | 5078/5758 [28:54<03:43,  3.05it/s]
 88%|████████▊ | 5079/5758 [28:55<03:36,  3.13it/s]
                                                   

 88%|████████▊ | 5079/5758 [28:55<03:36,  3.13it/s]
 88%|████████▊ | 5080/5758 [28:55<03:49,  2.95it/s]
                                                   

 88%|████████▊ | 5080/5758 [28:55<03:49,  2.95it/s]
 88%|████████▊ | 5081/5758 [28:55<04:01,  2.80it/s]
                                                   

 88%|████████▊ | 5081/5758 [28:55<04:01,  2.80it/s]
 88%|████████▊ | 5082/5758 [28:56<04:04,  2.77it/s]
                                                   

 88%|████████▊ | 5082/5758 [28:56<04:04,  2.77it/s]
 88%|████████▊ | 5083/5758 [28:56<03:55,  2.86it/s]
                                                   

 88%|████████▊ | 5083/5758 [28:56<03:55,  2.86it/s]
 88%|████████▊ | 5084/5758 [28:56<03:58,  2.82it/s]
                                                   

 88%|████████▊ | 5084/5758 [28:56<03:58,  2.82it/s]
 88%|████████▊ | 5085/5758 [28:57<03:50,  2.92it/s]
                                                   

 88%|████████▊ | 5085/5758 [28:57<03:50,  2.92it/s]
 88%|████████▊ | 5086/5758 [28:57<03:42,  3.02it/s]
                                                   

 88%|████████▊ | 5086/5758 [28:57<03:42,  3.02it/s]
 88%|████████▊ | 5087/5758 [28:57<03:35,  3.12it/s]
                                                   

 88%|████████▊ | 5087/5758 [28:57<03:35,  3.12it/s]
 88%|████████▊ | 5088/5758 [28:58<04:03,  2.75it/s]
                                                   

 88%|████████▊ | 5088/5758 [28:58<04:03,  2.75it/s]
 88%|████████▊ | 5089/5758 [28:58<04:19,  2.58it/s]
                                                   

 88%|████████▊ | 5089/5758 [28:58<04:19,  2.58it/s]
 88%|████████▊ | 5090/5758 [28:59<04:12,  2.65it/s]
                                                   

 88%|████████▊ | 5090/5758 [28:59<04:12,  2.65it/s]
 88%|████████▊ | 5091/5758 [28:59<04:13,  2.63it/s]
                                                   

 88%|████████▊ | 5091/5758 [28:59<04:13,  2.63it/s]
 88%|████████▊ | 5092/5758 [28:59<04:17,  2.59it/s]
                                                   

 88%|████████▊ | 5092/5758 [28:59<04:17,  2.59it/s]
 88%|█████���██▊ | 5093/5758 [29:00<04:33,  2.43it/s]
                                                   

 88%|████████▊ | 5093/5758 [29:00<04:33,  2.43it/s]
 88%|████████▊ | 5094/5758 [29:00<04:38,  2.38it/s]
                                                   

 88%|████████▊ | 5094/5758 [29:00<04:38,  2.38it/s]
 88%|████████▊ | 5095/5758 [29:01<04:23,  2.51it/s]
                                                   

 88%|████████▊ | 5095/5758 [29:01<04:23,  2.51it/s]
 89%|████████▊ | 5096/5758 [29:01<04:13,  2.61it/s]
                                                   

 89%|████████▊ | 5096/5758 [29:01<04:13,  2.61it/s]
 89%|████████▊ | 5097/5758 [29:01<04:03,  2.71it/s]
                                                   

 89%|████████▊ | 5097/5758 [29:01<04:03,  2.71it/s]
 89%|████████▊ | 5098/5758 [29:02<04:07,  2.67it/s]
                                                   

 89%|████████▊ | 5098/5758 [29:02<04:07,  2.67it/s]
 89%|████████▊ | 5099/5758 [29:02<04:01,  2.72it/s]
                                                   

 89%|████████▊ | 5099/5758 [29:02<04:01,  2.72it/s]
 89%|████████▊ | 5100/5758 [29:03<04:22,  2.50it/s]
                                                   

 89%|████████▊ | 5100/5758 [29:03<04:22,  2.50it/s]
 89%|████████▊ | 5101/5758 [29:03<04:14,  2.58it/s]
                                                   

 89%|████████▊ | 5101/5758 [29:03<04:14,  2.58it/s]
 89%|████████▊ | 5102/5758 [29:03<04:33,  2.40it/s]
                                                   

 89%|████████▊ | 5102/5758 [29:03<04:33,  2.40it/s]
 89%|████████▊ | 5103/5758 [29:04<04:23,  2.49it/s]
                                                   

 89%|████████▊ | 5103/5758 [29:04<04:23,  2.49it/s]
 89%|████████▊ | 5104/5758 [29:04<04:17,  2.54it/s]
                                                   

 89%|████████▊ | 5104/5758 [29:04<04:17,  2.54it/s]
 89%|████████▊ | 5105/5758 [29:05<04:16,  2.55it/s]
                                                   

 89%|████████▊ | 5105/5758 [29:05<04:16,  2.55it/s]
 89%|████████▊ | 5106/5758 [29:05<04:10,  2.60it/s]
                                                   

 89%|████████▊ | 5106/5758 [29:05<04:10,  2.60it/s]
 89%|████████▊ | 5107/5758 [29:05<04:00,  2.70it/s]
                                                   

 89%|████████▊ | 5107/5758 [29:05<04:00,  2.70it/s]
 89%|████████▊ | 5108/5758 [29:06<03:54,  2.78it/s]
                                                   

 89%|████████▊ | 5108/5758 [29:06<03:54,  2.78it/s]
 89%|████████▊ | 5109/5758 [29:06<03:41,  2.94it/s]
                                                   

 89%|████████▊ | 5109/5758 [29:06<03:41,  2.94it/s]
 89%|████████▊ | 5110/5758 [29:06<03:50,  2.81it/s]
                                                   

 89%|████████▊ | 5110/5758 [29:06<03:50,  2.81it/s]
 89%|████████▉ | 5111/5758 [29:07<03:52,  2.79it/s]
                                                   
{'loss': '2.457', 'grad_norm': '19', 'learning_rate': '1.221e-06', 'num_tokens': '3.36e+06', 'mean_token_accuracy': '0.5553', 'epoch': '0.8781'}
+{'loss': '1.207', 'grad_norm': '13.31', 'learning_rate': '1.219e-06', 'num_tokens': '3.361e+06', 'mean_token_accuracy': '0.6557', 'epoch': '0.8783'}
+{'loss': '0.5991', 'grad_norm': '8.938', 'learning_rate': '1.217e-06', 'num_tokens': '3.362e+06', 'mean_token_accuracy': '0.8134', 'epoch': '0.8784'}
+{'loss': '1.141', 'grad_norm': '33.5', 'learning_rate': '1.216e-06', 'num_tokens': '3.362e+06', 'mean_token_accuracy': '0.7458', 'epoch': '0.8786'}
+{'loss': '1.808', 'grad_norm': '46.25', 'learning_rate': '1.214e-06', 'num_tokens': '3.362e+06', 'mean_token_accuracy': '0.6842', 'epoch': '0.8788'}
+{'loss': '1.658', 'grad_norm': '11.19', 'learning_rate': '1.212e-06', 'num_tokens': '3.363e+06', 'mean_token_accuracy': '0.6703', 'epoch': '0.879'}
+{'loss': '0.9896', 'grad_norm': '14.25', 'learning_rate': '1.21e-06', 'num_tokens': '3.363e+06', 'mean_token_accuracy': '0.6981', 'epoch': '0.8791'}
+{'loss': '1.921', 'grad_norm': '17.12', 'learning_rate': '1.209e-06', 'num_tokens': '3.364e+06', 'mean_token_accuracy': '0.5923', 'epoch': '0.8793'}
+{'loss': '1.749', 'grad_norm': '33', 'learning_rate': '1.207e-06', 'num_tokens': '3.364e+06', 'mean_token_accuracy': '0.6525', 'epoch': '0.8795'}
+{'loss': '1.209', 'grad_norm': '11.81', 'learning_rate': '1.205e-06', 'num_tokens': '3.365e+06', 'mean_token_accuracy': '0.7173', 'epoch': '0.8796'}
+{'loss': '1.437', 'grad_norm': '11.94', 'learning_rate': '1.204e-06', 'num_tokens': '3.366e+06', 'mean_token_accuracy': '0.6183', 'epoch': '0.8798'}
+{'loss': '1.111', 'grad_norm': '30', 'learning_rate': '1.202e-06', 'num_tokens': '3.366e+06', 'mean_token_accuracy': '0.7194', 'epoch': '0.88'}
+{'loss': '1.335', 'grad_norm': '14.12', 'learning_rate': '1.2e-06', 'num_tokens': '3.366e+06', 'mean_token_accuracy': '0.6234', 'epoch': '0.8802'}
+{'loss': '1.171', 'grad_norm': '8.812', 'learning_rate': '1.198e-06', 'num_tokens': '3.368e+06', 'mean_token_accuracy': '0.7053', 'epoch': '0.8803'}
+{'loss': '1.839', 'grad_norm': '12.25', 'learning_rate': '1.197e-06', 'num_tokens': '3.369e+06', 'mean_token_accuracy': '0.6102', 'epoch': '0.8805'}
+{'loss': '1.281', 'grad_norm': '12.81', 'learning_rate': '1.195e-06', 'num_tokens': '3.369e+06', 'mean_token_accuracy': '0.6622', 'epoch': '0.8807'}
+{'loss': '1.028', 'grad_norm': '10.75', 'learning_rate': '1.193e-06', 'num_tokens': '3.37e+06', 'mean_token_accuracy': '0.7041', 'epoch': '0.8809'}
+{'loss': '1.209', 'grad_norm': '12.31', 'learning_rate': '1.191e-06', 'num_tokens': '3.371e+06', 'mean_token_accuracy': '0.678', 'epoch': '0.881'}
+{'loss': '1.91', 'grad_norm': '12.81', 'learning_rate': '1.19e-06', 'num_tokens': '3.371e+06', 'mean_token_accuracy': '0.6274', 'epoch': '0.8812'}
+{'loss': '1.458', 'grad_norm': '13.5', 'learning_rate': '1.188e-06', 'num_tokens': '3.372e+06', 'mean_token_accuracy': '0.6034', 'epoch': '0.8814'}
+{'loss': '1.082', 'grad_norm': '12.25', 'learning_rate': '1.186e-06', 'num_tokens': '3.373e+06', 'mean_token_accuracy': '0.696', 'epoch': '0.8816'}
+{'loss': '0.9825', 'grad_norm': '13.81', 'learning_rate': '1.184e-06', 'num_tokens': '3.373e+06', 'mean_token_accuracy': '0.7299', 'epoch': '0.8817'}
+{'loss': '1.16', 'grad_norm': '13.31', 'learning_rate': '1.183e-06', 'num_tokens': '3.374e+06', 'mean_token_accuracy': '0.6822', 'epoch': '0.8819'}
+{'loss': '0.8628', 'grad_norm': '24.38', 'learning_rate': '1.181e-06', 'num_tokens': '3.374e+06', 'mean_token_accuracy': '0.8089', 'epoch': '0.8821'}
+{'loss': '1.026', 'grad_norm': '13.31', 'learning_rate': '1.179e-06', 'num_tokens': '3.375e+06', 'mean_token_accuracy': '0.7111', 'epoch': '0.8823'}
+{'loss': '0.6862', 'grad_norm': '10.12', 'learning_rate': '1.177e-06', 'num_tokens': '3.376e+06', 'mean_token_accuracy': '0.7779', 'epoch': '0.8824'}
+{'loss': '1.465', 'grad_norm': '10.44', 'learning_rate': '1.176e-06', 'num_tokens': '3.377e+06', 'mean_token_accuracy': '0.6893', 'epoch': '0.8826'}
+{'loss': '1.161', 'grad_norm': '8.562', 'learning_rate': '1.174e-06', 'num_tokens': '3.379e+06', 'mean_token_accuracy': '0.7025', 'epoch': '0.8828'}
+{'loss': '0.976', 'grad_norm': '31', 'learning_rate': '1.172e-06', 'num_tokens': '3.379e+06', 'mean_token_accuracy': '0.7719', 'epoch': '0.8829'}
+{'loss': '1.026', 'grad_norm': '11.81', 'learning_rate': '1.171e-06', 'num_tokens': '3.379e+06', 'mean_token_accuracy': '0.7198', 'epoch': '0.8831'}
+{'loss': '1.374', 'grad_norm': '12.88', 'learning_rate': '1.169e-06', 'num_tokens': '3.38e+06', 'mean_token_accuracy': '0.6883', 'epoch': '0.8833'}
+{'loss': '0.9347', 'grad_norm': '12.75', 'learning_rate': '1.167e-06', 'num_tokens': '3.381e+06', 'mean_token_accuracy': '0.723', 'epoch': '0.8835'}
+{'loss': '0.6295', 'grad_norm': '7.031', 'learning_rate': '1.165e-06', 'num_tokens': '3.382e+06', 'mean_token_accuracy': '0.8182', 'epoch': '0.8836'}
+{'loss': '1.414', 'grad_norm': '5.5', 'learning_rate': '1.164e-06', 'num_tokens': '3.386e+06', 'mean_token_accuracy': '0.654', 'epoch': '0.8838'}
+{'loss': '1.977', 'grad_norm': '19.38', 'learning_rate': '1.162e-06', 'num_tokens': '3.387e+06', 'mean_token_accuracy': '0.6098', 'epoch': '0.884'}
+{'loss': '1.864', 'grad_norm': '18.88', 'learning_rate': '1.16e-06', 'num_tokens': '3.387e+06', 'mean_token_accuracy': '0.6277', 'epoch': '0.8842'}
+{'loss': '1.145', 'grad_norm': '28.62', 'learning_rate': '1.158e-06', 'num_tokens': '3.387e+06', 'mean_token_accuracy': '0.7959', 'epoch': '0.8843'}
+{'loss': '1.825', 'grad_norm': '18.38', 'learning_rate': '1.157e-06', 'num_tokens': '3.388e+06', 'mean_token_accuracy': '0.6055', 'epoch': '0.8845'}
+{'loss': '1.36', 'grad_norm': '16.88', 'learning_rate': '1.155e-06', 'num_tokens': '3.389e+06', 'mean_token_accuracy': '0.677', 'epoch': '0.8847'}
+{'loss': '0.9813', 'grad_norm': '10.38', 'learning_rate': '1.153e-06', 'num_tokens': '3.389e+06', 'mean_token_accuracy': '0.8061', 'epoch': '0.8849'}
+{'loss': '0.6577', 'grad_norm': '8.188', 'learning_rate': '1.151e-06', 'num_tokens': '3.39e+06', 'mean_token_accuracy': '0.8058', 'epoch': '0.885'}
+{'loss': '1.104', 'grad_norm': '13.06', 'learning_rate': '1.15e-06', 'num_tokens': '3.391e+06', 'mean_token_accuracy': '0.7346', 'epoch': '0.8852'}
+{'loss': '1.288', 'grad_norm': '11.69', 'learning_rate': '1.148e-06', 'num_tokens': '3.392e+06', 'mean_token_accuracy': '0.6461', 'epoch': '0.8854'}
+{'loss': '1.272', 'grad_norm': '9.188', 'learning_rate': '1.146e-06', 'num_tokens': '3.393e+06', 'mean_token_accuracy': '0.6545', 'epoch': '0.8856'}
+{'loss': '1.141', 'grad_norm': '6.719', 'learning_rate': '1.144e-06', 'num_tokens': '3.396e+06', 'mean_token_accuracy': '0.7128', 'epoch': '0.8857'}
+{'loss': '1.142', 'grad_norm': '13.12', 'learning_rate': '1.143e-06', 'num_tokens': '3.396e+06', 'mean_token_accuracy': '0.7001', 'epoch': '0.8859'}
+{'loss': '1.95', 'grad_norm': '41.75', 'learning_rate': '1.141e-06', 'num_tokens': '3.396e+06', 'mean_token_accuracy': '0.6331', 'epoch': '0.8861'}
+{'loss': '1.069', 'grad_norm': '36', 'learning_rate': '1.139e-06', 'num_tokens': '3.397e+06', 'mean_token_accuracy': '0.7527', 'epoch': '0.8862'}
+{'loss': '1.009', 'grad_norm': '29.88', 'learning_rate': '1.138e-06', 'num_tokens': '3.397e+06', 'mean_token_accuracy': '0.7903', 'epoch': '0.8864'}
+{'loss': '0.7318', 'grad_norm': '9.125', 'learning_rate': '1.136e-06', 'num_tokens': '3.398e+06', 'mean_token_accuracy': '0.7823', 'epoch': '0.8866'}
+{'loss': '1.95', 'grad_norm': '18.12', 'learning_rate': '1.134e-06', 'num_tokens': '3.398e+06', 'mean_token_accuracy': '0.6071', 'epoch': '0.8868'}
+{'loss': '1.83', 'grad_norm': '19.38', 'learning_rate': '1.132e-06', 'num_tokens': '3.399e+06', 'mean_token_accuracy': '0.6088', 'epoch': '0.8869'}
+{'loss': '1.317', 'grad_norm': '12', 'learning_rate': '1.131e-06', 'num_tokens': '3.4e+06', 'mean_token_accuracy': '0.6615', 'epoch': '0.8871'}
+{'loss': '0.7243', 'grad_norm': '10', 'learning_rate': '1.129e-06', 'num_tokens': '3.401e+06', 'mean_token_accuracy': '0.777', 'epoch': '0.8873'}
+{'loss': '0.6628', 'grad_norm': '25.75', 'learning_rate': '1.127e-06', 'num_tokens': '3.401e+06', 'mean_token_accuracy': '0.8714', 'epoch': '0.8875'}
+
 89%|████████▉ | 5111/5758 [29:07<03:52,  2.79it/s]
 89%|████████▉ | 5112/5758 [29:07<03:44,  2.88it/s]
                                                   

 89%|████████▉ | 5112/5758 [29:07<03:44,  2.88it/s]
 89%|████████▉ | 5113/5758 [29:07<03:40,  2.92it/s]
                                                   

 89%|████████▉ | 5113/5758 [29:07<03:40,  2.92it/s]
 89%|████████▉ | 5114/5758 [29:08<03:53,  2.76it/s]
                                                   

 89%|████████▉ | 5114/5758 [29:08<03:53,  2.76it/s]
 89%|████████▉ | 5115/5758 [29:08<03:55,  2.73it/s]
                                                   

 89%|████████▉ | 5115/5758 [29:08<03:55,  2.73it/s]
 89%|████████▉ | 5116/5758 [29:08<03:41,  2.89it/s]
                                                   

 89%|████████▉ | 5116/5758 [29:08<03:41,  2.89it/s]
 89%|████████▉ | 5117/5758 [29:09<03:34,  2.99it/s]
                                                   

 89%|████████▉ | 5117/5758 [29:09<03:34,  2.99it/s]
 89%|████████▉ | 5118/5758 [29:09<03:30,  3.03it/s]
                                                   

 89%|████████▉ | 5118/5758 [29:09<03:30,  3.03it/s]
 89%|████████▉ | 5119/5758 [29:09<03:34,  2.97it/s]
                                                   

 89%|████████▉ | 5119/5758 [29:09<03:34,  2.97it/s]
 89%|████████▉ | 5120/5758 [29:10<03:26,  3.09it/s]
                                                   

 89%|████████▉ | 5120/5758 [29:10<03:26,  3.09it/s]
 89%|████████▉ | 5121/5758 [29:10<03:35,  2.96it/s]
                                                   

 89%|████████▉ | 5121/5758 [29:10<03:35,  2.96it/s]
 89%|████████▉ | 5122/5758 [29:10<03:28,  3.05it/s]
                                                   

 89%|████████▉ | 5122/5758 [29:10<03:28,  3.05it/s]
 89%|████████▉ | 5123/5758 [29:11<03:22,  3.14it/s]
                                                   

 89%|████████▉ | 5123/5758 [29:11<03:22,  3.14it/s]
 89%|████████▉ | 5124/5758 [29:11<03:16,  3.23it/s]
                                                   

 89%|████████▉ | 5124/5758 [29:11<03:16,  3.23it/s]
 89%|████████▉ | 5125/5758 [29:11<03:15,  3.24it/s]
                                                   

 89%|████████▉ | 5125/5758 [29:11<03:15,  3.24it/s]
 89%|████████▉ | 5126/5758 [29:12<03:26,  3.06it/s]
                                                   

 89%|████████▉ | 5126/5758 [29:12<03:26,  3.06it/s]
 89%|████████▉ | 5127/5758 [29:12<03:29,  3.01it/s]
                                                   

 89%|████████▉ | 5127/5758 [29:12<03:29,  3.01it/s]
 89%|████████▉ | 5128/5758 [29:12<03:34,  2.94it/s]
                                                   

 89%|████████▉ | 5128/5758 [29:12<03:34,  2.94it/s]
 89%|████████▉ | 5129/5758 [29:13<03:31,  2.97it/s]
                                                   

 89%|████████▉ | 5129/5758 [29:13<03:31,  2.97it/s]
 89%|████████▉ | 5130/5758 [29:13<03:29,  3.00it/s]
                                                   

 89%|████████▉ | 5130/5758 [29:13<03:29,  3.00it/s]
 89%|████████▉ | 5131/5758 [29:13<03:29,  2.99it/s]
                                                   

 89%|████████▉ | 5131/5758 [29:13<03:29,  2.99it/s]
 89%|████████▉ | 5132/5758 [29:14<03:26,  3.03it/s]
                                                   

 89%|████████▉ | 5132/5758 [29:14<03:26,  3.03it/s]
 89%|████████▉ | 5133/5758 [29:14<03:37,  2.87it/s]
                                                   

 89%|████████▉ | 5133/5758 [29:14<03:37,  2.87it/s]
 89%|████████▉ | 5134/5758 [29:14<03:37,  2.86it/s]
                                                   

 89%|████████▉ | 5134/5758 [29:14<03:37,  2.86it/s]
 89%|████████▉ | 5135/5758 [29:15<03:40,  2.83it/s]
                                                   

 89%|████████▉ | 5135/5758 [29:15<03:40,  2.83it/s]
 89%|████████▉ | 5136/5758 [29:15<03:29,  2.97it/s]
                                                   

 89%|████████▉ | 5136/5758 [29:15<03:29,  2.97it/s]
 89%|████████▉ | 5137/5758 [29:15<03:24,  3.03it/s]
                                                   

 89%|████████▉ | 5137/5758 [29:15<03:24,  3.03it/s]
 89%|████████▉ | 5138/5758 [29:16<03:33,  2.90it/s]
                                                   

 89%|████████▉ | 5138/5758 [29:16<03:33,  2.90it/s]
 89%|████████▉ | 5139/5758 [29:16<03:38,  2.83it/s]
                                                   

 89%|████████▉ | 5139/5758 [29:16<03:38,  2.83it/s]
 89%|████████▉ | 5140/5758 [29:17<03:57,  2.60it/s]
                                                   

 89%|████████▉ | 5140/5758 [29:17<03:57,  2.60it/s]
 89%|████████▉ | 5141/5758 [29:17<03:48,  2.70it/s]
                                                   

 89%|████████▉ | 5141/5758 [29:17<03:48,  2.70it/s]
 89%|████████▉ | 5142/5758 [29:17<03:39,  2.81it/s]
                                                   

 89%|████████▉ | 5142/5758 [29:17<03:39,  2.81it/s]
 89%|████████▉ | 5143/5758 [29:17<03:28,  2.95it/s]
                                                   

 89%|████████▉ | 5143/5758 [29:17<03:28,  2.95it/s]
 89%|████████▉ | 5144/5758 [29:18<03:23,  3.01it/s]
                                                   

 89%|████████▉ | 5144/5758 [29:18<03:23,  3.01it/s]
 89%|████████▉ | 5145/5758 [29:18<03:15,  3.13it/s]
                                                   

 89%|████████▉ | 5145/5758 [29:18<03:15,  3.13it/s]
 89%|████████▉ | 5146/5758 [29:18<03:15,  3.13it/s]
                                                   

 89%|████████▉ | 5146/5758 [29:18<03:15,  3.13it/s]
 89%|████████▉ | 5147/5758 [29:19<03:18,  3.08it/s]
                                                   

 89%|████████▉ | 5147/5758 [29:19<03:18,  3.08it/s]
 89%|████████▉ | 5148/5758 [29:19<03:18,  3.07it/s]
                                                   

 89%|████████▉ | 5148/5758 [29:19<03:18,  3.07it/s]
 89%|████████▉ | 5149/5758 [29:19<03:12,  3.16it/s]
                                                   

 89%|████████▉ | 5149/5758 [29:19<03:12,  3.16it/s]
 89%|████████▉ | 5150/5758 [29:20<03:19,  3.04it/s]
                                                   

 89%|████████▉ | 5150/5758 [29:20<03:19,  3.04it/s]
 89%|████████▉ | 5151/5758 [29:20<03:18,  3.06it/s]
                                                   

 89%|████████▉ | 5151/5758 [29:20<03:18,  3.06it/s]
 89%|████████▉ | 5152/5758 [29:20<03:16,  3.09it/s]
                                                   

 89%|████████▉ | 5152/5758 [29:20<03:16,  3.09it/s]
 89%|████████▉ | 5153/5758 [29:21<03:20,  3.02it/s]
                                                   

 89%|████████▉ | 5153/5758 [29:21<03:20,  3.02it/s]
 90%|████████▉ | 5154/5758 [29:21<03:28,  2.89it/s]
                                                   

 90%|████████▉ | 5154/5758 [29:21<03:28,  2.89it/s]
 90%|████████▉ | 5155/5758 [29:21<03:22,  2.98it/s]
                                                   

 90%|████████▉ | 5155/5758 [29:21<03:22,  2.98it/s]
 90%|████████▉ | 5156/5758 [29:22<03:30,  2.86it/s]
                                                   

 90%|██���█████▉ | 5156/5758 [29:22<03:30,  2.86it/s]
 90%|████████▉ | 5157/5758 [29:22<03:17,  3.04it/s]
                                                   

 90%|████████▉ | 5157/5758 [29:22<03:17,  3.04it/s]
 90%|████████▉ | 5158/5758 [29:22<03:14,  3.08it/s]
                                                   

 90%|████████▉ | 5158/5758 [29:22<03:14,  3.08it/s]
 90%|████████▉ | 5159/5758 [29:23<03:26,  2.91it/s]
                                                   

 90%|████████▉ | 5159/5758 [29:23<03:26,  2.91it/s]
 90%|████████▉ | 5160/5758 [29:23<03:23,  2.94it/s]
                                                   

 90%|████████▉ | 5160/5758 [29:23<03:23,  2.94it/s]
 90%|████████▉ | 5161/5758 [29:23<03:31,  2.82it/s]
                                                   

 90%|████████▉ | 5161/5758 [29:23<03:31,  2.82it/s]
 90%|████████▉ | 5162/5758 [29:24<03:37,  2.74it/s]
                                                   

 90%|████████▉ | 5162/5758 [29:24<03:37,  2.74it/s]
 90%|████████▉ | 5163/5758 [29:24<03:27,  2.86it/s]
                                                   

 90%|████████▉ | 5163/5758 [29:24<03:27,  2.86it/s]
 90%|████████▉ | 5164/5758 [29:24<03:17,  3.01it/s]
                                                   

 90%|████████▉ | 5164/5758 [29:24<03:17,  3.01it/s]
 90%|████████▉ | 5165/5758 [29:25<03:18,  2.99it/s]
                                                   

 90%|████████▉ | 5165/5758 [29:25<03:18,  2.99it/s]
 90%|████████▉ | 5166/5758 [29:25<03:31,  2.80it/s]
                                                   
{'loss': '1.284', 'grad_norm': '10.19', 'learning_rate': '1.125e-06', 'num_tokens': '3.402e+06', 'mean_token_accuracy': '0.68', 'epoch': '0.8876'}
+{'loss': '1.37', 'grad_norm': '12.56', 'learning_rate': '1.124e-06', 'num_tokens': '3.403e+06', 'mean_token_accuracy': '0.6346', 'epoch': '0.8878'}
+{'loss': '2.202', 'grad_norm': '22.5', 'learning_rate': '1.122e-06', 'num_tokens': '3.403e+06', 'mean_token_accuracy': '0.5571', 'epoch': '0.888'}
+{'loss': '0.8154', 'grad_norm': '32.5', 'learning_rate': '1.12e-06', 'num_tokens': '3.403e+06', 'mean_token_accuracy': '0.7971', 'epoch': '0.8882'}
+{'loss': '1.749', 'grad_norm': '18.62', 'learning_rate': '1.118e-06', 'num_tokens': '3.404e+06', 'mean_token_accuracy': '0.6333', 'epoch': '0.8883'}
+{'loss': '0.5678', 'grad_norm': '15.69', 'learning_rate': '1.117e-06', 'num_tokens': '3.404e+06', 'mean_token_accuracy': '0.8609', 'epoch': '0.8885'}
+{'loss': '1.725', 'grad_norm': '10.44', 'learning_rate': '1.115e-06', 'num_tokens': '3.405e+06', 'mean_token_accuracy': '0.648', 'epoch': '0.8887'}
+{'loss': '0.497', 'grad_norm': '6.844', 'learning_rate': '1.113e-06', 'num_tokens': '3.407e+06', 'mean_token_accuracy': '0.8563', 'epoch': '0.8889'}
+{'loss': '1.13', 'grad_norm': '16.5', 'learning_rate': '1.111e-06', 'num_tokens': '3.407e+06', 'mean_token_accuracy': '0.7182', 'epoch': '0.889'}
+{'loss': '1.6', 'grad_norm': '13.44', 'learning_rate': '1.11e-06', 'num_tokens': '3.408e+06', 'mean_token_accuracy': '0.6746', 'epoch': '0.8892'}
+{'loss': '1.053', 'grad_norm': '13.69', 'learning_rate': '1.108e-06', 'num_tokens': '3.408e+06', 'mean_token_accuracy': '0.7336', 'epoch': '0.8894'}
+{'loss': '1.181', 'grad_norm': '12.5', 'learning_rate': '1.106e-06', 'num_tokens': '3.409e+06', 'mean_token_accuracy': '0.6872', 'epoch': '0.8895'}
+{'loss': '1.676', 'grad_norm': '12.69', 'learning_rate': '1.105e-06', 'num_tokens': '3.41e+06', 'mean_token_accuracy': '0.6538', 'epoch': '0.8897'}
+{'loss': '1.302', 'grad_norm': '14', 'learning_rate': '1.103e-06', 'num_tokens': '3.41e+06', 'mean_token_accuracy': '0.6768', 'epoch': '0.8899'}
+{'loss': '1.186', 'grad_norm': '15.75', 'learning_rate': '1.101e-06', 'num_tokens': '3.411e+06', 'mean_token_accuracy': '0.675', 'epoch': '0.8901'}
+{'loss': '1.489', 'grad_norm': '28.88', 'learning_rate': '1.099e-06', 'num_tokens': '3.411e+06', 'mean_token_accuracy': '0.6992', 'epoch': '0.8902'}
+{'loss': '1.277', 'grad_norm': '16.38', 'learning_rate': '1.098e-06', 'num_tokens': '3.412e+06', 'mean_token_accuracy': '0.7441', 'epoch': '0.8904'}
+{'loss': '1.31', 'grad_norm': '24', 'learning_rate': '1.096e-06', 'num_tokens': '3.412e+06', 'mean_token_accuracy': '0.7412', 'epoch': '0.8906'}
+{'loss': '1.463', 'grad_norm': '13.75', 'learning_rate': '1.094e-06', 'num_tokens': '3.413e+06', 'mean_token_accuracy': '0.6533', 'epoch': '0.8908'}
+{'loss': '1.851', 'grad_norm': '15.81', 'learning_rate': '1.092e-06', 'num_tokens': '3.413e+06', 'mean_token_accuracy': '0.6029', 'epoch': '0.8909'}
+{'loss': '1.156', 'grad_norm': '25.38', 'learning_rate': '1.091e-06', 'num_tokens': '3.413e+06', 'mean_token_accuracy': '0.7523', 'epoch': '0.8911'}
+{'loss': '0.994', 'grad_norm': '28', 'learning_rate': '1.089e-06', 'num_tokens': '3.414e+06', 'mean_token_accuracy': '0.7708', 'epoch': '0.8913'}
+{'loss': '0.9024', 'grad_norm': '13.44', 'learning_rate': '1.087e-06', 'num_tokens': '3.414e+06', 'mean_token_accuracy': '0.7247', 'epoch': '0.8915'}
+{'loss': '1.237', 'grad_norm': '7.312', 'learning_rate': '1.085e-06', 'num_tokens': '3.416e+06', 'mean_token_accuracy': '0.7066', 'epoch': '0.8916'}
+{'loss': '1.473', 'grad_norm': '21.62', 'learning_rate': '1.084e-06', 'num_tokens': '3.417e+06', 'mean_token_accuracy': '0.6098', 'epoch': '0.8918'}
+{'loss': '1.796', 'grad_norm': '18.5', 'learning_rate': '1.082e-06', 'num_tokens': '3.417e+06', 'mean_token_accuracy': '0.6502', 'epoch': '0.892'}
+{'loss': '0.76', 'grad_norm': '8.062', 'learning_rate': '1.08e-06', 'num_tokens': '3.419e+06', 'mean_token_accuracy': '0.7654', 'epoch': '0.8922'}
+{'loss': '0.8001', 'grad_norm': '9.562', 'learning_rate': '1.078e-06', 'num_tokens': '3.42e+06', 'mean_token_accuracy': '0.7641', 'epoch': '0.8923'}
+{'loss': '1.499', 'grad_norm': '25', 'learning_rate': '1.077e-06', 'num_tokens': '3.42e+06', 'mean_token_accuracy': '0.717', 'epoch': '0.8925'}
+{'loss': '2.155', 'grad_norm': '18', 'learning_rate': '1.075e-06', 'num_tokens': '3.42e+06', 'mean_token_accuracy': '0.5767', 'epoch': '0.8927'}
+{'loss': '0.8687', 'grad_norm': '18.12', 'learning_rate': '1.073e-06', 'num_tokens': '3.421e+06', 'mean_token_accuracy': '0.7652', 'epoch': '0.8928'}
+{'loss': '0.6668', 'grad_norm': '34.25', 'learning_rate': '1.072e-06', 'num_tokens': '3.421e+06', 'mean_token_accuracy': '0.8242', 'epoch': '0.893'}
+{'loss': '1.116', 'grad_norm': '12.5', 'learning_rate': '1.07e-06', 'num_tokens': '3.422e+06', 'mean_token_accuracy': '0.6986', 'epoch': '0.8932'}
+{'loss': '0.6281', 'grad_norm': '7.469', 'learning_rate': '1.068e-06', 'num_tokens': '3.423e+06', 'mean_token_accuracy': '0.809', 'epoch': '0.8934'}
+{'loss': '1.651', 'grad_norm': '31.88', 'learning_rate': '1.066e-06', 'num_tokens': '3.423e+06', 'mean_token_accuracy': '0.6159', 'epoch': '0.8935'}
+{'loss': '2', 'grad_norm': '15.81', 'learning_rate': '1.065e-06', 'num_tokens': '3.424e+06', 'mean_token_accuracy': '0.5885', 'epoch': '0.8937'}
+{'loss': '0.9745', 'grad_norm': '12.31', 'learning_rate': '1.063e-06', 'num_tokens': '3.424e+06', 'mean_token_accuracy': '0.7845', 'epoch': '0.8939'}
+{'loss': '1.063', 'grad_norm': '7.781', 'learning_rate': '1.061e-06', 'num_tokens': '3.426e+06', 'mean_token_accuracy': '0.7396', 'epoch': '0.8941'}
+{'loss': '1.147', 'grad_norm': '13.62', 'learning_rate': '1.059e-06', 'num_tokens': '3.427e+06', 'mean_token_accuracy': '0.7114', 'epoch': '0.8942'}
+{'loss': '2.009', 'grad_norm': '14.38', 'learning_rate': '1.058e-06', 'num_tokens': '3.427e+06', 'mean_token_accuracy': '0.5683', 'epoch': '0.8944'}
+{'loss': '1.09', 'grad_norm': '8.25', 'learning_rate': '1.056e-06', 'num_tokens': '3.429e+06', 'mean_token_accuracy': '0.737', 'epoch': '0.8946'}
+{'loss': '0.6321', 'grad_norm': '8.125', 'learning_rate': '1.054e-06', 'num_tokens': '3.43e+06', 'mean_token_accuracy': '0.8216', 'epoch': '0.8948'}
+{'loss': '2.241', 'grad_norm': '15.5', 'learning_rate': '1.052e-06', 'num_tokens': '3.431e+06', 'mean_token_accuracy': '0.5614', 'epoch': '0.8949'}
+{'loss': '1.164', 'grad_norm': '13.62', 'learning_rate': '1.051e-06', 'num_tokens': '3.431e+06', 'mean_token_accuracy': '0.6876', 'epoch': '0.8951'}
+{'loss': '0.9736', 'grad_norm': '15.75', 'learning_rate': '1.049e-06', 'num_tokens': '3.432e+06', 'mean_token_accuracy': '0.7624', 'epoch': '0.8953'}
+{'loss': '1.738', 'grad_norm': '20', 'learning_rate': '1.047e-06', 'num_tokens': '3.433e+06', 'mean_token_accuracy': '0.6209', 'epoch': '0.8954'}
+{'loss': '0.9547', 'grad_norm': '40.5', 'learning_rate': '1.046e-06', 'num_tokens': '3.433e+06', 'mean_token_accuracy': '0.807', 'epoch': '0.8956'}
+{'loss': '0.3125', 'grad_norm': '28.75', 'learning_rate': '1.044e-06', 'num_tokens': '3.433e+06', 'mean_token_accuracy': '0.9036', 'epoch': '0.8958'}
+{'loss': '0.6115', 'grad_norm': '8.812', 'learning_rate': '1.042e-06', 'num_tokens': '3.434e+06', 'mean_token_accuracy': '0.8193', 'epoch': '0.896'}
+{'loss': '1.755', 'grad_norm': '30.62', 'learning_rate': '1.04e-06', 'num_tokens': '3.434e+06', 'mean_token_accuracy': '0.6143', 'epoch': '0.8961'}
+{'loss': '1.541', 'grad_norm': '11.94', 'learning_rate': '1.039e-06', 'num_tokens': '3.435e+06', 'mean_token_accuracy': '0.6072', 'epoch': '0.8963'}
+{'loss': '1.203', 'grad_norm': '30.5', 'learning_rate': '1.037e-06', 'num_tokens': '3.435e+06', 'mean_token_accuracy': '0.7462', 'epoch': '0.8965'}
+{'loss': '1.397', 'grad_norm': '13.56', 'learning_rate': '1.035e-06', 'num_tokens': '3.436e+06', 'mean_token_accuracy': '0.6485', 'epoch': '0.8967'}
+{'loss': '2.014', 'grad_norm': '17.25', 'learning_rate': '1.033e-06', 'num_tokens': '3.437e+06', 'mean_token_accuracy': '0.6072', 'epoch': '0.8968'}
+{'loss': '0.5575', 'grad_norm': '40.5', 'learning_rate': '1.032e-06', 'num_tokens': '3.437e+06', 'mean_token_accuracy': '0.8462', 'epoch': '0.897'}
+
 90%|████████▉ | 5166/5758 [29:25<03:31,  2.80it/s]
 90%|████████▉ | 5167/5758 [29:26<03:25,  2.88it/s]
                                                   

 90%|████████▉ | 5167/5758 [29:26<03:25,  2.88it/s]
 90%|████████▉ | 5168/5758 [29:26<03:21,  2.92it/s]
                                                   

 90%|████████▉ | 5168/5758 [29:26<03:21,  2.92it/s]
 90%|████████▉ | 5169/5758 [29:26<03:20,  2.93it/s]
                                                   

 90%|████████▉ | 5169/5758 [29:26<03:20,  2.93it/s]
 90%|████████▉ | 5170/5758 [29:27<03:24,  2.88it/s]
                                                   

 90%|████████▉ | 5170/5758 [29:27<03:24,  2.88it/s]
 90%|████████▉ | 5171/5758 [29:27<03:14,  3.03it/s]
                                                   

 90%|████████▉ | 5171/5758 [29:27<03:14,  3.03it/s]
 90%|████████▉ | 5172/5758 [29:27<03:24,  2.87it/s]
                                                   

 90%|████████▉ | 5172/5758 [29:27<03:24,  2.87it/s]
 90%|████████▉ | 5173/5758 [29:28<03:28,  2.81it/s]
                                                   

 90%|████████▉ | 5173/5758 [29:28<03:28,  2.81it/s]
 90%|████████▉ | 5174/5758 [29:28<03:30,  2.77it/s]
                                                   

 90%|████████▉ | 5174/5758 [29:28<03:30,  2.77it/s]
 90%|████████▉ | 5175/5758 [29:28<03:25,  2.84it/s]
                                                   

 90%|████████▉ | 5175/5758 [29:28<03:25,  2.84it/s]
 90%|████████▉ | 5176/5758 [29:29<03:20,  2.90it/s]
                                                   

 90%|████████▉ | 5176/5758 [29:29<03:20,  2.90it/s]
 90%|████████▉ | 5177/5758 [29:29<03:25,  2.82it/s]
                                                   

 90%|████████▉ | 5177/5758 [29:29<03:25,  2.82it/s]
 90%|████████▉ | 5178/5758 [29:29<03:19,  2.91it/s]
                                                   

 90%|████████▉ | 5178/5758 [29:29<03:19,  2.91it/s]
 90%|████████▉ | 5179/5758 [29:30<03:11,  3.03it/s]
                                                   

 90%|████████▉ | 5179/5758 [29:30<03:11,  3.03it/s]
 90%|████████▉ | 5180/5758 [29:30<03:00,  3.19it/s]
                                                   

 90%|████████▉ | 5180/5758 [29:30<03:00,  3.19it/s]
 90%|████████▉ | 5181/5758 [29:30<03:01,  3.18it/s]
                                                   

 90%|████████▉ | 5181/5758 [29:30<03:01,  3.18it/s]
 90%|████████▉ | 5182/5758 [29:31<03:10,  3.02it/s]
                                                   

 90%|████████▉ | 5182/5758 [29:31<03:10,  3.02it/s]
 90%|█████████ | 5183/5758 [29:31<03:18,  2.89it/s]
                                                   

 90%|█████████ | 5183/5758 [29:31<03:18,  2.89it/s]
 90%|█████████ | 5184/5758 [29:31<03:13,  2.97it/s]
                                                   

 90%|█████████ | 5184/5758 [29:31<03:13,  2.97it/s]
 90%|█████████ | 5185/5758 [29:32<03:16,  2.91it/s]
                                                   

 90%|█████████ | 5185/5758 [29:32<03:16,  2.91it/s]
 90%|█████████ | 5186/5758 [29:32<03:12,  2.97it/s]
                                                   

 90%|█████████ | 5186/5758 [29:32<03:12,  2.97it/s]
 90%|█████████ | 5187/5758 [29:32<03:08,  3.04it/s]
                                                   

 90%|█████████ | 5187/5758 [29:32<03:08,  3.04it/s]
 90%|█████████ | 5188/5758 [29:33<03:02,  3.13it/s]
                                                   

 90%|█████████ | 5188/5758 [29:33<03:02,  3.13it/s]
 90%|█████████ | 5189/5758 [29:33<03:08,  3.02it/s]
                                                   

 90%|█████████ | 5189/5758 [29:33<03:08,  3.02it/s]
 90%|█████████ | 5190/5758 [29:33<03:14,  2.92it/s]
                                                   

 90%|█████████ | 5190/5758 [29:33<03:14,  2.92it/s]
 90%|█████████ | 5191/5758 [29:34<03:13,  2.93it/s]
                                                   

 90%|█████████ | 5191/5758 [29:34<03:13,  2.93it/s]
 90%|█████████ | 5192/5758 [29:34<03:16,  2.88it/s]
                                                   

 90%|█████████ | 5192/5758 [29:34<03:16,  2.88it/s]
 90%|█████████ | 5193/5758 [29:35<03:40,  2.56it/s]
                                                   

 90%|█████████ | 5193/5758 [29:35<03:40,  2.56it/s]
 90%|█████████ | 5194/5758 [29:35<03:31,  2.67it/s]
                                                   

 90%|█████████ | 5194/5758 [29:35<03:31,  2.67it/s]
 90%|█████████ | 5195/5758 [29:35<03:17,  2.86it/s]
                                                   

 90%|█████████ | 5195/5758 [29:35<03:17,  2.86it/s]
 90%|█████████ | 5196/5758 [29:36<03:20,  2.81it/s]
                                                   

 90%|█████████ | 5196/5758 [29:36<03:20,  2.81it/s]
 90%|█████████ | 5197/5758 [29:36<03:09,  2.96it/s]
                                                   

 90%|█████████ | 5197/5758 [29:36<03:09,  2.96it/s]
 90%|█████████ | 5198/5758 [29:36<03:06,  3.01it/s]
                                                   

 90%|█████████ | 5198/5758 [29:36<03:06,  3.01it/s]
 90%|█████████ | 5199/5758 [29:37<03:15,  2.86it/s]
                                                   

 90%|█████████ | 5199/5758 [29:37<03:15,  2.86it/s]
 90%|█████████ | 5200/5758 [29:37<03:12,  2.89it/s]
                                                   

 90%|█████████ | 5200/5758 [29:37<03:12,  2.89it/s]
 90%|█████████ | 5201/5758 [29:37<03:17,  2.82it/s]
                                                   

 90%|█████████ | 5201/5758 [29:37<03:17,  2.82it/s]
 90%|█████████ | 5202/5758 [29:38<03:22,  2.74it/s]
                                                   

 90%|█████████ | 5202/5758 [29:38<03:22,  2.74it/s]
 90%|█████████ | 5203/5758 [29:38<03:22,  2.73it/s]
                                                   

 90%|█████████ | 5203/5758 [29:38<03:22,  2.73it/s]
 90%|█████████ | 5204/5758 [29:38<03:11,  2.89it/s]
                                                   

 90%|█████████ | 5204/5758 [29:38<03:11,  2.89it/s]
 90%|█████████ | 5205/5758 [29:39<03:04,  3.00it/s]
                                                   

 90%|█████████ | 5205/5758 [29:39<03:04,  3.00it/s]
 90%|█████████ | 5206/5758 [29:39<03:08,  2.93it/s]
                                                   

 90%|█████████ | 5206/5758 [29:39<03:08,  2.93it/s]
 90%|█████████ | 5207/5758 [29:39<03:04,  2.99it/s]
                                                   

 90%|█████████ | 5207/5758 [29:39<03:04,  2.99it/s]
 90%|█████████ | 5208/5758 [29:40<03:15,  2.82it/s]
                                                   

 90%|█████████ | 5208/5758 [29:40<03:15,  2.82it/s]
 90%|█████████ | 5209/5758 [29:40<03:10,  2.88it/s]
                                                   

 90%|█████████ | 5209/5758 [29:40<03:10,  2.88it/s]
 90%|█████████ | 5210/5758 [29:40<03:08,  2.90it/s]
                                                   

 90%|█████████ | 5210/5758 [29:40<03:08,  2.90it/s]
 91%|█████████ | 5211/5758 [29:41<03:11,  2.85it/s]
                                                   

 91%|█████████ | 5211/5758 [29:41<03:11,  2.85it/s]
 91%|█████████ | 5212/5758 [29:41<03:00,  3.02it/s]
                                                   

 91%|█████████ | 5212/5758 [29:41<03:00,  3.02it/s]
 91%|█████████ | 5213/5758 [29:41<03:12,  2.83it/s]
                                                   

 91%|█████████ | 5213/5758 [29:41<03:12,  2.83it/s]
 91%|█████████ | 5214/5758 [29:42<03:09,  2.87it/s]
                                                   

 91%|█████████ | 5214/5758 [29:42<03:09,  2.87it/s]
 91%|█████████ | 5215/5758 [29:42<03:12,  2.82it/s]
                                                   

 91%|█████████ | 5215/5758 [29:42<03:12,  2.82it/s]
 91%|█████████ | 5216/5758 [29:42<03:15,  2.78it/s]
                                                   

 91%|█████████ | 5216/5758 [29:42<03:15,  2.78it/s]
 91%|█████████ | 5217/5758 [29:43<03:19,  2.71it/s]
                                                   

 91%|█████████ | 5217/5758 [29:43<03:19,  2.71it/s]
 91%|█████████ | 5218/5758 [29:43<03:08,  2.87it/s]
                                                   

 91%|█████████ | 5218/5758 [29:43<03:08,  2.87it/s]
 91%|█████████ | 5219/5758 [29:44<03:15,  2.76it/s]
                                                   

 91%|█████████ | 5219/5758 [29:44<03:15,  2.76it/s]
 91%|█████████ | 5220/5758 [29:44<03:09,  2.84it/s]
                                                   

 91%|█████████ | 5220/5758 [29:44<03:09,  2.84it/s]
 91%|█████████ | 5221/5758 [29:44<03:33,  2.51it/s]
                                                   
{'loss': '1.45', 'grad_norm': '9.812', 'learning_rate': '1.03e-06', 'num_tokens': '3.438e+06', 'mean_token_accuracy': '0.6579', 'epoch': '0.8972'}
+{'loss': '1.226', 'grad_norm': '5.875', 'learning_rate': '1.028e-06', 'num_tokens': '3.441e+06', 'mean_token_accuracy': '0.7057', 'epoch': '0.8974'}
+{'loss': '0.5872', 'grad_norm': '7.438', 'learning_rate': '1.026e-06', 'num_tokens': '3.443e+06', 'mean_token_accuracy': '0.823', 'epoch': '0.8975'}
+{'loss': '1.117', 'grad_norm': '11.69', 'learning_rate': '1.025e-06', 'num_tokens': '3.444e+06', 'mean_token_accuracy': '0.6781', 'epoch': '0.8977'}
+{'loss': '0.9408', 'grad_norm': '9.438', 'learning_rate': '1.023e-06', 'num_tokens': '3.445e+06', 'mean_token_accuracy': '0.7812', 'epoch': '0.8979'}
+{'loss': '0.993', 'grad_norm': '15.5', 'learning_rate': '1.021e-06', 'num_tokens': '3.445e+06', 'mean_token_accuracy': '0.7209', 'epoch': '0.8981'}
+{'loss': '1.525', 'grad_norm': '10.44', 'learning_rate': '1.019e-06', 'num_tokens': '3.447e+06', 'mean_token_accuracy': '0.6533', 'epoch': '0.8982'}
+{'loss': '1.116', 'grad_norm': '14.31', 'learning_rate': '1.018e-06', 'num_tokens': '3.447e+06', 'mean_token_accuracy': '0.6599', 'epoch': '0.8984'}
+{'loss': '1.767', 'grad_norm': '17.75', 'learning_rate': '1.016e-06', 'num_tokens': '3.448e+06', 'mean_token_accuracy': '0.6353', 'epoch': '0.8986'}
+{'loss': '1.119', 'grad_norm': '11.31', 'learning_rate': '1.014e-06', 'num_tokens': '3.449e+06', 'mean_token_accuracy': '0.6758', 'epoch': '0.8987'}
+{'loss': '1.329', 'grad_norm': '11.88', 'learning_rate': '1.013e-06', 'num_tokens': '3.449e+06', 'mean_token_accuracy': '0.6351', 'epoch': '0.8989'}
+{'loss': '1.534', 'grad_norm': '21', 'learning_rate': '1.011e-06', 'num_tokens': '3.45e+06', 'mean_token_accuracy': '0.6942', 'epoch': '0.8991'}
+{'loss': '2.256', 'grad_norm': '21.12', 'learning_rate': '1.009e-06', 'num_tokens': '3.45e+06', 'mean_token_accuracy': '0.5722', 'epoch': '0.8993'}
+{'loss': '0.489', 'grad_norm': '7.688', 'learning_rate': '1.007e-06', 'num_tokens': '3.451e+06', 'mean_token_accuracy': '0.8518', 'epoch': '0.8994'}
+{'loss': '0.6564', 'grad_norm': '57.5', 'learning_rate': '1.006e-06', 'num_tokens': '3.451e+06', 'mean_token_accuracy': '0.8393', 'epoch': '0.8996'}
+{'loss': '1.588', 'grad_norm': '21.62', 'learning_rate': '1.004e-06', 'num_tokens': '3.452e+06', 'mean_token_accuracy': '0.6524', 'epoch': '0.8998'}
+{'loss': '1.05', 'grad_norm': '22.62', 'learning_rate': '1.002e-06', 'num_tokens': '3.452e+06', 'mean_token_accuracy': '0.7365', 'epoch': '0.9'}
+{'loss': '1.423', 'grad_norm': '14.81', 'learning_rate': '1e-06', 'num_tokens': '3.452e+06', 'mean_token_accuracy': '0.6105', 'epoch': '0.9001'}
+{'loss': '1.337', 'grad_norm': '5.531', 'learning_rate': '9.986e-07', 'num_tokens': '3.455e+06', 'mean_token_accuracy': '0.712', 'epoch': '0.9003'}
+{'loss': '1.296', 'grad_norm': '13.75', 'learning_rate': '9.969e-07', 'num_tokens': '3.456e+06', 'mean_token_accuracy': '0.7278', 'epoch': '0.9005'}
+{'loss': '1.248', 'grad_norm': '13.06', 'learning_rate': '9.951e-07', 'num_tokens': '3.456e+06', 'mean_token_accuracy': '0.6451', 'epoch': '0.9007'}
+{'loss': '2.061', 'grad_norm': '13.94', 'learning_rate': '9.934e-07', 'num_tokens': '3.457e+06', 'mean_token_accuracy': '0.6032', 'epoch': '0.9008'}
+{'loss': '1.456', 'grad_norm': '13.44', 'learning_rate': '9.917e-07', 'num_tokens': '3.458e+06', 'mean_token_accuracy': '0.6282', 'epoch': '0.901'}
+{'loss': '2.19', 'grad_norm': '25.5', 'learning_rate': '9.899e-07', 'num_tokens': '3.458e+06', 'mean_token_accuracy': '0.504', 'epoch': '0.9012'}
+{'loss': '1.747', 'grad_norm': '7.719', 'learning_rate': '9.882e-07', 'num_tokens': '3.46e+06', 'mean_token_accuracy': '0.6372', 'epoch': '0.9014'}
+{'loss': '1.985', 'grad_norm': '18.88', 'learning_rate': '9.865e-07', 'num_tokens': '3.461e+06', 'mean_token_accuracy': '0.605', 'epoch': '0.9015'}
+{'loss': '1.674', 'grad_norm': '22', 'learning_rate': '9.847e-07', 'num_tokens': '3.461e+06', 'mean_token_accuracy': '0.6561', 'epoch': '0.9017'}
+{'loss': '1.296', 'grad_norm': '13.69', 'learning_rate': '9.83e-07', 'num_tokens': '3.461e+06', 'mean_token_accuracy': '0.7278', 'epoch': '0.9019'}
+{'loss': '0.9714', 'grad_norm': '11.38', 'learning_rate': '9.812e-07', 'num_tokens': '3.462e+06', 'mean_token_accuracy': '0.7539', 'epoch': '0.902'}
+{'loss': '1.637', 'grad_norm': '26.5', 'learning_rate': '9.795e-07', 'num_tokens': '3.462e+06', 'mean_token_accuracy': '0.6471', 'epoch': '0.9022'}
+{'loss': '2.046', 'grad_norm': '22', 'learning_rate': '9.778e-07', 'num_tokens': '3.463e+06', 'mean_token_accuracy': '0.5522', 'epoch': '0.9024'}
+{'loss': '1.068', 'grad_norm': '12.19', 'learning_rate': '9.76e-07', 'num_tokens': '3.463e+06', 'mean_token_accuracy': '0.6907', 'epoch': '0.9026'}
+{'loss': '1.181', 'grad_norm': '12.19', 'learning_rate': '9.743e-07', 'num_tokens': '3.464e+06', 'mean_token_accuracy': '0.6869', 'epoch': '0.9027'}
+{'loss': '1.651', 'grad_norm': '10.88', 'learning_rate': '9.726e-07', 'num_tokens': '3.465e+06', 'mean_token_accuracy': '0.6332', 'epoch': '0.9029'}
+{'loss': '0.6445', 'grad_norm': '7.969', 'learning_rate': '9.708e-07', 'num_tokens': '3.467e+06', 'mean_token_accuracy': '0.8036', 'epoch': '0.9031'}
+{'loss': '0.9317', 'grad_norm': '35', 'learning_rate': '9.691e-07', 'num_tokens': '3.467e+06', 'mean_token_accuracy': '0.8145', 'epoch': '0.9033'}
+{'loss': '0.637', 'grad_norm': '30.62', 'learning_rate': '9.673e-07', 'num_tokens': '3.467e+06', 'mean_token_accuracy': '0.8485', 'epoch': '0.9034'}
+{'loss': '1.416', 'grad_norm': '19.38', 'learning_rate': '9.656e-07', 'num_tokens': '3.467e+06', 'mean_token_accuracy': '0.6792', 'epoch': '0.9036'}
+{'loss': '1.207', 'grad_norm': '6.25', 'learning_rate': '9.639e-07', 'num_tokens': '3.47e+06', 'mean_token_accuracy': '0.7148', 'epoch': '0.9038'}
+{'loss': '1.073', 'grad_norm': '28.62', 'learning_rate': '9.621e-07', 'num_tokens': '3.47e+06', 'mean_token_accuracy': '0.7746', 'epoch': '0.904'}
+{'loss': '2.198', 'grad_norm': '17.38', 'learning_rate': '9.604e-07', 'num_tokens': '3.471e+06', 'mean_token_accuracy': '0.5124', 'epoch': '0.9041'}
+{'loss': '2.038', 'grad_norm': '20.62', 'learning_rate': '9.587e-07', 'num_tokens': '3.471e+06', 'mean_token_accuracy': '0.5882', 'epoch': '0.9043'}
+{'loss': '0.8048', 'grad_norm': '24.5', 'learning_rate': '9.569e-07', 'num_tokens': '3.471e+06', 'mean_token_accuracy': '0.773', 'epoch': '0.9045'}
+{'loss': '2.233', 'grad_norm': '20', 'learning_rate': '9.552e-07', 'num_tokens': '3.471e+06', 'mean_token_accuracy': '0.621', 'epoch': '0.9047'}
+{'loss': '1.997', 'grad_norm': '25', 'learning_rate': '9.535e-07', 'num_tokens': '3.472e+06', 'mean_token_accuracy': '0.5877', 'epoch': '0.9048'}
+{'loss': '1.041', 'grad_norm': '18.62', 'learning_rate': '9.517e-07', 'num_tokens': '3.472e+06', 'mean_token_accuracy': '0.7782', 'epoch': '0.905'}
+{'loss': '1.11', 'grad_norm': '14.5', 'learning_rate': '9.5e-07', 'num_tokens': '3.473e+06', 'mean_token_accuracy': '0.6995', 'epoch': '0.9052'}
+{'loss': '1.975', 'grad_norm': '40.5', 'learning_rate': '9.482e-07', 'num_tokens': '3.473e+06', 'mean_token_accuracy': '0.6301', 'epoch': '0.9053'}
+{'loss': '0.9121', 'grad_norm': '17.5', 'learning_rate': '9.465e-07', 'num_tokens': '3.473e+06', 'mean_token_accuracy': '0.7974', 'epoch': '0.9055'}
+{'loss': '0.9346', 'grad_norm': '9.125', 'learning_rate': '9.448e-07', 'num_tokens': '3.474e+06', 'mean_token_accuracy': '0.7743', 'epoch': '0.9057'}
+{'loss': '1.112', 'grad_norm': '15.12', 'learning_rate': '9.43e-07', 'num_tokens': '3.475e+06', 'mean_token_accuracy': '0.7074', 'epoch': '0.9059'}
+{'loss': '2.108', 'grad_norm': '42', 'learning_rate': '9.413e-07', 'num_tokens': '3.475e+06', 'mean_token_accuracy': '0.5942', 'epoch': '0.906'}
+{'loss': '0.4806', 'grad_norm': '6.938', 'learning_rate': '9.396e-07', 'num_tokens': '3.476e+06', 'mean_token_accuracy': '0.8523', 'epoch': '0.9062'}
+{'loss': '0.6415', 'grad_norm': '21.88', 'learning_rate': '9.378e-07', 'num_tokens': '3.476e+06', 'mean_token_accuracy': '0.8511', 'epoch': '0.9064'}
+{'loss': '1.579', 'grad_norm': '16.12', 'learning_rate': '9.361e-07', 'num_tokens': '3.477e+06', 'mean_token_accuracy': '0.6627', 'epoch': '0.9066'}
+
 91%|█████████ | 5221/5758 [29:44<03:33,  2.51it/s]
 91%|█████████ | 5222/5758 [29:45<03:18,  2.70it/s]
                                                   

 91%|█████████ | 5222/5758 [29:45<03:18,  2.70it/s]
 91%|█████████ | 5223/5758 [29:45<03:17,  2.71it/s]
                                                   

 91%|█████████ | 5223/5758 [29:45<03:17,  2.71it/s]
 91%|█████████ | 5224/5758 [29:45<03:05,  2.88it/s]
                                                   

 91%|█████████ | 5224/5758 [29:45<03:05,  2.88it/s]
 91%|█████████ | 5225/5758 [29:46<03:03,  2.91it/s]
                                                   

 91%|█████████ | 5225/5758 [29:46<03:03,  2.91it/s]
 91%|█████████ | 5226/5758 [29:46<03:00,  2.94it/s]
                                                   

 91%|█████████ | 5226/5758 [29:46<03:00,  2.94it/s]
 91%|█████████ | 5227/5758 [29:46<03:00,  2.94it/s]
                                                   

 91%|█████████ | 5227/5758 [29:46<03:00,  2.94it/s]
 91%|█████████ | 5228/5758 [29:47<02:52,  3.07it/s]
                                                   

 91%|█████████ | 5228/5758 [29:47<02:52,  3.07it/s]
 91%|█████████ | 5229/5758 [29:47<02:54,  3.02it/s]
                                                   

 91%|█████████ | 5229/5758 [29:47<02:54,  3.02it/s]
 91%|█████████ | 5230/5758 [29:47<03:04,  2.86it/s]
                                                   

 91%|█████████ | 5230/5758 [29:47<03:04,  2.86it/s]
 91%|█████████ | 5231/5758 [29:48<03:00,  2.92it/s]
                                                   

 91%|█████████ | 5231/5758 [29:48<03:00,  2.92it/s]
 91%|█████████ | 5232/5758 [29:48<02:51,  3.06it/s]
                                                   

 91%|█████████ | 5232/5758 [29:48<02:51,  3.06it/s]
 91%|█████████ | 5233/5758 [29:48<02:45,  3.17it/s]
                                                   

 91%|█████████ | 5233/5758 [29:48<02:45,  3.17it/s]
 91%|█████████ | 5234/5758 [29:49<02:42,  3.21it/s]
                                                   

 91%|█████████ | 5234/5758 [29:49<02:42,  3.21it/s]
 91%|█████████ | 5235/5758 [29:49<02:44,  3.19it/s]
                                                   

 91%|█████████ | 5235/5758 [29:49<02:44,  3.19it/s]
 91%|█████████ | 5236/5758 [29:49<02:44,  3.18it/s]
                                                   

 91%|█████████ | 5236/5758 [29:49<02:44,  3.18it/s]
 91%|█████████ | 5237/5758 [29:50<02:44,  3.17it/s]
                                                   

 91%|█████████ | 5237/5758 [29:50<02:44,  3.17it/s]
 91%|█████████ | 5238/5758 [29:50<02:41,  3.23it/s]
                                                   

 91%|█████████ | 5238/5758 [29:50<02:41,  3.23it/s]
 91%|█████████ | 5239/5758 [29:50<02:39,  3.25it/s]
                                                   

 91%|█████████ | 5239/5758 [29:50<02:39,  3.25it/s]
 91%|█████████ | 5240/5758 [29:51<02:43,  3.17it/s]
                                                   

 91%|█████████ | 5240/5758 [29:51<02:43,  3.17it/s]
 91%|█████████ | 5241/5758 [29:51<02:39,  3.24it/s]
                                                   

 91%|█████████ | 5241/5758 [29:51<02:39,  3.24it/s]
 91%|█████████ | 5242/5758 [29:51<02:47,  3.08it/s]
                                                   

 91%|█████████ | 5242/5758 [29:51<02:47,  3.08it/s]
 91%|█████████ | 5243/5758 [29:51<02:46,  3.10it/s]
                                                   

 91%|█████████ | 5243/5758 [29:51<02:46,  3.10it/s]
 91%|█████████ | 5244/5758 [29:52<02:44,  3.13it/s]
                                                   

 91%|█████████ | 5244/5758 [29:52<02:44,  3.13it/s]
 91%|█████████ | 5245/5758 [29:52<02:40,  3.21it/s]
                                                   

 91%|█████████ | 5245/5758 [29:52<02:40,  3.21it/s]
 91%|█████████ | 5246/5758 [29:52<02:52,  2.97it/s]
                                                   

 91%|█████████ | 5246/5758 [29:52<02:52,  2.97it/s]
 91%|█████████ | 5247/5758 [29:53<02:52,  2.96it/s]
                                                   

 91%|█████████ | 5247/5758 [29:53<02:52,  2.96it/s]
 91%|█████████ | 5248/5758 [29:53<02:53,  2.94it/s]
                                                   

 91%|█████████ | 5248/5758 [29:53<02:53,  2.94it/s]
 91%|█████████ | 5249/5758 [29:54<03:14,  2.62it/s]
                                                   

 91%|█████████ | 5249/5758 [29:54<03:14,  2.62it/s]
 91%|█████████ | 5250/5758 [29:54<03:00,  2.81it/s]
                                                   

 91%|█████████ | 5250/5758 [29:54<03:00,  2.81it/s]
 91%|█████████ | 5251/5758 [29:54<02:57,  2.86it/s]
                                                   

 91%|█████████ | 5251/5758 [29:54<02:57,  2.86it/s]
 91%|█████████ | 5252/5758 [29:55<02:50,  2.98it/s]
                                                   

 91%|█████████ | 5252/5758 [29:55<02:50,  2.98it/s]
 91%|█████████ | 5253/5758 [29:55<02:38,  3.18it/s]
                                                   

 91%|█████████ | 5253/5758 [29:55<02:38,  3.18it/s]
 91%|█████████ | 5254/5758 [29:55<02:39,  3.16it/s]
                                                   

 91%|█████████ | 5254/5758 [29:55<02:39,  3.16it/s]
 91%|█████████▏| 5255/5758 [29:56<02:43,  3.08it/s]
                                                   

 91%|█████████▏| 5255/5758 [29:56<02:43,  3.08it/s]
 91%|█████████▏| 5256/5758 [29:56<02:42,  3.09it/s]
                                                   

 91%|█████████▏| 5256/5758 [29:56<02:42,  3.09it/s]
 91%|█████████▏| 5257/5758 [29:56<02:40,  3.13it/s]
                                                   

 91%|█████████▏| 5257/5758 [29:56<02:40,  3.13it/s]
 91%|█████████▏| 5258/5758 [29:57<02:49,  2.95it/s]
                                                   

 91%|█████████▏| 5258/5758 [29:57<02:49,  2.95it/s]
 91%|█████████▏| 5259/5758 [29:57<02:55,  2.85it/s]
                                                   

 91%|█████████▏| 5259/5758 [29:57<02:55,  2.85it/s]
 91%|█████████▏| 5260/5758 [29:57<02:47,  2.97it/s]
                                                   

 91%|█████████▏| 5260/5758 [29:57<02:47,  2.97it/s]
 91%|█████████▏| 5261/5758 [29:58<02:43,  3.04it/s]
                                                   

 91%|█████████▏| 5261/5758 [29:58<02:43,  3.04it/s]
 91%|█████████▏| 5262/5758 [29:58<02:38,  3.13it/s]
                                                   

 91%|█████████▏| 5262/5758 [29:58<02:38,  3.13it/s]
 91%|█████████▏| 5263/5758 [29:58<02:37,  3.15it/s]
                                                   

 91%|█████████▏| 5263/5758 [29:58<02:37,  3.15it/s]
 91%|█████████▏| 5264/5758 [29:59<02:47,  2.95it/s]
                                                   

 91%|█████████▏| 5264/5758 [29:59<02:47,  2.95it/s]
 91%|█████████▏| 5265/5758 [29:59<02:44,  3.00it/s]
                                                   

 91%|█████████▏| 5265/5758 [29:59<02:44,  3.00it/s]
 91%|█████████▏| 5266/5758 [29:59<02:48,  2.92it/s]
                                                   

 91%|█████████▏| 5266/5758 [29:59<02:48,  2.92it/s]
 91%|█████████▏| 5267/5758 [29:59<02:39,  3.07it/s]
                                                   

 91%|█████████▏| 5267/5758 [29:59<02:39,  3.07it/s]
 91%|█████████▏| 5268/5758 [30:00<02:36,  3.14it/s]
                                                   

 91%|█████████▏| 5268/5758 [30:00<02:36,  3.14it/s]
 92%|█████████▏| 5269/5758 [30:00<02:37,  3.10it/s]
                                                   

 92%|█████████▏| 5269/5758 [30:00<02:37,  3.10it/s]
 92%|█████████▏| 5270/5758 [30:00<02:33,  3.17it/s]
                                                   

 92%|█████████▏| 5270/5758 [30:00<02:33,  3.17it/s]
 92%|█████████▏| 5271/5758 [30:01<02:30,  3.23it/s]
                                                   

 92%|█████████▏| 5271/5758 [30:01<02:30,  3.23it/s]
 92%|█████████▏| 5272/5758 [30:01<02:41,  3.00it/s]
                                                   

 92%|█████████▏| 5272/5758 [30:01<02:41,  3.00it/s]
 92%|█████████▏| 5273/5758 [30:01<02:47,  2.89it/s]
                                                   

 92%|█████████▏| 5273/5758 [30:01<02:47,  2.89it/s]
 92%|█████████▏| 5274/5758 [30:02<02:56,  2.74it/s]
                                                   

 92%|█████████▏| 5274/5758 [30:02<02:56,  2.74it/s]
 92%|█████████▏| 5275/5758 [30:02<02:59,  2.69it/s]
                                                   

 92%|█████████▏| 5275/5758 [30:02<02:59,  2.69it/s]
 92%|█████████▏| 5276/5758 [30:03<02:56,  2.74it/s]
                                                   
{'loss': '2.204', 'grad_norm': '17.75', 'learning_rate': '9.344e-07', 'num_tokens': '3.477e+06', 'mean_token_accuracy': '0.5801', 'epoch': '0.9067'}
+{'loss': '0.7004', 'grad_norm': '9.562', 'learning_rate': '9.326e-07', 'num_tokens': '3.479e+06', 'mean_token_accuracy': '0.7819', 'epoch': '0.9069'}
+{'loss': '1.154', 'grad_norm': '11.12', 'learning_rate': '9.309e-07', 'num_tokens': '3.48e+06', 'mean_token_accuracy': '0.6997', 'epoch': '0.9071'}
+{'loss': '1.703', 'grad_norm': '12.06', 'learning_rate': '9.291e-07', 'num_tokens': '3.48e+06', 'mean_token_accuracy': '0.6305', 'epoch': '0.9073'}
+{'loss': '0.8108', 'grad_norm': '30.12', 'learning_rate': '9.274e-07', 'num_tokens': '3.48e+06', 'mean_token_accuracy': '0.7857', 'epoch': '0.9074'}
+{'loss': '0.398', 'grad_norm': '19.5', 'learning_rate': '9.257e-07', 'num_tokens': '3.481e+06', 'mean_token_accuracy': '0.9005', 'epoch': '0.9076'}
+{'loss': '0.6035', 'grad_norm': '30.38', 'learning_rate': '9.239e-07', 'num_tokens': '3.481e+06', 'mean_token_accuracy': '0.8485', 'epoch': '0.9078'}
+{'loss': '0.6788', 'grad_norm': '9.188', 'learning_rate': '9.222e-07', 'num_tokens': '3.482e+06', 'mean_token_accuracy': '0.7831', 'epoch': '0.908'}
+{'loss': '1.583', 'grad_norm': '12.56', 'learning_rate': '9.205e-07', 'num_tokens': '3.483e+06', 'mean_token_accuracy': '0.6582', 'epoch': '0.9081'}
+{'loss': '1.749', 'grad_norm': '12.94', 'learning_rate': '9.187e-07', 'num_tokens': '3.484e+06', 'mean_token_accuracy': '0.6639', 'epoch': '0.9083'}
+{'loss': '1.309', 'grad_norm': '12.06', 'learning_rate': '9.17e-07', 'num_tokens': '3.484e+06', 'mean_token_accuracy': '0.6538', 'epoch': '0.9085'}
+{'loss': '1.198', 'grad_norm': '13.06', 'learning_rate': '9.152e-07', 'num_tokens': '3.485e+06', 'mean_token_accuracy': '0.6942', 'epoch': '0.9086'}
+{'loss': '1.06', 'grad_norm': '27.25', 'learning_rate': '9.135e-07', 'num_tokens': '3.485e+06', 'mean_token_accuracy': '0.7368', 'epoch': '0.9088'}
+{'loss': '1.701', 'grad_norm': '49.75', 'learning_rate': '9.118e-07', 'num_tokens': '3.485e+06', 'mean_token_accuracy': '0.6625', 'epoch': '0.909'}
+{'loss': '0.9408', 'grad_norm': '12.5', 'learning_rate': '9.1e-07', 'num_tokens': '3.486e+06', 'mean_token_accuracy': '0.7106', 'epoch': '0.9092'}
+{'loss': '1.185', 'grad_norm': '11.44', 'learning_rate': '9.083e-07', 'num_tokens': '3.487e+06', 'mean_token_accuracy': '0.6727', 'epoch': '0.9093'}
+{'loss': '0.6478', 'grad_norm': '14.44', 'learning_rate': '9.066e-07', 'num_tokens': '3.487e+06', 'mean_token_accuracy': '0.8372', 'epoch': '0.9095'}
+{'loss': '1.224', 'grad_norm': '15.31', 'learning_rate': '9.048e-07', 'num_tokens': '3.487e+06', 'mean_token_accuracy': '0.7433', 'epoch': '0.9097'}
+{'loss': '1.542', 'grad_norm': '13.62', 'learning_rate': '9.031e-07', 'num_tokens': '3.488e+06', 'mean_token_accuracy': '0.5964', 'epoch': '0.9099'}
+{'loss': '1.153', 'grad_norm': '11.25', 'learning_rate': '9.014e-07', 'num_tokens': '3.489e+06', 'mean_token_accuracy': '0.6763', 'epoch': '0.91'}
+{'loss': '1.739', 'grad_norm': '14.81', 'learning_rate': '8.996e-07', 'num_tokens': '3.49e+06', 'mean_token_accuracy': '0.6579', 'epoch': '0.9102'}
+{'loss': '1.408', 'grad_norm': '4.875', 'learning_rate': '8.979e-07', 'num_tokens': '3.494e+06', 'mean_token_accuracy': '0.6618', 'epoch': '0.9104'}
+{'loss': '1.472', 'grad_norm': '14.88', 'learning_rate': '8.961e-07', 'num_tokens': '3.494e+06', 'mean_token_accuracy': '0.6959', 'epoch': '0.9106'}
+{'loss': '1.983', 'grad_norm': '16.12', 'learning_rate': '8.944e-07', 'num_tokens': '3.495e+06', 'mean_token_accuracy': '0.5695', 'epoch': '0.9107'}
+{'loss': '0.8219', 'grad_norm': '12.06', 'learning_rate': '8.927e-07', 'num_tokens': '3.495e+06', 'mean_token_accuracy': '0.7941', 'epoch': '0.9109'}
+{'loss': '0.5184', 'grad_norm': '25.88', 'learning_rate': '8.909e-07', 'num_tokens': '3.495e+06', 'mean_token_accuracy': '0.8686', 'epoch': '0.9111'}
+{'loss': '1.289', 'grad_norm': '16.25', 'learning_rate': '8.892e-07', 'num_tokens': '3.496e+06', 'mean_token_accuracy': '0.6565', 'epoch': '0.9113'}
+{'loss': '0.6558', 'grad_norm': '12.44', 'learning_rate': '8.875e-07', 'num_tokens': '3.496e+06', 'mean_token_accuracy': '0.8467', 'epoch': '0.9114'}
+{'loss': '1.707', 'grad_norm': '34', 'learning_rate': '8.857e-07', 'num_tokens': '3.496e+06', 'mean_token_accuracy': '0.6759', 'epoch': '0.9116'}
+{'loss': '0.9533', 'grad_norm': '12.44', 'learning_rate': '8.84e-07', 'num_tokens': '3.497e+06', 'mean_token_accuracy': '0.7724', 'epoch': '0.9118'}
+{'loss': '2.001', 'grad_norm': '17.25', 'learning_rate': '8.823e-07', 'num_tokens': '3.497e+06', 'mean_token_accuracy': '0.6005', 'epoch': '0.9119'}
+{'loss': '1.263', 'grad_norm': '11.81', 'learning_rate': '8.805e-07', 'num_tokens': '3.498e+06', 'mean_token_accuracy': '0.6584', 'epoch': '0.9121'}
+{'loss': '0.7205', 'grad_norm': '35.5', 'learning_rate': '8.788e-07', 'num_tokens': '3.498e+06', 'mean_token_accuracy': '0.8704', 'epoch': '0.9123'}
+{'loss': '0.6389', 'grad_norm': '13.62', 'learning_rate': '8.77e-07', 'num_tokens': '3.499e+06', 'mean_token_accuracy': '0.8145', 'epoch': '0.9125'}
+{'loss': '0.7173', 'grad_norm': '32', 'learning_rate': '8.753e-07', 'num_tokens': '3.499e+06', 'mean_token_accuracy': '0.8571', 'epoch': '0.9126'}
+{'loss': '1.088', 'grad_norm': '11.25', 'learning_rate': '8.736e-07', 'num_tokens': '3.499e+06', 'mean_token_accuracy': '0.6958', 'epoch': '0.9128'}
+{'loss': '2.078', 'grad_norm': '12.94', 'learning_rate': '8.718e-07', 'num_tokens': '3.5e+06', 'mean_token_accuracy': '0.5759', 'epoch': '0.913'}
+{'loss': '1.144', 'grad_norm': '14.38', 'learning_rate': '8.701e-07', 'num_tokens': '3.501e+06', 'mean_token_accuracy': '0.6893', 'epoch': '0.9132'}
+{'loss': '1.294', 'grad_norm': '13.5', 'learning_rate': '8.684e-07', 'num_tokens': '3.501e+06', 'mean_token_accuracy': '0.6944', 'epoch': '0.9133'}
+{'loss': '0.604', 'grad_norm': '17.62', 'learning_rate': '8.666e-07', 'num_tokens': '3.502e+06', 'mean_token_accuracy': '0.8545', 'epoch': '0.9135'}
+{'loss': '0.8822', 'grad_norm': '14.56', 'learning_rate': '8.649e-07', 'num_tokens': '3.502e+06', 'mean_token_accuracy': '0.7772', 'epoch': '0.9137'}
+{'loss': '1.916', 'grad_norm': '17.62', 'learning_rate': '8.631e-07', 'num_tokens': '3.503e+06', 'mean_token_accuracy': '0.5908', 'epoch': '0.9139'}
+{'loss': '1.915', 'grad_norm': '16.88', 'learning_rate': '8.614e-07', 'num_tokens': '3.503e+06', 'mean_token_accuracy': '0.6158', 'epoch': '0.914'}
+{'loss': '2.004', 'grad_norm': '14', 'learning_rate': '8.597e-07', 'num_tokens': '3.504e+06', 'mean_token_accuracy': '0.599', 'epoch': '0.9142'}
+{'loss': '1.352', 'grad_norm': '18', 'learning_rate': '8.579e-07', 'num_tokens': '3.504e+06', 'mean_token_accuracy': '0.7368', 'epoch': '0.9144'}
+{'loss': '1.135', 'grad_norm': '19.62', 'learning_rate': '8.562e-07', 'num_tokens': '3.504e+06', 'mean_token_accuracy': '0.7615', 'epoch': '0.9146'}
+{'loss': '0.599', 'grad_norm': '49.75', 'learning_rate': '8.545e-07', 'num_tokens': '3.504e+06', 'mean_token_accuracy': '0.8485', 'epoch': '0.9147'}
+{'loss': '1.966', 'grad_norm': '17', 'learning_rate': '8.527e-07', 'num_tokens': '3.505e+06', 'mean_token_accuracy': '0.5807', 'epoch': '0.9149'}
+{'loss': '0.649', 'grad_norm': '11.19', 'learning_rate': '8.51e-07', 'num_tokens': '3.505e+06', 'mean_token_accuracy': '0.8592', 'epoch': '0.9151'}
+{'loss': '1.169', 'grad_norm': '12.69', 'learning_rate': '8.493e-07', 'num_tokens': '3.506e+06', 'mean_token_accuracy': '0.6772', 'epoch': '0.9152'}
+{'loss': '1.566', 'grad_norm': '41.75', 'learning_rate': '8.475e-07', 'num_tokens': '3.506e+06', 'mean_token_accuracy': '0.72', 'epoch': '0.9154'}
+{'loss': '1.008', 'grad_norm': '15.56', 'learning_rate': '8.458e-07', 'num_tokens': '3.506e+06', 'mean_token_accuracy': '0.7705', 'epoch': '0.9156'}
+{'loss': '1.16', 'grad_norm': '19.25', 'learning_rate': '8.44e-07', 'num_tokens': '3.507e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.9158'}
+{'loss': '1.58', 'grad_norm': '6.812', 'learning_rate': '8.423e-07', 'num_tokens': '3.509e+06', 'mean_token_accuracy': '0.6436', 'epoch': '0.9159'}
+{'loss': '2.034', 'grad_norm': '33', 'learning_rate': '8.406e-07', 'num_tokens': '3.509e+06', 'mean_token_accuracy': '0.5891', 'epoch': '0.9161'}
+
 92%|█████████▏| 5276/5758 [30:03<02:56,  2.74it/s]
 92%|█████████▏| 5277/5758 [30:03<02:44,  2.93it/s]
                                                   

 92%|█████████▏| 5277/5758 [30:03<02:44,  2.93it/s]
 92%|█████████▏| 5278/5758 [30:03<02:37,  3.05it/s]
                                                   

 92%|█████████▏| 5278/5758 [30:03<02:37,  3.05it/s]
 92%|█████████▏| 5279/5758 [30:04<02:32,  3.14it/s]
                                                   

 92%|█████████▏| 5279/5758 [30:04<02:32,  3.14it/s]
 92%|█████████▏| 5280/5758 [30:04<02:34,  3.09it/s]
                                                   

 92%|█████████▏| 5280/5758 [30:04<02:34,  3.09it/s]
 92%|█████████▏| 5281/5758 [30:04<02:35,  3.08it/s]
                                                   

 92%|█████████▏| 5281/5758 [30:04<02:35,  3.08it/s]
 92%|█████████▏| 5282/5758 [30:04<02:31,  3.15it/s]
                                                   

 92%|█████████▏| 5282/5758 [30:04<02:31,  3.15it/s]
 92%|█████████▏| 5283/5758 [30:05<02:39,  2.97it/s]
                                                   

 92%|█████████▏| 5283/5758 [30:05<02:39,  2.97it/s]
 92%|█████████▏| 5284/5758 [30:05<02:35,  3.05it/s]
                                                   

 92%|█████████▏| 5284/5758 [30:05<02:35,  3.05it/s]
 92%|█████████▏| 5285/5758 [30:05<02:29,  3.16it/s]
                                                   

 92%|█████████▏| 5285/5758 [30:05<02:29,  3.16it/s]
 92%|█████████▏| 5286/5758 [30:06<02:35,  3.03it/s]
                                                   

 92%|█████████▏| 5286/5758 [30:06<02:35,  3.03it/s]
 92%|█████████▏| 5287/5758 [30:06<02:30,  3.13it/s]
                                                   

 92%|█████████▏| 5287/5758 [30:06<02:30,  3.13it/s]
 92%|█████████▏| 5288/5758 [30:06<02:35,  3.02it/s]
                                                   

 92%|█████████▏| 5288/5758 [30:06<02:35,  3.02it/s]
 92%|█████████▏| 5289/5758 [30:07<02:32,  3.07it/s]
                                                   

 92%|█████████▏| 5289/5758 [30:07<02:32,  3.07it/s]
 92%|█████████▏| 5290/5758 [30:07<02:27,  3.17it/s]
                                                   

 92%|█████████▏| 5290/5758 [30:07<02:27,  3.17it/s]
 92%|█████████▏| 5291/5758 [30:07<02:33,  3.04it/s]
                                                   

 92%|█████████▏| 5291/5758 [30:07<02:33,  3.04it/s]
 92%|█████████▏| 5292/5758 [30:08<02:30,  3.10it/s]
                                                   

 92%|█████████▏| 5292/5758 [30:08<02:30,  3.10it/s]
 92%|█████████▏| 5293/5758 [30:08<02:26,  3.17it/s]
                                                   

 92%|█████████▏| 5293/5758 [30:08<02:26,  3.17it/s]
 92%|█████████▏| 5294/5758 [30:08<02:26,  3.17it/s]
                                                   

 92%|█████████▏| 5294/5758 [30:08<02:26,  3.17it/s]
 92%|█████████▏| 5295/5758 [30:09<02:29,  3.11it/s]
                                                   

 92%|█████████▏| 5295/5758 [30:09<02:29,  3.11it/s]
 92%|█████████▏| 5296/5758 [30:09<02:27,  3.13it/s]
                                                   

 92%|█████████▏| 5296/5758 [30:09<02:27,  3.13it/s]
 92%|█████████▏| 5297/5758 [30:09<02:24,  3.19it/s]
                                                   

 92%|█████████▏| 5297/5758 [30:09<02:24,  3.19it/s]
 92%|█████████▏| 5298/5758 [30:10<02:21,  3.24it/s]
                                                   

 92%|█████████▏| 5298/5758 [30:10<02:21,  3.24it/s]
 92%|█████████▏| 5299/5758 [30:10<02:21,  3.23it/s]
                                                   

 92%|█████████▏| 5299/5758 [30:10<02:21,  3.23it/s]
 92%|█████████▏| 5300/5758 [30:10<02:48,  2.73it/s]
                                                   

 92%|█████████▏| 5300/5758 [30:10<02:48,  2.73it/s]
 92%|█████████▏| 5301/5758 [30:11<02:40,  2.85it/s]
                                                   

 92%|█████████▏| 5301/5758 [30:11<02:40,  2.85it/s]
 92%|█████████▏| 5302/5758 [30:11<02:32,  2.99it/s]
                                                   

 92%|█████████▏| 5302/5758 [30:11<02:32,  2.99it/s]
 92%|█████████▏| 5303/5758 [30:11<02:36,  2.92it/s]
                                                   

 92%|█████████▏| 5303/5758 [30:11<02:36,  2.92it/s]
 92%|█████████▏| 5304/5758 [30:12<02:41,  2.81it/s]
                                                   

 92%|█████████▏| 5304/5758 [30:12<02:41,  2.81it/s]
 92%|█████████▏| 5305/5758 [30:12<02:42,  2.78it/s]
                                                   

 92%|█████████▏| 5305/5758 [30:12<02:42,  2.78it/s]
 92%|█████████▏| 5306/5758 [30:13<02:43,  2.77it/s]
                                                   

 92%|█████████▏| 5306/5758 [30:13<02:43,  2.77it/s]
 92%|█████████▏| 5307/5758 [30:13<02:32,  2.95it/s]
                                                   

 92%|█████████▏| 5307/5758 [30:13<02:32,  2.95it/s]
 92%|█████████▏| 5308/5758 [30:13<02:37,  2.86it/s]
                                                   

 92%|█████████▏| 5308/5758 [30:13<02:37,  2.86it/s]
 92%|█████████▏| 5309/5758 [30:13<02:29,  3.01it/s]
                                                   

 92%|█████████▏| 5309/5758 [30:13<02:29,  3.01it/s]
 92%|█████████▏| 5310/5758 [30:14<02:28,  3.01it/s]
                                                   

 92%|█████████▏| 5310/5758 [30:14<02:28,  3.01it/s]
 92%|█████████▏| 5311/5758 [30:14<02:25,  3.06it/s]
                                                   

 92%|█████████▏| 5311/5758 [30:14<02:25,  3.06it/s]
 92%|█████████▏| 5312/5758 [30:14<02:22,  3.12it/s]
                                                   

 92%|█████████▏| 5312/5758 [30:14<02:22,  3.12it/s]
 92%|█████████▏| 5313/5758 [30:15<02:28,  2.99it/s]
                                                   

 92%|█████████▏| 5313/5758 [30:15<02:28,  2.99it/s]
 92%|█████████▏| 5314/5758 [30:15<02:24,  3.07it/s]
                                                   

 92%|█████████▏| 5314/5758 [30:15<02:24,  3.07it/s]
 92%|█████████▏| 5315/5758 [30:15<02:20,  3.16it/s]
                                                   

 92%|█████████▏| 5315/5758 [30:15<02:20,  3.16it/s]
 92%|█████████▏| 5316/5758 [30:16<02:22,  3.10it/s]
                                                   

 92%|█████████▏| 5316/5758 [30:16<02:22,  3.10it/s]
 92%|█████████▏| 5317/5758 [30:16<02:24,  3.04it/s]
                                                   

 92%|█████████▏| 5317/5758 [30:16<02:24,  3.04it/s]
 92%|█████████▏| 5318/5758 [30:16<02:25,  3.02it/s]
                                                   

 92%|█████████▏| 5318/5758 [30:16<02:25,  3.02it/s]
 92%|█████████▏| 5319/5758 [30:17<02:22,  3.08it/s]
                                                   

 92%|█████████▏| 5319/5758 [30:17<02:22,  3.08it/s]
 92%|█████████▏| 5320/5758 [30:17<02:29,  2.92it/s]
                                                   

 92%|█████████▏| 5320/5758 [30:17<02:29,  2.92it/s]
 92%|█████████▏| 5321/5758 [30:17<02:28,  2.94it/s]
                                                   

 92%|█████████▏| 5321/5758 [30:17<02:28,  2.94it/s]
 92%|█████████▏| 5322/5758 [30:18<02:23,  3.04it/s]
                                                   

 92%|█████████▏| 5322/5758 [30:18<02:23,  3.04it/s]
 92%|█████████▏| 5323/5758 [30:18<02:19,  3.12it/s]
                                                   

 92%|█████████▏| 5323/5758 [30:18<02:19,  3.12it/s]
 92%|█████████▏| 5324/5758 [30:18<02:26,  2.96it/s]
                                                   

 92%|█████████▏| 5324/5758 [30:18<02:26,  2.96it/s]
 92%|█████████▏| 5325/5758 [30:19<02:21,  3.05it/s]
                                                   

 92%|█████████▏| 5325/5758 [30:19<02:21,  3.05it/s]
 92%|█████████▏| 5326/5758 [30:19<02:31,  2.84it/s]
                                                   

 92%|█████████▏| 5326/5758 [30:19<02:31,  2.84it/s]
 93%|█████████▎| 5327/5758 [30:19<02:26,  2.95it/s]
                                                   

 93%|█████████▎| 5327/5758 [30:19<02:26,  2.95it/s]
 93%|█████████▎| 5328/5758 [30:20<02:23,  2.99it/s]
                                                   

 93%|█████████▎| 5328/5758 [30:20<02:23,  2.99it/s]
 93%|█████████▎| 5329/5758 [30:20<02:20,  3.05it/s]
                                                   

 93%|█████████▎| 5329/5758 [30:20<02:20,  3.05it/s]
 93%|█████████▎| 5330/5758 [30:20<02:16,  3.15it/s]
                                                   

 93%|█████████▎| 5330/5758 [30:20<02:16,  3.15it/s]
 93%|█████████▎| 5331/5758 [30:21<02:20,  3.04it/s]
                                                   
{'loss': '1.187', 'grad_norm': '12.38', 'learning_rate': '8.388e-07', 'num_tokens': '3.51e+06', 'mean_token_accuracy': '0.672', 'epoch': '0.9163'}
+{'loss': '0.9282', 'grad_norm': '36.25', 'learning_rate': '8.371e-07', 'num_tokens': '3.51e+06', 'mean_token_accuracy': '0.7895', 'epoch': '0.9165'}
+{'loss': '1.9', 'grad_norm': '14.12', 'learning_rate': '8.354e-07', 'num_tokens': '3.511e+06', 'mean_token_accuracy': '0.6027', 'epoch': '0.9166'}
+{'loss': '0.9642', 'grad_norm': '12.31', 'learning_rate': '8.336e-07', 'num_tokens': '3.512e+06', 'mean_token_accuracy': '0.73', 'epoch': '0.9168'}
+{'loss': '1.163', 'grad_norm': '10.88', 'learning_rate': '8.319e-07', 'num_tokens': '3.513e+06', 'mean_token_accuracy': '0.6885', 'epoch': '0.917'}
+{'loss': '1.382', 'grad_norm': '13.19', 'learning_rate': '8.301e-07', 'num_tokens': '3.513e+06', 'mean_token_accuracy': '0.6149', 'epoch': '0.9172'}
+{'loss': '0.7392', 'grad_norm': '8.625', 'learning_rate': '8.284e-07', 'num_tokens': '3.515e+06', 'mean_token_accuracy': '0.7776', 'epoch': '0.9173'}
+{'loss': '0.9062', 'grad_norm': '13.5', 'learning_rate': '8.267e-07', 'num_tokens': '3.515e+06', 'mean_token_accuracy': '0.7372', 'epoch': '0.9175'}
+{'loss': '1.077', 'grad_norm': '12.44', 'learning_rate': '8.249e-07', 'num_tokens': '3.516e+06', 'mean_token_accuracy': '0.6915', 'epoch': '0.9177'}
+{'loss': '0.6285', 'grad_norm': '8', 'learning_rate': '8.232e-07', 'num_tokens': '3.517e+06', 'mean_token_accuracy': '0.8078', 'epoch': '0.9179'}
+{'loss': '1.516', 'grad_norm': '19.12', 'learning_rate': '8.215e-07', 'num_tokens': '3.518e+06', 'mean_token_accuracy': '0.67', 'epoch': '0.918'}
+{'loss': '1.281', 'grad_norm': '13', 'learning_rate': '8.197e-07', 'num_tokens': '3.518e+06', 'mean_token_accuracy': '0.6577', 'epoch': '0.9182'}
+{'loss': '0.877', 'grad_norm': '12.81', 'learning_rate': '8.18e-07', 'num_tokens': '3.519e+06', 'mean_token_accuracy': '0.8039', 'epoch': '0.9184'}
+{'loss': '1.274', 'grad_norm': '12.44', 'learning_rate': '8.163e-07', 'num_tokens': '3.52e+06', 'mean_token_accuracy': '0.6569', 'epoch': '0.9185'}
+{'loss': '1.025', 'grad_norm': '17.12', 'learning_rate': '8.145e-07', 'num_tokens': '3.52e+06', 'mean_token_accuracy': '0.7675', 'epoch': '0.9187'}
+{'loss': '1.613', 'grad_norm': '18.75', 'learning_rate': '8.128e-07', 'num_tokens': '3.521e+06', 'mean_token_accuracy': '0.6542', 'epoch': '0.9189'}
+{'loss': '0.7859', 'grad_norm': '21.62', 'learning_rate': '8.11e-07', 'num_tokens': '3.521e+06', 'mean_token_accuracy': '0.7955', 'epoch': '0.9191'}
+{'loss': '0.9366', 'grad_norm': '12.12', 'learning_rate': '8.093e-07', 'num_tokens': '3.521e+06', 'mean_token_accuracy': '0.7266', 'epoch': '0.9192'}
+{'loss': '1.481', 'grad_norm': '12.19', 'learning_rate': '8.076e-07', 'num_tokens': '3.522e+06', 'mean_token_accuracy': '0.6847', 'epoch': '0.9194'}
+{'loss': '2.294', 'grad_norm': '12.75', 'learning_rate': '8.058e-07', 'num_tokens': '3.523e+06', 'mean_token_accuracy': '0.5394', 'epoch': '0.9196'}
+{'loss': '1.242', 'grad_norm': '13.69', 'learning_rate': '8.041e-07', 'num_tokens': '3.524e+06', 'mean_token_accuracy': '0.6582', 'epoch': '0.9198'}
+{'loss': '1.273', 'grad_norm': '14.25', 'learning_rate': '8.024e-07', 'num_tokens': '3.524e+06', 'mean_token_accuracy': '0.6701', 'epoch': '0.9199'}
+{'loss': '1.049', 'grad_norm': '13.25', 'learning_rate': '8.006e-07', 'num_tokens': '3.525e+06', 'mean_token_accuracy': '0.6979', 'epoch': '0.9201'}
+{'loss': '1.08', 'grad_norm': '11.56', 'learning_rate': '7.989e-07', 'num_tokens': '3.526e+06', 'mean_token_accuracy': '0.7021', 'epoch': '0.9203'}
+{'loss': '0.8533', 'grad_norm': '14.19', 'learning_rate': '7.972e-07', 'num_tokens': '3.526e+06', 'mean_token_accuracy': '0.7954', 'epoch': '0.9205'}
+{'loss': '1.288', 'grad_norm': '12.5', 'learning_rate': '7.954e-07', 'num_tokens': '3.527e+06', 'mean_token_accuracy': '0.6587', 'epoch': '0.9206'}
+{'loss': '1.082', 'grad_norm': '13.31', 'learning_rate': '7.937e-07', 'num_tokens': '3.528e+06', 'mean_token_accuracy': '0.6994', 'epoch': '0.9208'}
+{'loss': '0.8817', 'grad_norm': '17.5', 'learning_rate': '7.919e-07', 'num_tokens': '3.528e+06', 'mean_token_accuracy': '0.8031', 'epoch': '0.921'}
+{'loss': '1.092', 'grad_norm': '14.56', 'learning_rate': '7.902e-07', 'num_tokens': '3.529e+06', 'mean_token_accuracy': '0.6859', 'epoch': '0.9212'}
+{'loss': '1.833', 'grad_norm': '17.5', 'learning_rate': '7.885e-07', 'num_tokens': '3.529e+06', 'mean_token_accuracy': '0.6117', 'epoch': '0.9213'}
+{'loss': '1.769', 'grad_norm': '19.62', 'learning_rate': '7.867e-07', 'num_tokens': '3.53e+06', 'mean_token_accuracy': '0.6266', 'epoch': '0.9215'}
+{'loss': '0.5744', 'grad_norm': '8.062', 'learning_rate': '7.85e-07', 'num_tokens': '3.531e+06', 'mean_token_accuracy': '0.8164', 'epoch': '0.9217'}
+{'loss': '1.023', 'grad_norm': '25.25', 'learning_rate': '7.833e-07', 'num_tokens': '3.531e+06', 'mean_token_accuracy': '0.7597', 'epoch': '0.9218'}
+{'loss': '1.921', 'grad_norm': '13.81', 'learning_rate': '7.815e-07', 'num_tokens': '3.532e+06', 'mean_token_accuracy': '0.6065', 'epoch': '0.922'}
+{'loss': '0.9054', 'grad_norm': '14.31', 'learning_rate': '7.798e-07', 'num_tokens': '3.532e+06', 'mean_token_accuracy': '0.7801', 'epoch': '0.9222'}
+{'loss': '1.072', 'grad_norm': '12.12', 'learning_rate': '7.78e-07', 'num_tokens': '3.533e+06', 'mean_token_accuracy': '0.6968', 'epoch': '0.9224'}
+{'loss': '0.843', 'grad_norm': '18.75', 'learning_rate': '7.763e-07', 'num_tokens': '3.533e+06', 'mean_token_accuracy': '0.8194', 'epoch': '0.9225'}
+{'loss': '2.102', 'grad_norm': '15.75', 'learning_rate': '7.746e-07', 'num_tokens': '3.534e+06', 'mean_token_accuracy': '0.586', 'epoch': '0.9227'}
+{'loss': '0.5046', 'grad_norm': '6.562', 'learning_rate': '7.728e-07', 'num_tokens': '3.535e+06', 'mean_token_accuracy': '0.8438', 'epoch': '0.9229'}
+{'loss': '1.144', 'grad_norm': '14.25', 'learning_rate': '7.711e-07', 'num_tokens': '3.536e+06', 'mean_token_accuracy': '0.6601', 'epoch': '0.9231'}
+{'loss': '1.281', 'grad_norm': '11.44', 'learning_rate': '7.694e-07', 'num_tokens': '3.537e+06', 'mean_token_accuracy': '0.6674', 'epoch': '0.9232'}
+{'loss': '0.7089', 'grad_norm': '28', 'learning_rate': '7.676e-07', 'num_tokens': '3.537e+06', 'mean_token_accuracy': '0.8288', 'epoch': '0.9234'}
+{'loss': '0.8203', 'grad_norm': '18.88', 'learning_rate': '7.659e-07', 'num_tokens': '3.537e+06', 'mean_token_accuracy': '0.8158', 'epoch': '0.9236'}
+{'loss': '0.8261', 'grad_norm': '9.438', 'learning_rate': '7.642e-07', 'num_tokens': '3.538e+06', 'mean_token_accuracy': '0.7956', 'epoch': '0.9238'}
+{'loss': '0.862', 'grad_norm': '14.12', 'learning_rate': '7.624e-07', 'num_tokens': '3.539e+06', 'mean_token_accuracy': '0.7691', 'epoch': '0.9239'}
+{'loss': '2.273', 'grad_norm': '26', 'learning_rate': '7.607e-07', 'num_tokens': '3.539e+06', 'mean_token_accuracy': '0.5333', 'epoch': '0.9241'}
+{'loss': '1.961', 'grad_norm': '12.81', 'learning_rate': '7.589e-07', 'num_tokens': '3.539e+06', 'mean_token_accuracy': '0.5437', 'epoch': '0.9243'}
+{'loss': '1.565', 'grad_norm': '14.25', 'learning_rate': '7.572e-07', 'num_tokens': '3.54e+06', 'mean_token_accuracy': '0.6118', 'epoch': '0.9245'}
+{'loss': '0.6609', 'grad_norm': '11.75', 'learning_rate': '7.555e-07', 'num_tokens': '3.541e+06', 'mean_token_accuracy': '0.8123', 'epoch': '0.9246'}
+{'loss': '1.336', 'grad_norm': '12.69', 'learning_rate': '7.537e-07', 'num_tokens': '3.541e+06', 'mean_token_accuracy': '0.6456', 'epoch': '0.9248'}
+{'loss': '1.089', 'grad_norm': '13.31', 'learning_rate': '7.52e-07', 'num_tokens': '3.542e+06', 'mean_token_accuracy': '0.6661', 'epoch': '0.925'}
+{'loss': '0.6346', 'grad_norm': '7.625', 'learning_rate': '7.503e-07', 'num_tokens': '3.544e+06', 'mean_token_accuracy': '0.8012', 'epoch': '0.9251'}
+{'loss': '1.911', 'grad_norm': '11.5', 'learning_rate': '7.485e-07', 'num_tokens': '3.544e+06', 'mean_token_accuracy': '0.5985', 'epoch': '0.9253'}
+{'loss': '1.293', 'grad_norm': '5.781', 'learning_rate': '7.468e-07', 'num_tokens': '3.547e+06', 'mean_token_accuracy': '0.7183', 'epoch': '0.9255'}
+{'loss': '1.376', 'grad_norm': '16.12', 'learning_rate': '7.451e-07', 'num_tokens': '3.548e+06', 'mean_token_accuracy': '0.692', 'epoch': '0.9257'}
+
 93%|█████████▎| 5331/5758 [30:21<02:20,  3.04it/s]
 93%|█████████▎| 5332/5758 [30:21<02:23,  2.97it/s]
                                                   

 93%|█████████▎| 5332/5758 [30:21<02:23,  2.97it/s]
 93%|█████████▎| 5333/5758 [30:21<02:21,  2.99it/s]
                                                   

 93%|█████████▎| 5333/5758 [30:21<02:21,  2.99it/s]
 93%|█████████▎| 5334/5758 [30:22<02:18,  3.05it/s]
                                                   

 93%|█████████▎| 5334/5758 [30:22<02:18,  3.05it/s]
 93%|█████████▎| 5335/5758 [30:22<02:16,  3.10it/s]
                                                   

 93%|█████████▎| 5335/5758 [30:22<02:16,  3.10it/s]
 93%|█████████▎| 5336/5758 [30:22<02:21,  2.98it/s]
                                                   

 93%|█████████▎| 5336/5758 [30:22<02:21,  2.98it/s]
 93%|█████████▎| 5337/5758 [30:23<02:24,  2.91it/s]
                                                   

 93%|█████████▎| 5337/5758 [30:23<02:24,  2.91it/s]
 93%|█████████▎| 5338/5758 [30:23<02:25,  2.88it/s]
                                                   

 93%|█████████▎| 5338/5758 [30:23<02:25,  2.88it/s]
 93%|█████████▎| 5339/5758 [30:23<02:20,  2.99it/s]
                                                   

 93%|█████████▎| 5339/5758 [30:23<02:20,  2.99it/s]
 93%|█████████▎| 5340/5758 [30:24<02:22,  2.94it/s]
                                                   

 93%|█████████▎| 5340/5758 [30:24<02:22,  2.94it/s]
 93%|█████████▎| 5341/5758 [30:24<02:17,  3.04it/s]
                                                   

 93%|█████████▎| 5341/5758 [30:24<02:17,  3.04it/s]
 93%|█████████▎| 5342/5758 [30:24<02:15,  3.08it/s]
                                                   

 93%|█████████▎| 5342/5758 [30:24<02:15,  3.08it/s]
 93%|█████████▎| 5343/5758 [30:25<02:21,  2.93it/s]
                                                   

 93%|█████████▎| 5343/5758 [30:25<02:21,  2.93it/s]
 93%|█████████▎| 5344/5758 [30:25<02:15,  3.05it/s]
                                                   

 93%|█████████▎| 5344/5758 [30:25<02:15,  3.05it/s]
 93%|█████████▎| 5345/5758 [30:25<02:10,  3.15it/s]
                                                   

 93%|█████████▎| 5345/5758 [30:25<02:10,  3.15it/s]
 93%|█████████▎| 5346/5758 [30:26<02:10,  3.17it/s]
                                                   

 93%|█████████▎| 5346/5758 [30:26<02:10,  3.17it/s]
 93%|█████████▎| 5347/5758 [30:26<02:08,  3.20it/s]
                                                   

 93%|█████████▎| 5347/5758 [30:26<02:08,  3.20it/s]
 93%|█████████▎| 5348/5758 [30:26<02:09,  3.17it/s]
                                                   

 93%|█████████▎| 5348/5758 [30:26<02:09,  3.17it/s]
 93%|█████████▎| 5349/5758 [30:27<02:10,  3.13it/s]
                                                   

 93%|█████████▎| 5349/5758 [30:27<02:10,  3.13it/s]
 93%|█████████▎| 5350/5758 [30:27<02:09,  3.14it/s]
                                                   

 93%|█████████▎| 5350/5758 [30:27<02:09,  3.14it/s]
 93%|█████████▎| 5351/5758 [30:27<02:15,  2.99it/s]
                                                   

 93%|█████████▎| 5351/5758 [30:27<02:15,  2.99it/s]
 93%|█████████▎| 5352/5758 [30:28<02:36,  2.59it/s]
                                                   

 93%|█████████▎| 5352/5758 [30:28<02:36,  2.59it/s]
 93%|█████████▎| 5353/5758 [30:28<02:33,  2.65it/s]
                                                   

 93%|█████████▎| 5353/5758 [30:28<02:33,  2.65it/s]
 93%|█████████▎| 5354/5758 [30:28<02:26,  2.77it/s]
                                                   

 93%|█████████▎| 5354/5758 [30:28<02:26,  2.77it/s]
 93%|█████████▎| 5355/5758 [30:29<02:19,  2.89it/s]
                                                   

 93%|█████████▎| 5355/5758 [30:29<02:19,  2.89it/s]
 93%|█████████▎| 5356/5758 [30:29<02:18,  2.89it/s]
                                                   

 93%|█████████▎| 5356/5758 [30:29<02:18,  2.89it/s]
 93%|█████████▎| 5357/5758 [30:30<02:20,  2.85it/s]
                                                   

 93%|█████████▎| 5357/5758 [30:30<02:20,  2.85it/s]
 93%|█████████▎| 5358/5758 [30:30<02:13,  2.99it/s]
                                                   

 93%|█████████▎| 5358/5758 [30:30<02:13,  2.99it/s]
 93%|█████████▎| 5359/5758 [30:30<02:11,  3.03it/s]
                                                   

 93%|█████████▎| 5359/5758 [30:30<02:11,  3.03it/s]
 93%|█████████▎| 5360/5758 [30:30<02:11,  3.02it/s]
                                                   

 93%|█████████▎| 5360/5758 [30:30<02:11,  3.02it/s]
 93%|█████████▎| 5361/5758 [30:31<02:08,  3.09it/s]
                                                   

 93%|█████████▎| 5361/5758 [30:31<02:08,  3.09it/s]
 93%|█████████▎| 5362/5758 [30:31<02:06,  3.14it/s]
                                                   

 93%|█████████▎| 5362/5758 [30:31<02:06,  3.14it/s]
 93%|█████████▎| 5363/5758 [30:31<02:04,  3.18it/s]
                                                   

 93%|█████████▎| 5363/5758 [30:31<02:04,  3.18it/s]
 93%|█████████▎| 5364/5758 [30:32<02:02,  3.22it/s]
                                                   

 93%|█████████▎| 5364/5758 [30:32<02:02,  3.22it/s]
 93%|█████████▎| 5365/5758 [30:32<02:05,  3.12it/s]
                                                   

 93%|█████████▎| 5365/5758 [30:32<02:05,  3.12it/s]
 93%|█████████▎| 5366/5758 [30:32<02:02,  3.19it/s]
                                                   

 93%|█████████▎| 5366/5758 [30:32<02:02,  3.19it/s]
 93%|█████████▎| 5367/5758 [30:33<02:03,  3.17it/s]
                                                   

 93%|█████████▎| 5367/5758 [30:33<02:03,  3.17it/s]
 93%|█████████▎| 5368/5758 [30:33<02:08,  3.03it/s]
                                                   

 93%|█████████▎| 5368/5758 [30:33<02:08,  3.03it/s]
 93%|█████████▎| 5369/5758 [30:33<02:04,  3.12it/s]
                                                   

 93%|█████████▎| 5369/5758 [30:33<02:04,  3.12it/s]
 93%|█████████▎| 5370/5758 [30:34<02:06,  3.07it/s]
                                                   

 93%|█████████▎| 5370/5758 [30:34<02:06,  3.07it/s]
 93%|█████████▎| 5371/5758 [30:34<02:11,  2.95it/s]
                                                   

 93%|█████████▎| 5371/5758 [30:34<02:11,  2.95it/s]
 93%|█████████▎| 5372/5758 [30:34<02:11,  2.93it/s]
                                                   

 93%|█████████▎| 5372/5758 [30:34<02:11,  2.93it/s]
 93%|█████████▎| 5373/5758 [30:35<02:08,  2.99it/s]
                                                   

 93%|█████████▎| 5373/5758 [30:35<02:08,  2.99it/s]
 93%|█████████▎| 5374/5758 [30:35<02:03,  3.10it/s]
                                                   

 93%|█████████▎| 5374/5758 [30:35<02:03,  3.10it/s]
 93%|█████████▎| 5375/5758 [30:35<02:03,  3.09it/s]
                                                   

 93%|█████████▎| 5375/5758 [30:35<02:03,  3.09it/s]
 93%|█████████▎| 5376/5758 [30:36<02:02,  3.12it/s]
                                                   

 93%|█████████▎| 5376/5758 [30:36<02:02,  3.12it/s]
 93%|█████████▎| 5377/5758 [30:36<02:04,  3.05it/s]
                                                   

 93%|█████████▎| 5377/5758 [30:36<02:04,  3.05it/s]
 93%|█████████▎| 5378/5758 [30:36<02:20,  2.71it/s]
                                                   

 93%|█████████▎| 5378/5758 [30:36<02:20,  2.71it/s]
 93%|█████████▎| 5379/5758 [30:37<02:13,  2.84it/s]
                                                   

 93%|█████████▎| 5379/5758 [30:37<02:13,  2.84it/s]
 93%|█████████▎| 5380/5758 [30:37<02:13,  2.83it/s]
                                                   

 93%|█████████▎| 5380/5758 [30:37<02:13,  2.83it/s]
 93%|█████████▎| 5381/5758 [30:37<02:10,  2.89it/s]
                                                   

 93%|█████████▎| 5381/5758 [30:37<02:10,  2.89it/s]
 93%|█████████▎| 5382/5758 [30:38<02:05,  2.99it/s]
                                                   

 93%|█████████▎| 5382/5758 [30:38<02:05,  2.99it/s]
 93%|█████████▎| 5383/5758 [30:38<02:02,  3.06it/s]
                                                   

 93%|█████████▎| 5383/5758 [30:38<02:02,  3.06it/s]
 94%|█████████▎| 5384/5758 [30:38<02:05,  2.97it/s]
                                                   

 94%|█████████▎| 5384/5758 [30:38<02:05,  2.97it/s]
 94%|█████████▎| 5385/5758 [30:39<02:08,  2.91it/s]
                                                   

 94%|█████████▎| 5385/5758 [30:39<02:08,  2.91it/s]
 94%|█████████▎| 5386/5758 [30:39<02:04,  3.00it/s]
                                                   
{'loss': '1.159', 'grad_norm': '7.094', 'learning_rate': '7.433e-07', 'num_tokens': '3.55e+06', 'mean_token_accuracy': '0.7151', 'epoch': '0.9258'}
+{'loss': '1.939', 'grad_norm': '9.938', 'learning_rate': '7.416e-07', 'num_tokens': '3.551e+06', 'mean_token_accuracy': '0.6064', 'epoch': '0.926'}
+{'loss': '1.805', 'grad_norm': '10.31', 'learning_rate': '7.398e-07', 'num_tokens': '3.552e+06', 'mean_token_accuracy': '0.6071', 'epoch': '0.9262'}
+{'loss': '0.6169', 'grad_norm': '14.81', 'learning_rate': '7.381e-07', 'num_tokens': '3.552e+06', 'mean_token_accuracy': '0.8516', 'epoch': '0.9264'}
+{'loss': '1.266', 'grad_norm': '12.12', 'learning_rate': '7.364e-07', 'num_tokens': '3.553e+06', 'mean_token_accuracy': '0.6477', 'epoch': '0.9265'}
+{'loss': '0.8236', 'grad_norm': '27.62', 'learning_rate': '7.346e-07', 'num_tokens': '3.553e+06', 'mean_token_accuracy': '0.8182', 'epoch': '0.9267'}
+{'loss': '1.04', 'grad_norm': '17.25', 'learning_rate': '7.329e-07', 'num_tokens': '3.554e+06', 'mean_token_accuracy': '0.7196', 'epoch': '0.9269'}
+{'loss': '1.097', 'grad_norm': '10.5', 'learning_rate': '7.312e-07', 'num_tokens': '3.555e+06', 'mean_token_accuracy': '0.7077', 'epoch': '0.9271'}
+{'loss': '0.7396', 'grad_norm': '16.12', 'learning_rate': '7.294e-07', 'num_tokens': '3.555e+06', 'mean_token_accuracy': '0.8356', 'epoch': '0.9272'}
+{'loss': '0.782', 'grad_norm': '11.88', 'learning_rate': '7.277e-07', 'num_tokens': '3.556e+06', 'mean_token_accuracy': '0.8369', 'epoch': '0.9274'}
+{'loss': '1.066', 'grad_norm': '13.31', 'learning_rate': '7.259e-07', 'num_tokens': '3.556e+06', 'mean_token_accuracy': '0.691', 'epoch': '0.9276'}
+{'loss': '1.082', 'grad_norm': '7.156', 'learning_rate': '7.242e-07', 'num_tokens': '3.558e+06', 'mean_token_accuracy': '0.7524', 'epoch': '0.9278'}
+{'loss': '0.6154', 'grad_norm': '11.06', 'learning_rate': '7.225e-07', 'num_tokens': '3.558e+06', 'mean_token_accuracy': '0.8374', 'epoch': '0.9279'}
+{'loss': '2', 'grad_norm': '8.5', 'learning_rate': '7.207e-07', 'num_tokens': '3.56e+06', 'mean_token_accuracy': '0.5746', 'epoch': '0.9281'}
+{'loss': '0.8944', 'grad_norm': '46', 'learning_rate': '7.19e-07', 'num_tokens': '3.56e+06', 'mean_token_accuracy': '0.8056', 'epoch': '0.9283'}
+{'loss': '1.205', 'grad_norm': '24.5', 'learning_rate': '7.173e-07', 'num_tokens': '3.56e+06', 'mean_token_accuracy': '0.7', 'epoch': '0.9284'}
+{'loss': '1.239', 'grad_norm': '13.5', 'learning_rate': '7.155e-07', 'num_tokens': '3.561e+06', 'mean_token_accuracy': '0.6639', 'epoch': '0.9286'}
+{'loss': '1.213', 'grad_norm': '13.19', 'learning_rate': '7.138e-07', 'num_tokens': '3.562e+06', 'mean_token_accuracy': '0.6831', 'epoch': '0.9288'}
+{'loss': '2.328', 'grad_norm': '14.62', 'learning_rate': '7.121e-07', 'num_tokens': '3.562e+06', 'mean_token_accuracy': '0.5435', 'epoch': '0.929'}
+{'loss': '1.213', 'grad_norm': '10.31', 'learning_rate': '7.103e-07', 'num_tokens': '3.563e+06', 'mean_token_accuracy': '0.7224', 'epoch': '0.9291'}
+{'loss': '1.674', 'grad_norm': '14.69', 'learning_rate': '7.086e-07', 'num_tokens': '3.563e+06', 'mean_token_accuracy': '0.6461', 'epoch': '0.9293'}
+{'loss': '1.076', 'grad_norm': '31.38', 'learning_rate': '7.068e-07', 'num_tokens': '3.563e+06', 'mean_token_accuracy': '0.7926', 'epoch': '0.9295'}
+{'loss': '1.202', 'grad_norm': '26.88', 'learning_rate': '7.051e-07', 'num_tokens': '3.564e+06', 'mean_token_accuracy': '0.7522', 'epoch': '0.9297'}
+{'loss': '0.4742', 'grad_norm': '6.062', 'learning_rate': '7.034e-07', 'num_tokens': '3.565e+06', 'mean_token_accuracy': '0.8571', 'epoch': '0.9298'}
+{'loss': '1.188', 'grad_norm': '11.38', 'learning_rate': '7.016e-07', 'num_tokens': '3.566e+06', 'mean_token_accuracy': '0.6905', 'epoch': '0.93'}
+{'loss': '1.998', 'grad_norm': '43.75', 'learning_rate': '6.999e-07', 'num_tokens': '3.566e+06', 'mean_token_accuracy': '0.5981', 'epoch': '0.9302'}
+{'loss': '1.14', 'grad_norm': '18.75', 'learning_rate': '6.982e-07', 'num_tokens': '3.566e+06', 'mean_token_accuracy': '0.7557', 'epoch': '0.9304'}
+{'loss': '1.249', 'grad_norm': '12.31', 'learning_rate': '6.964e-07', 'num_tokens': '3.567e+06', 'mean_token_accuracy': '0.652', 'epoch': '0.9305'}
+{'loss': '0.6707', 'grad_norm': '31.88', 'learning_rate': '6.947e-07', 'num_tokens': '3.567e+06', 'mean_token_accuracy': '0.8462', 'epoch': '0.9307'}
+{'loss': '0.8858', 'grad_norm': '30.38', 'learning_rate': '6.929e-07', 'num_tokens': '3.567e+06', 'mean_token_accuracy': '0.7879', 'epoch': '0.9309'}
+{'loss': '0.5985', 'grad_norm': '14.31', 'learning_rate': '6.912e-07', 'num_tokens': '3.567e+06', 'mean_token_accuracy': '0.8497', 'epoch': '0.9311'}
+{'loss': '0.5023', 'grad_norm': '7.281', 'learning_rate': '6.895e-07', 'num_tokens': '3.569e+06', 'mean_token_accuracy': '0.8432', 'epoch': '0.9312'}
+{'loss': '1.372', 'grad_norm': '10.88', 'learning_rate': '6.877e-07', 'num_tokens': '3.57e+06', 'mean_token_accuracy': '0.7143', 'epoch': '0.9314'}
+{'loss': '1.023', 'grad_norm': '13.25', 'learning_rate': '6.86e-07', 'num_tokens': '3.57e+06', 'mean_token_accuracy': '0.6942', 'epoch': '0.9316'}
+{'loss': '1.382', 'grad_norm': '16.75', 'learning_rate': '6.843e-07', 'num_tokens': '3.571e+06', 'mean_token_accuracy': '0.7127', 'epoch': '0.9317'}
+{'loss': '1.206', 'grad_norm': '15', 'learning_rate': '6.825e-07', 'num_tokens': '3.571e+06', 'mean_token_accuracy': '0.7447', 'epoch': '0.9319'}
+{'loss': '1.482', 'grad_norm': '12.5', 'learning_rate': '6.808e-07', 'num_tokens': '3.572e+06', 'mean_token_accuracy': '0.6095', 'epoch': '0.9321'}
+{'loss': '1.897', 'grad_norm': '14.44', 'learning_rate': '6.791e-07', 'num_tokens': '3.572e+06', 'mean_token_accuracy': '0.6065', 'epoch': '0.9323'}
+{'loss': '0.7914', 'grad_norm': '33', 'learning_rate': '6.773e-07', 'num_tokens': '3.572e+06', 'mean_token_accuracy': '0.8462', 'epoch': '0.9324'}
+{'loss': '1.883', 'grad_norm': '23.25', 'learning_rate': '6.756e-07', 'num_tokens': '3.573e+06', 'mean_token_accuracy': '0.6062', 'epoch': '0.9326'}
+{'loss': '1.713', 'grad_norm': '13.94', 'learning_rate': '6.738e-07', 'num_tokens': '3.573e+06', 'mean_token_accuracy': '0.6287', 'epoch': '0.9328'}
+{'loss': '2.034', 'grad_norm': '19.88', 'learning_rate': '6.721e-07', 'num_tokens': '3.574e+06', 'mean_token_accuracy': '0.6004', 'epoch': '0.933'}
+{'loss': '1.735', 'grad_norm': '8.438', 'learning_rate': '6.704e-07', 'num_tokens': '3.575e+06', 'mean_token_accuracy': '0.6217', 'epoch': '0.9331'}
+{'loss': '0.6038', 'grad_norm': '7.406', 'learning_rate': '6.686e-07', 'num_tokens': '3.577e+06', 'mean_token_accuracy': '0.8284', 'epoch': '0.9333'}
+{'loss': '1.079', 'grad_norm': '10.88', 'learning_rate': '6.669e-07', 'num_tokens': '3.577e+06', 'mean_token_accuracy': '0.707', 'epoch': '0.9335'}
+{'loss': '0.6864', 'grad_norm': '17.25', 'learning_rate': '6.652e-07', 'num_tokens': '3.578e+06', 'mean_token_accuracy': '0.8382', 'epoch': '0.9337'}
+{'loss': '1.88', 'grad_norm': '9.938', 'learning_rate': '6.634e-07', 'num_tokens': '3.579e+06', 'mean_token_accuracy': '0.6285', 'epoch': '0.9338'}
+{'loss': '1.278', 'grad_norm': '10.5', 'learning_rate': '6.617e-07', 'num_tokens': '3.58e+06', 'mean_token_accuracy': '0.6451', 'epoch': '0.934'}
+{'loss': '0.9265', 'grad_norm': '17.25', 'learning_rate': '6.6e-07', 'num_tokens': '3.58e+06', 'mean_token_accuracy': '0.8012', 'epoch': '0.9342'}
+{'loss': '1.142', 'grad_norm': '10.44', 'learning_rate': '6.582e-07', 'num_tokens': '3.581e+06', 'mean_token_accuracy': '0.6844', 'epoch': '0.9344'}
+{'loss': '2.035', 'grad_norm': '16.75', 'learning_rate': '6.565e-07', 'num_tokens': '3.581e+06', 'mean_token_accuracy': '0.5501', 'epoch': '0.9345'}
+{'loss': '1.391', 'grad_norm': '29.5', 'learning_rate': '6.547e-07', 'num_tokens': '3.582e+06', 'mean_token_accuracy': '0.7257', 'epoch': '0.9347'}
+{'loss': '1.198', 'grad_norm': '8.875', 'learning_rate': '6.53e-07', 'num_tokens': '3.583e+06', 'mean_token_accuracy': '0.7112', 'epoch': '0.9349'}
+{'loss': '1.095', 'grad_norm': '19.5', 'learning_rate': '6.513e-07', 'num_tokens': '3.583e+06', 'mean_token_accuracy': '0.7549', 'epoch': '0.935'}
+{'loss': '0.8214', 'grad_norm': '28.12', 'learning_rate': '6.495e-07', 'num_tokens': '3.583e+06', 'mean_token_accuracy': '0.823', 'epoch': '0.9352'}
+
 94%|█████████▎| 5386/5758 [30:39<02:04,  3.00it/s]
 94%|█████████▎| 5387/5758 [30:39<02:01,  3.05it/s]
                                                   

 94%|█████████▎| 5387/5758 [30:39<02:01,  3.05it/s]
 94%|█████████▎| 5388/5758 [30:40<02:01,  3.05it/s]
                                                   

 94%|█████████▎| 5388/5758 [30:40<02:01,  3.05it/s]
 94%|█████████▎| 5389/5758 [30:40<01:57,  3.14it/s]
                                                   

 94%|█████████▎| 5389/5758 [30:40<01:57,  3.14it/s]
 94%|█████████▎| 5390/5758 [30:40<01:56,  3.17it/s]
                                                   

 94%|█████████▎| 5390/5758 [30:40<01:56,  3.17it/s]
 94%|█████████▎| 5391/5758 [30:41<01:55,  3.17it/s]
                                                   

 94%|█████████▎| 5391/5758 [30:41<01:55,  3.17it/s]
 94%|████████��▎| 5392/5758 [30:41<01:55,  3.17it/s]
                                                   

 94%|█████████▎| 5392/5758 [30:41<01:55,  3.17it/s]
 94%|█████████▎| 5393/5758 [30:41<01:54,  3.20it/s]
                                                   

 94%|█████████▎| 5393/5758 [30:41<01:54,  3.20it/s]
 94%|█████████▎| 5394/5758 [30:42<01:52,  3.24it/s]
                                                   

 94%|█████████▎| 5394/5758 [30:42<01:52,  3.24it/s]
 94%|█████████▎| 5395/5758 [30:42<01:58,  3.06it/s]
                                                   

 94%|█████████▎| 5395/5758 [30:42<01:58,  3.06it/s]
 94%|█████████▎| 5396/5758 [30:42<01:55,  3.15it/s]
                                                   

 94%|█████████▎| 5396/5758 [30:42<01:55,  3.15it/s]
 94%|█████████▎| 5397/5758 [30:43<02:00,  3.00it/s]
                                                   

 94%|█████████▎| 5397/5758 [30:43<02:00,  3.00it/s]
 94%|█████████▎| 5398/5758 [30:43<01:56,  3.08it/s]
                                                   

 94%|█████████▎| 5398/5758 [30:43<01:56,  3.08it/s]
 94%|█████████▍| 5399/5758 [30:43<01:59,  3.01it/s]
                                                   

 94%|█████████▍| 5399/5758 [30:43<01:59,  3.01it/s]
 94%|█████████▍| 5400/5758 [30:44<01:56,  3.07it/s]
                                                   

 94%|█████████▍| 5400/5758 [30:44<01:56,  3.07it/s]
 94%|█████████▍| 5401/5758 [30:44<01:55,  3.10it/s]
                                                   

 94%|█████████▍| 5401/5758 [30:44<01:55,  3.10it/s]
 94%|█████████▍| 5402/5758 [30:44<01:51,  3.18it/s]
                                                   

 94%|█████████▍| 5402/5758 [30:44<01:51,  3.18it/s]
 94%|█████████▍| 5403/5758 [30:44<01:51,  3.18it/s]
                                                   

 94%|█████████▍| 5403/5758 [30:44<01:51,  3.18it/s]
 94%|█████████▍| 5404/5758 [30:45<01:52,  3.15it/s]
                                                   

 94%|█████████▍| 5404/5758 [30:45<01:52,  3.15it/s]
 94%|█████████▍| 5405/5758 [30:45<02:04,  2.84it/s]
                                                   

 94%|█████████▍| 5405/5758 [30:45<02:04,  2.84it/s]
 94%|█████████▍| 5406/5758 [30:46<01:59,  2.94it/s]
                                                   

 94%|█████████▍| 5406/5758 [30:46<01:59,  2.94it/s]
 94%|█████████▍| 5407/5758 [30:46<02:05,  2.80it/s]
                                                   

 94%|█████████▍| 5407/5758 [30:46<02:05,  2.80it/s]
 94%|█████████▍| 5408/5758 [30:46<01:59,  2.93it/s]
                                                   

 94%|█████████▍| 5408/5758 [30:46<01:59,  2.93it/s]
 94%|█████████▍| 5409/5758 [30:47<01:57,  2.97it/s]
                                                   

 94%|█████████▍| 5409/5758 [30:47<01:57,  2.97it/s]
 94%|█████████▍| 5410/5758 [30:47<01:54,  3.05it/s]
                                                   

 94%|█████████▍| 5410/5758 [30:47<01:54,  3.05it/s]
 94%|█████████▍| 5411/5758 [30:47<01:59,  2.90it/s]
                                                   

 94%|█████████▍| 5411/5758 [30:47<01:59,  2.90it/s]
 94%|█████████▍| 5412/5758 [30:48<02:01,  2.85it/s]
                                                   

 94%|█████████▍| 5412/5758 [30:48<02:01,  2.85it/s]
 94%|█████████▍| 5413/5758 [30:48<01:56,  2.95it/s]
                                                   

 94%|█████████▍| 5413/5758 [30:48<01:56,  2.95it/s]
 94%|█████████▍| 5414/5758 [30:48<01:53,  3.02it/s]
                                                   

 94%|█████████▍| 5414/5758 [30:48<01:53,  3.02it/s]
 94%|█████████▍| 5415/5758 [30:49<01:51,  3.08it/s]
                                                   

 94%|█████████▍| 5415/5758 [30:49<01:51,  3.08it/s]
 94%|█████████▍| 5416/5758 [30:49<01:55,  2.97it/s]
                                                   

 94%|█████████▍| 5416/5758 [30:49<01:55,  2.97it/s]
 94%|█████████▍| 5417/5758 [30:49<01:56,  2.93it/s]
                                                   

 94%|█████████▍| 5417/5758 [30:49<01:56,  2.93it/s]
 94%|█████████▍| 5418/5758 [30:50<01:58,  2.86it/s]
                                                   

 94%|█████████▍| 5418/5758 [30:50<01:58,  2.86it/s]
 94%|█████████▍| 5419/5758 [30:50<01:53,  3.00it/s]
                                                   

 94%|█████████▍| 5419/5758 [30:50<01:53,  3.00it/s]
 94%|█████████▍| 5420/5758 [30:50<01:52,  3.00it/s]
                                                   

 94%|█████████▍| 5420/5758 [30:50<01:52,  3.00it/s]
 94%|█████████▍| 5421/5758 [30:51<01:48,  3.11it/s]
                                                   

 94%|█████████▍| 5421/5758 [30:51<01:48,  3.11it/s]
 94%|█████████▍| 5422/5758 [30:51<01:49,  3.06it/s]
                                                   

 94%|█████████▍| 5422/5758 [30:51<01:49,  3.06it/s]
 94%|█████████▍| 5423/5758 [30:51<01:53,  2.96it/s]
                                                   

 94%|█████████▍| 5423/5758 [30:51<01:53,  2.96it/s]
 94%|█████████▍| 5424/5758 [30:52<01:56,  2.88it/s]
                                                   

 94%|█████████▍| 5424/5758 [30:52<01:56,  2.88it/s]
 94%|█████████▍| 5425/5758 [30:52<01:53,  2.94it/s]
                                                   

 94%|█████████▍| 5425/5758 [30:52<01:53,  2.94it/s]
 94%|█████████▍| 5426/5758 [30:52<01:57,  2.83it/s]
                                                   

 94%|█████████▍| 5426/5758 [30:52<01:57,  2.83it/s]
 94%|█████████▍| 5427/5758 [30:53<01:55,  2.88it/s]
                                                   

 94%|█████████▍| 5427/5758 [30:53<01:55,  2.88it/s]
 94%|█████████▍| 5428/5758 [30:53<01:54,  2.87it/s]
                                                   

 94%|█████████▍| 5428/5758 [30:53<01:54,  2.87it/s]
 94%|█████████▍| 5429/5758 [30:53<01:56,  2.83it/s]
                                                   

 94%|█████████▍| 5429/5758 [30:53<01:56,  2.83it/s]
 94%|█████████▍| 5430/5758 [30:54<01:59,  2.75it/s]
                                                   

 94%|█████████▍| 5430/5758 [30:54<01:59,  2.75it/s]
 94%|█████████▍| 5431/5758 [30:54<01:58,  2.75it/s]
                                                   

 94%|█████████▍| 5431/5758 [30:54<01:58,  2.75it/s]
 94%|█████████▍| 5432/5758 [30:55<01:59,  2.72it/s]
                                                   

 94%|█████████▍| 5432/5758 [30:55<01:59,  2.72it/s]
 94%|█████████▍| 5433/5758 [30:55<02:04,  2.60it/s]
                                                   

 94%|█████████▍| 5433/5758 [30:55<02:04,  2.60it/s]
 94%|█████████▍| 5434/5758 [30:55<02:00,  2.70it/s]
                                                   

 94%|█████████▍| 5434/5758 [30:55<02:00,  2.70it/s]
 94%|█████████▍| 5435/5758 [30:56<01:54,  2.82it/s]
                                                   

 94%|█████████▍| 5435/5758 [30:56<01:54,  2.82it/s]
 94%|█████████▍| 5436/5758 [30:56<01:53,  2.84it/s]
                                                   

 94%|█████████▍| 5436/5758 [30:56<01:53,  2.84it/s]
 94%|█████████▍| 5437/5758 [30:56<01:55,  2.77it/s]
                                                   

 94%|█████████▍| 5437/5758 [30:56<01:55,  2.77it/s]
 94%|█████████▍| 5438/5758 [30:57<01:48,  2.94it/s]
                                                   

 94%|█████████▍| 5438/5758 [30:57<01:48,  2.94it/s]
 94%|█████████▍| 5439/5758 [30:57<01:48,  2.94it/s]
                                                   

 94%|█████████▍| 5439/5758 [30:57<01:48,  2.94it/s]
 94%|█████████▍| 5440/5758 [30:57<01:49,  2.89it/s]
                                                   

 94%|█████████▍| 5440/5758 [30:57<01:49,  2.89it/s]
 94%|█████████▍| 5441/5758 [30:58<01:46,  2.99it/s]
                                                   
{'loss': '1.308', 'grad_norm': '49', 'learning_rate': '6.478e-07', 'num_tokens': '3.584e+06', 'mean_token_accuracy': '0.7241', 'epoch': '0.9354'}
+{'loss': '1.212', 'grad_norm': '12.5', 'learning_rate': '6.461e-07', 'num_tokens': '3.584e+06', 'mean_token_accuracy': '0.682', 'epoch': '0.9356'}
+{'loss': '1.883', 'grad_norm': '26.38', 'learning_rate': '6.443e-07', 'num_tokens': '3.584e+06', 'mean_token_accuracy': '0.639', 'epoch': '0.9357'}
+{'loss': '1.443', 'grad_norm': '13.69', 'learning_rate': '6.426e-07', 'num_tokens': '3.585e+06', 'mean_token_accuracy': '0.6465', 'epoch': '0.9359'}
+{'loss': '1.204', 'grad_norm': '15.38', 'learning_rate': '6.408e-07', 'num_tokens': '3.586e+06', 'mean_token_accuracy': '0.7022', 'epoch': '0.9361'}
+{'loss': '1.141', 'grad_norm': '8.375', 'learning_rate': '6.391e-07', 'num_tokens': '3.587e+06', 'mean_token_accuracy': '0.7109', 'epoch': '0.9363'}
+{'loss': '1.577', 'grad_norm': '9.625', 'learning_rate': '6.374e-07', 'num_tokens': '3.589e+06', 'mean_token_accuracy': '0.5604', 'epoch': '0.9364'}
+{'loss': '2.211', 'grad_norm': '28.38', 'learning_rate': '6.356e-07', 'num_tokens': '3.589e+06', 'mean_token_accuracy': '0.5543', 'epoch': '0.9366'}
+{'loss': '1.415', 'grad_norm': '13.5', 'learning_rate': '6.339e-07', 'num_tokens': '3.59e+06', 'mean_token_accuracy': '0.617', 'epoch': '0.9368'}
+{'loss': '1.87', 'grad_norm': '20.12', 'learning_rate': '6.322e-07', 'num_tokens': '3.59e+06', 'mean_token_accuracy': '0.5959', 'epoch': '0.937'}
+{'loss': '1.807', 'grad_norm': '12.75', 'learning_rate': '6.304e-07', 'num_tokens': '3.591e+06', 'mean_token_accuracy': '0.6066', 'epoch': '0.9371'}
+{'loss': '2.063', 'grad_norm': '17', 'learning_rate': '6.287e-07', 'num_tokens': '3.591e+06', 'mean_token_accuracy': '0.6098', 'epoch': '0.9373'}
+{'loss': '0.5023', 'grad_norm': '7.312', 'learning_rate': '6.27e-07', 'num_tokens': '3.593e+06', 'mean_token_accuracy': '0.8453', 'epoch': '0.9375'}
+{'loss': '0.5444', 'grad_norm': '6.062', 'learning_rate': '6.252e-07', 'num_tokens': '3.595e+06', 'mean_token_accuracy': '0.827', 'epoch': '0.9377'}
+{'loss': '2.034', 'grad_norm': '14.5', 'learning_rate': '6.235e-07', 'num_tokens': '3.595e+06', 'mean_token_accuracy': '0.5972', 'epoch': '0.9378'}
+{'loss': '0.7815', 'grad_norm': '11.5', 'learning_rate': '6.217e-07', 'num_tokens': '3.596e+06', 'mean_token_accuracy': '0.7456', 'epoch': '0.938'}
+{'loss': '1.429', 'grad_norm': '13.31', 'learning_rate': '6.2e-07', 'num_tokens': '3.597e+06', 'mean_token_accuracy': '0.6281', 'epoch': '0.9382'}
+{'loss': '0.7663', 'grad_norm': '7.969', 'learning_rate': '6.183e-07', 'num_tokens': '3.598e+06', 'mean_token_accuracy': '0.7948', 'epoch': '0.9383'}
+{'loss': '1.291', 'grad_norm': '11.94', 'learning_rate': '6.165e-07', 'num_tokens': '3.599e+06', 'mean_token_accuracy': '0.6557', 'epoch': '0.9385'}
+{'loss': '1.163', 'grad_norm': '12.12', 'learning_rate': '6.148e-07', 'num_tokens': '3.6e+06', 'mean_token_accuracy': '0.7062', 'epoch': '0.9387'}
+{'loss': '0.9776', 'grad_norm': '11.75', 'learning_rate': '6.131e-07', 'num_tokens': '3.6e+06', 'mean_token_accuracy': '0.7254', 'epoch': '0.9389'}
+{'loss': '1.622', 'grad_norm': '34.75', 'learning_rate': '6.113e-07', 'num_tokens': '3.601e+06', 'mean_token_accuracy': '0.6884', 'epoch': '0.939'}
+{'loss': '0.7491', 'grad_norm': '8.938', 'learning_rate': '6.096e-07', 'num_tokens': '3.602e+06', 'mean_token_accuracy': '0.768', 'epoch': '0.9392'}
+{'loss': '1.998', 'grad_norm': '19.88', 'learning_rate': '6.078e-07', 'num_tokens': '3.602e+06', 'mean_token_accuracy': '0.5569', 'epoch': '0.9394'}
+{'loss': '1.329', 'grad_norm': '15.25', 'learning_rate': '6.061e-07', 'num_tokens': '3.603e+06', 'mean_token_accuracy': '0.6921', 'epoch': '0.9396'}
+{'loss': '1.533', 'grad_norm': '21.38', 'learning_rate': '6.044e-07', 'num_tokens': '3.603e+06', 'mean_token_accuracy': '0.6784', 'epoch': '0.9397'}
+{'loss': '1.925', 'grad_norm': '19.38', 'learning_rate': '6.026e-07', 'num_tokens': '3.604e+06', 'mean_token_accuracy': '0.6096', 'epoch': '0.9399'}
+{'loss': '1.284', 'grad_norm': '13.5', 'learning_rate': '6.009e-07', 'num_tokens': '3.605e+06', 'mean_token_accuracy': '0.6332', 'epoch': '0.9401'}
+{'loss': '1.087', 'grad_norm': '12.88', 'learning_rate': '5.992e-07', 'num_tokens': '3.605e+06', 'mean_token_accuracy': '0.6883', 'epoch': '0.9403'}
+{'loss': '1.699', 'grad_norm': '11.88', 'learning_rate': '5.974e-07', 'num_tokens': '3.606e+06', 'mean_token_accuracy': '0.6117', 'epoch': '0.9404'}
+{'loss': '1.582', 'grad_norm': '19.12', 'learning_rate': '5.957e-07', 'num_tokens': '3.606e+06', 'mean_token_accuracy': '0.6772', 'epoch': '0.9406'}
+{'loss': '1.264', 'grad_norm': '7.406', 'learning_rate': '5.94e-07', 'num_tokens': '3.608e+06', 'mean_token_accuracy': '0.7117', 'epoch': '0.9408'}
+{'loss': '0.8632', 'grad_norm': '12.75', 'learning_rate': '5.922e-07', 'num_tokens': '3.609e+06', 'mean_token_accuracy': '0.7757', 'epoch': '0.941'}
+{'loss': '0.5604', 'grad_norm': '35.5', 'learning_rate': '5.905e-07', 'num_tokens': '3.609e+06', 'mean_token_accuracy': '0.863', 'epoch': '0.9411'}
+{'loss': '0.8394', 'grad_norm': '17.25', 'learning_rate': '5.887e-07', 'num_tokens': '3.609e+06', 'mean_token_accuracy': '0.8', 'epoch': '0.9413'}
+{'loss': '0.6221', 'grad_norm': '8.062', 'learning_rate': '5.87e-07', 'num_tokens': '3.61e+06', 'mean_token_accuracy': '0.8127', 'epoch': '0.9415'}
+{'loss': '2.205', 'grad_norm': '10.44', 'learning_rate': '5.853e-07', 'num_tokens': '3.611e+06', 'mean_token_accuracy': '0.5611', 'epoch': '0.9416'}
+{'loss': '1.613', 'grad_norm': '19.88', 'learning_rate': '5.835e-07', 'num_tokens': '3.612e+06', 'mean_token_accuracy': '0.6455', 'epoch': '0.9418'}
+{'loss': '1.256', 'grad_norm': '21.25', 'learning_rate': '5.818e-07', 'num_tokens': '3.612e+06', 'mean_token_accuracy': '0.7146', 'epoch': '0.942'}
+{'loss': '1.034', 'grad_norm': '11.12', 'learning_rate': '5.801e-07', 'num_tokens': '3.613e+06', 'mean_token_accuracy': '0.6897', 'epoch': '0.9422'}
+{'loss': '0.965', 'grad_norm': '15', 'learning_rate': '5.783e-07', 'num_tokens': '3.614e+06', 'mean_token_accuracy': '0.7751', 'epoch': '0.9423'}
+{'loss': '0.7282', 'grad_norm': '13.62', 'learning_rate': '5.766e-07', 'num_tokens': '3.614e+06', 'mean_token_accuracy': '0.8345', 'epoch': '0.9425'}
+{'loss': '2.12', 'grad_norm': '20.38', 'learning_rate': '5.749e-07', 'num_tokens': '3.615e+06', 'mean_token_accuracy': '0.5669', 'epoch': '0.9427'}
+{'loss': '1.271', 'grad_norm': '15.12', 'learning_rate': '5.731e-07', 'num_tokens': '3.615e+06', 'mean_token_accuracy': '0.6418', 'epoch': '0.9429'}
+{'loss': '1.178', 'grad_norm': '13.88', 'learning_rate': '5.714e-07', 'num_tokens': '3.616e+06', 'mean_token_accuracy': '0.7179', 'epoch': '0.943'}
+{'loss': '0.4904', 'grad_norm': '5.781', 'learning_rate': '5.696e-07', 'num_tokens': '3.618e+06', 'mean_token_accuracy': '0.8602', 'epoch': '0.9432'}
+{'loss': '1.141', 'grad_norm': '13.88', 'learning_rate': '5.679e-07', 'num_tokens': '3.618e+06', 'mean_token_accuracy': '0.7041', 'epoch': '0.9434'}
+{'loss': '1.695', 'grad_norm': '31.38', 'learning_rate': '5.662e-07', 'num_tokens': '3.618e+06', 'mean_token_accuracy': '0.7169', 'epoch': '0.9436'}
+{'loss': '1.44', 'grad_norm': '20.5', 'learning_rate': '5.644e-07', 'num_tokens': '3.619e+06', 'mean_token_accuracy': '0.7207', 'epoch': '0.9437'}
+{'loss': '0.5007', 'grad_norm': '6.688', 'learning_rate': '5.627e-07', 'num_tokens': '3.62e+06', 'mean_token_accuracy': '0.8465', 'epoch': '0.9439'}
+{'loss': '0.8816', 'grad_norm': '20.38', 'learning_rate': '5.61e-07', 'num_tokens': '3.62e+06', 'mean_token_accuracy': '0.819', 'epoch': '0.9441'}
+{'loss': '1.438', 'grad_norm': '25.25', 'learning_rate': '5.592e-07', 'num_tokens': '3.621e+06', 'mean_token_accuracy': '0.7308', 'epoch': '0.9443'}
+{'loss': '1.07', 'grad_norm': '13', 'learning_rate': '5.575e-07', 'num_tokens': '3.621e+06', 'mean_token_accuracy': '0.6932', 'epoch': '0.9444'}
+{'loss': '0.9857', 'grad_norm': '15.94', 'learning_rate': '5.557e-07', 'num_tokens': '3.622e+06', 'mean_token_accuracy': '0.7581', 'epoch': '0.9446'}
+{'loss': '1.454', 'grad_norm': '11.19', 'learning_rate': '5.54e-07', 'num_tokens': '3.623e+06', 'mean_token_accuracy': '0.6234', 'epoch': '0.9448'}
+
 94%|█████████▍| 5441/5758 [30:58<01:46,  2.99it/s]
 95%|█████████▍| 5442/5758 [30:58<01:41,  3.11it/s]
                                                   

 95%|█████████▍| 5442/5758 [30:58<01:41,  3.11it/s]
 95%|█████████▍| 5443/5758 [30:58<01:42,  3.08it/s]
                                                   

 95%|█████████▍| 5443/5758 [30:58<01:42,  3.08it/s]
 95%|█████████▍| 5444/5758 [30:59<01:43,  3.03it/s]
                                                   

 95%|█████████▍| 5444/5758 [30:59<01:43,  3.03it/s]
 95%|█████████▍| 5445/5758 [30:59<01:44,  2.98it/s]
                                                   

 95%|█████████▍| 5445/5758 [30:59<01:44,  2.98it/s]
 95%|█████████▍| 5446/5758 [30:59<01:42,  3.05it/s]
                                                   

 95%|█████████▍| 5446/5758 [30:59<01:42,  3.05it/s]
 95%|█████████▍| 5447/5758 [31:00<01:48,  2.88it/s]
                                                   

 95%|█████████▍| 5447/5758 [31:00<01:48,  2.88it/s]
 95%|█████████▍| 5448/5758 [31:00<01:45,  2.93it/s]
                                                   

 95%|█████████▍| 5448/5758 [31:00<01:45,  2.93it/s]
 95%|█████████▍| 5449/5758 [31:00<01:49,  2.83it/s]
                                                   

 95%|█████████▍| 5449/5758 [31:00<01:49,  2.83it/s]
 95%|█████████▍| 5450/5758 [31:01<01:49,  2.82it/s]
                                                   

 95%|█████████▍| 5450/5758 [31:01<01:49,  2.82it/s]
 95%|█████████▍| 5451/5758 [31:01<01:43,  2.98it/s]
                                                   

 95%|█████████▍| 5451/5758 [31:01<01:43,  2.98it/s]
 95%|█████████▍| 5452/5758 [31:01<01:47,  2.85it/s]
                                                   

 95%|█████████▍| 5452/5758 [31:01<01:47,  2.85it/s]
 95%|█████████▍| 5453/5758 [31:02<01:43,  2.96it/s]
                                                   

 95%|█████████▍| 5453/5758 [31:02<01:43,  2.96it/s]
 95%|█████████▍| 5454/5758 [31:02<01:39,  3.05it/s]
                                                   

 95%|█████████▍| 5454/5758 [31:02<01:39,  3.05it/s]
 95%|█████████▍| 5455/5758 [31:02<01:42,  2.97it/s]
                                                   

 95%|█████████▍| 5455/5758 [31:02<01:42,  2.97it/s]
 95%|█████████▍| 5456/5758 [31:03<01:36,  3.14it/s]
                                                   

 95%|█████████▍| 5456/5758 [31:03<01:36,  3.14it/s]
 95%|█████████▍| 5457/5758 [31:03<01:36,  3.13it/s]
                                                   

 95%|█████████▍| 5457/5758 [31:03<01:36,  3.13it/s]
 95%|█████████▍| 5458/5758 [31:03<01:40,  2.97it/s]
                                                   

 95%|█████████▍| 5458/5758 [31:03<01:40,  2.97it/s]
 95%|█████████▍| 5459/5758 [31:04<01:37,  3.08it/s]
                                                   

 95%|█████████▍| 5459/5758 [31:04<01:37,  3.08it/s]
 95%|█████████▍| 5460/5758 [31:04<01:35,  3.12it/s]
                                                   

 95%|█████████▍| 5460/5758 [31:04<01:35,  3.12it/s]
 95%|█████████▍| 5461/5758 [31:04<01:44,  2.84it/s]
                                                   

 95%|█████████▍| 5461/5758 [31:04<01:44,  2.84it/s]
 95%|█████████▍| 5462/5758 [31:05<01:42,  2.90it/s]
                                                   

 95%|█████████▍| 5462/5758 [31:05<01:42,  2.90it/s]
 95%|█████████▍| 5463/5758 [31:05<01:36,  3.05it/s]
                                                   

 95%|█████████▍| 5463/5758 [31:05<01:36,  3.05it/s]
 95%|█████████▍| 5464/5758 [31:05<01:30,  3.24it/s]
                                                   

 95%|█████████▍| 5464/5758 [31:05<01:30,  3.24it/s]
 95%|█████████▍| 5465/5758 [31:06<01:30,  3.25it/s]
                                                   

 95%|█████████▍| 5465/5758 [31:06<01:30,  3.25it/s]
 95%|█████████▍| 5466/5758 [31:06<01:28,  3.30it/s]
                                                   

 95%|█████████▍| 5466/5758 [31:06<01:28,  3.30it/s]
 95%|█████████▍| 5467/5758 [31:06<01:34,  3.09it/s]
                                                   

 95%|█████████▍| 5467/5758 [31:06<01:34,  3.09it/s]
 95%|█████████▍| 5468/5758 [31:07<01:39,  2.93it/s]
                                                   

 95%|█████████▍| 5468/5758 [31:07<01:39,  2.93it/s]
 95%|█████████▍| 5469/5758 [31:07<01:36,  2.98it/s]
                                                   

 95%|█████████▍| 5469/5758 [31:07<01:36,  2.98it/s]
 95%|█████████▍| 5470/5758 [31:07<01:32,  3.11it/s]
                                                   

 95%|█████████▍| 5470/5758 [31:07<01:32,  3.11it/s]
 95%|█████████▌| 5471/5758 [31:07<01:29,  3.21it/s]
                                                   

 95%|█████████▌| 5471/5758 [31:07<01:29,  3.21it/s]
 95%|█████████▌| 5472/5758 [31:08<01:27,  3.26it/s]
                                                   

 95%|█████████▌| 5472/5758 [31:08<01:27,  3.26it/s]
 95%|█████████▌| 5473/5758 [31:08<01:29,  3.18it/s]
                                                   

 95%|█████████▌| 5473/5758 [31:08<01:29,  3.18it/s]
 95%|█████████▌| 5474/5758 [31:08<01:33,  3.04it/s]
                                                   

 95%|█████████▌| 5474/5758 [31:08<01:33,  3.04it/s]
 95%|█████████▌| 5475/5758 [31:09<01:29,  3.15it/s]
                                                   

 95%|█████████▌| 5475/5758 [31:09<01:29,  3.15it/s]
 95%|█████████▌| 5476/5758 [31:09<01:27,  3.24it/s]
                                                   

 95%|█████████▌| 5476/5758 [31:09<01:27,  3.24it/s]
 95%|█████████▌| 5477/5758 [31:09<01:27,  3.23it/s]
                                                   

 95%|█████████▌| 5477/5758 [31:09<01:27,  3.23it/s]
 95%|█████████▌| 5478/5758 [31:10<01:29,  3.13it/s]
                                                   

 95%|█████████▌| 5478/5758 [31:10<01:29,  3.13it/s]
 95%|█████████▌| 5479/5758 [31:10<01:30,  3.07it/s]
                                                   

 95%|█████████▌| 5479/5758 [31:10<01:30,  3.07it/s]
 95%|█████████▌| 5480/5758 [31:10<01:26,  3.20it/s]
                                                   

 95%|█████████▌| 5480/5758 [31:10<01:26,  3.20it/s]
 95%|█████████▌| 5481/5758 [31:11<01:26,  3.19it/s]
                                                   

 95%|█████████▌| 5481/5758 [31:11<01:26,  3.19it/s]
 95%|█████████▌| 5482/5758 [31:11<01:26,  3.21it/s]
                                                   

 95%|█████████▌| 5482/5758 [31:11<01:26,  3.21it/s]
 95%|█████████▌| 5483/5758 [31:11<01:31,  2.99it/s]
                                                   

 95%|█████████▌| 5483/5758 [31:11<01:31,  2.99it/s]
 95%|█████████▌| 5484/5758 [31:12<01:31,  3.00it/s]
                                                   

 95%|█████████▌| 5484/5758 [31:12<01:31,  3.00it/s]
 95%|█████████▌| 5485/5758 [31:12<01:32,  2.96it/s]
                                                   

 95%|█████████▌| 5485/5758 [31:12<01:32,  2.96it/s]
 95%|█████████▌| 5486/5758 [31:13<01:42,  2.65it/s]
                                                   

 95%|█████████▌| 5486/5758 [31:13<01:42,  2.65it/s]
 95%|█████████▌| 5487/5758 [31:13<01:37,  2.78it/s]
                                                   

 95%|█████████▌| 5487/5758 [31:13<01:37,  2.78it/s]
 95%|█████████▌| 5488/5758 [31:13<01:32,  2.91it/s]
                                                   

 95%|█████████▌| 5488/5758 [31:13<01:32,  2.91it/s]
 95%|█████████▌| 5489/5758 [31:13<01:30,  2.97it/s]
                                                   

 95%|█████████▌| 5489/5758 [31:13<01:30,  2.97it/s]
 95%|█████████▌| 5490/5758 [31:14<01:31,  2.94it/s]
                                                   

 95%|█████████▌| 5490/5758 [31:14<01:31,  2.94it/s]
 95%|█████████▌| 5491/5758 [31:14<01:30,  2.96it/s]
                                                   

 95%|█████████▌| 5491/5758 [31:14<01:30,  2.96it/s]
 95%|█████████▌| 5492/5758 [31:14<01:28,  3.00it/s]
                                                   

 95%|█████████▌| 5492/5758 [31:14<01:28,  3.00it/s]
 95%|█████████▌| 5493/5758 [31:15<01:24,  3.12it/s]
                                                   

 95%|█████████▌| 5493/5758 [31:15<01:24,  3.12it/s]
 95%|█████████▌| 5494/5758 [31:15<01:23,  3.17it/s]
                                                   

 95%|█████████▌| 5494/5758 [31:15<01:23,  3.17it/s]
 95%|█████████▌| 5495/5758 [31:15<01:25,  3.09it/s]
                                                   

 95%|█████████▌| 5495/5758 [31:15<01:25,  3.09it/s]
 95%|█████████▌| 5496/5758 [31:16<01:25,  3.08it/s]
                                                   
{'loss': '1.112', 'grad_norm': '13.62', 'learning_rate': '5.523e-07', 'num_tokens': '3.623e+06', 'mean_token_accuracy': '0.6716', 'epoch': '0.9449'}
+{'loss': '2.002', 'grad_norm': '13.44', 'learning_rate': '5.505e-07', 'num_tokens': '3.624e+06', 'mean_token_accuracy': '0.6224', 'epoch': '0.9451'}
+{'loss': '1.72', 'grad_norm': '18.75', 'learning_rate': '5.488e-07', 'num_tokens': '3.624e+06', 'mean_token_accuracy': '0.6083', 'epoch': '0.9453'}
+{'loss': '1.6', 'grad_norm': '9.812', 'learning_rate': '5.471e-07', 'num_tokens': '3.625e+06', 'mean_token_accuracy': '0.6629', 'epoch': '0.9455'}
+{'loss': '1.215', 'grad_norm': '21.25', 'learning_rate': '5.453e-07', 'num_tokens': '3.626e+06', 'mean_token_accuracy': '0.7303', 'epoch': '0.9456'}
+{'loss': '0.8444', 'grad_norm': '21.5', 'learning_rate': '5.436e-07', 'num_tokens': '3.626e+06', 'mean_token_accuracy': '0.7747', 'epoch': '0.9458'}
+{'loss': '1.293', 'grad_norm': '35', 'learning_rate': '5.419e-07', 'num_tokens': '3.626e+06', 'mean_token_accuracy': '0.6541', 'epoch': '0.946'}
+{'loss': '0.9253', 'grad_norm': '15.5', 'learning_rate': '5.401e-07', 'num_tokens': '3.626e+06', 'mean_token_accuracy': '0.7786', 'epoch': '0.9462'}
+{'loss': '1.765', 'grad_norm': '21.38', 'learning_rate': '5.384e-07', 'num_tokens': '3.627e+06', 'mean_token_accuracy': '0.6113', 'epoch': '0.9463'}
+{'loss': '2.162', 'grad_norm': '14.81', 'learning_rate': '5.366e-07', 'num_tokens': '3.627e+06', 'mean_token_accuracy': '0.5761', 'epoch': '0.9465'}
+{'loss': '1.96', 'grad_norm': '16.5', 'learning_rate': '5.349e-07', 'num_tokens': '3.628e+06', 'mean_token_accuracy': '0.5795', 'epoch': '0.9467'}
+{'loss': '1.814', 'grad_norm': '17.62', 'learning_rate': '5.332e-07', 'num_tokens': '3.628e+06', 'mean_token_accuracy': '0.6091', 'epoch': '0.9469'}
+{'loss': '1.451', 'grad_norm': '17', 'learning_rate': '5.314e-07', 'num_tokens': '3.629e+06', 'mean_token_accuracy': '0.7224', 'epoch': '0.947'}
+{'loss': '0.9451', 'grad_norm': '12.19', 'learning_rate': '5.297e-07', 'num_tokens': '3.629e+06', 'mean_token_accuracy': '0.7388', 'epoch': '0.9472'}
+{'loss': '1.266', 'grad_norm': '13.44', 'learning_rate': '5.28e-07', 'num_tokens': '3.63e+06', 'mean_token_accuracy': '0.739', 'epoch': '0.9474'}
+{'loss': '1.159', 'grad_norm': '46.25', 'learning_rate': '5.262e-07', 'num_tokens': '3.63e+06', 'mean_token_accuracy': '0.7167', 'epoch': '0.9476'}
+{'loss': '1.452', 'grad_norm': '37.5', 'learning_rate': '5.245e-07', 'num_tokens': '3.63e+06', 'mean_token_accuracy': '0.7222', 'epoch': '0.9477'}
+{'loss': '0.8857', 'grad_norm': '12.81', 'learning_rate': '5.228e-07', 'num_tokens': '3.63e+06', 'mean_token_accuracy': '0.7505', 'epoch': '0.9479'}
+{'loss': '1.009', 'grad_norm': '11.44', 'learning_rate': '5.21e-07', 'num_tokens': '3.631e+06', 'mean_token_accuracy': '0.7618', 'epoch': '0.9481'}
+{'loss': '0.9486', 'grad_norm': '23.38', 'learning_rate': '5.193e-07', 'num_tokens': '3.631e+06', 'mean_token_accuracy': '0.7809', 'epoch': '0.9482'}
+{'loss': '0.6008', 'grad_norm': '6.781', 'learning_rate': '5.175e-07', 'num_tokens': '3.633e+06', 'mean_token_accuracy': '0.8313', 'epoch': '0.9484'}
+{'loss': '1.319', 'grad_norm': '8.688', 'learning_rate': '5.158e-07', 'num_tokens': '3.634e+06', 'mean_token_accuracy': '0.6873', 'epoch': '0.9486'}
+{'loss': '3.368', 'grad_norm': '39.25', 'learning_rate': '5.141e-07', 'num_tokens': '3.634e+06', 'mean_token_accuracy': '0.4371', 'epoch': '0.9488'}
+{'loss': '0.9221', 'grad_norm': '46', 'learning_rate': '5.123e-07', 'num_tokens': '3.635e+06', 'mean_token_accuracy': '0.7925', 'epoch': '0.9489'}
+{'loss': '0.4372', 'grad_norm': '23.75', 'learning_rate': '5.106e-07', 'num_tokens': '3.635e+06', 'mean_token_accuracy': '0.9244', 'epoch': '0.9491'}
+{'loss': '1.526', 'grad_norm': '13.19', 'learning_rate': '5.089e-07', 'num_tokens': '3.635e+06', 'mean_token_accuracy': '0.6635', 'epoch': '0.9493'}
+{'loss': '1.777', 'grad_norm': '13.75', 'learning_rate': '5.071e-07', 'num_tokens': '3.636e+06', 'mean_token_accuracy': '0.6413', 'epoch': '0.9495'}
+{'loss': '3.522', 'grad_norm': '56.5', 'learning_rate': '5.054e-07', 'num_tokens': '3.636e+06', 'mean_token_accuracy': '0.5227', 'epoch': '0.9496'}
+{'loss': '3.056', 'grad_norm': '20.75', 'learning_rate': '5.036e-07', 'num_tokens': '3.636e+06', 'mean_token_accuracy': '0.443', 'epoch': '0.9498'}
+{'loss': '1.88', 'grad_norm': '11.94', 'learning_rate': '5.019e-07', 'num_tokens': '3.637e+06', 'mean_token_accuracy': '0.6044', 'epoch': '0.95'}
+{'loss': '0.548', 'grad_norm': '9.062', 'learning_rate': '5.002e-07', 'num_tokens': '3.638e+06', 'mean_token_accuracy': '0.8282', 'epoch': '0.9502'}
+{'loss': '1.279', 'grad_norm': '13.56', 'learning_rate': '4.984e-07', 'num_tokens': '3.639e+06', 'mean_token_accuracy': '0.6443', 'epoch': '0.9503'}
+{'loss': '1.065', 'grad_norm': '12', 'learning_rate': '4.967e-07', 'num_tokens': '3.64e+06', 'mean_token_accuracy': '0.7289', 'epoch': '0.9505'}
+{'loss': '2.166', 'grad_norm': '19.12', 'learning_rate': '4.95e-07', 'num_tokens': '3.64e+06', 'mean_token_accuracy': '0.5526', 'epoch': '0.9507'}
+{'loss': '0.6995', 'grad_norm': '8.375', 'learning_rate': '4.932e-07', 'num_tokens': '3.642e+06', 'mean_token_accuracy': '0.7985', 'epoch': '0.9509'}
+{'loss': '0.8743', 'grad_norm': '8.062', 'learning_rate': '4.915e-07', 'num_tokens': '3.643e+06', 'mean_token_accuracy': '0.7878', 'epoch': '0.951'}
+{'loss': '1.036', 'grad_norm': '11.69', 'learning_rate': '4.898e-07', 'num_tokens': '3.644e+06', 'mean_token_accuracy': '0.7139', 'epoch': '0.9512'}
+{'loss': '1.42', 'grad_norm': '13.94', 'learning_rate': '4.88e-07', 'num_tokens': '3.644e+06', 'mean_token_accuracy': '0.7028', 'epoch': '0.9514'}
+{'loss': '0.8416', 'grad_norm': '17.88', 'learning_rate': '4.863e-07', 'num_tokens': '3.644e+06', 'mean_token_accuracy': '0.8261', 'epoch': '0.9515'}
+{'loss': '1.334', 'grad_norm': '20.75', 'learning_rate': '4.845e-07', 'num_tokens': '3.645e+06', 'mean_token_accuracy': '0.7081', 'epoch': '0.9517'}
+{'loss': '1.694', 'grad_norm': '15.19', 'learning_rate': '4.828e-07', 'num_tokens': '3.645e+06', 'mean_token_accuracy': '0.6519', 'epoch': '0.9519'}
+{'loss': '1.123', 'grad_norm': '12.44', 'learning_rate': '4.811e-07', 'num_tokens': '3.646e+06', 'mean_token_accuracy': '0.6952', 'epoch': '0.9521'}
+{'loss': '0.9632', 'grad_norm': '12.88', 'learning_rate': '4.793e-07', 'num_tokens': '3.646e+06', 'mean_token_accuracy': '0.6981', 'epoch': '0.9522'}
+{'loss': '1.029', 'grad_norm': '11.38', 'learning_rate': '4.776e-07', 'num_tokens': '3.647e+06', 'mean_token_accuracy': '0.7224', 'epoch': '0.9524'}
+{'loss': '0.6381', 'grad_norm': '27.25', 'learning_rate': '4.759e-07', 'num_tokens': '3.647e+06', 'mean_token_accuracy': '0.8774', 'epoch': '0.9526'}
+{'loss': '1.906', 'grad_norm': '19.12', 'learning_rate': '4.741e-07', 'num_tokens': '3.648e+06', 'mean_token_accuracy': '0.6321', 'epoch': '0.9528'}
+{'loss': '2.035', 'grad_norm': '8.625', 'learning_rate': '4.724e-07', 'num_tokens': '3.649e+06', 'mean_token_accuracy': '0.5663', 'epoch': '0.9529'}
+{'loss': '2.014', 'grad_norm': '15.5', 'learning_rate': '4.706e-07', 'num_tokens': '3.65e+06', 'mean_token_accuracy': '0.5691', 'epoch': '0.9531'}
+{'loss': '0.5722', 'grad_norm': '6.594', 'learning_rate': '4.689e-07', 'num_tokens': '3.652e+06', 'mean_token_accuracy': '0.8268', 'epoch': '0.9533'}
+{'loss': '1.701', 'grad_norm': '20.62', 'learning_rate': '4.672e-07', 'num_tokens': '3.652e+06', 'mean_token_accuracy': '0.6532', 'epoch': '0.9535'}
+{'loss': '1.288', 'grad_norm': '8.25', 'learning_rate': '4.654e-07', 'num_tokens': '3.654e+06', 'mean_token_accuracy': '0.6991', 'epoch': '0.9536'}
+{'loss': '1.484', 'grad_norm': '11.94', 'learning_rate': '4.637e-07', 'num_tokens': '3.654e+06', 'mean_token_accuracy': '0.6285', 'epoch': '0.9538'}
+{'loss': '0.5584', 'grad_norm': '8.062', 'learning_rate': '4.62e-07', 'num_tokens': '3.656e+06', 'mean_token_accuracy': '0.8276', 'epoch': '0.954'}
+{'loss': '1.158', 'grad_norm': '13.12', 'learning_rate': '4.602e-07', 'num_tokens': '3.656e+06', 'mean_token_accuracy': '0.6906', 'epoch': '0.9542'}
+{'loss': '1.057', 'grad_norm': '6.375', 'learning_rate': '4.585e-07', 'num_tokens': '3.658e+06', 'mean_token_accuracy': '0.7456', 'epoch': '0.9543'}
+
 95%|█████████▌| 5496/5758 [31:16<01:25,  3.08it/s]
 95%|█████████▌| 5497/5758 [31:16<01:22,  3.15it/s]
                                                   

 95%|█████████▌| 5497/5758 [31:16<01:22,  3.15it/s]
 95%|█████████▌| 5498/5758 [31:16<01:22,  3.16it/s]
                                                   

 95%|█████████▌| 5498/5758 [31:16<01:22,  3.16it/s]
 96%|█████████▌| 5499/5758 [31:17<01:26,  3.00it/s]
                                                   

 96%|█████████▌| 5499/5758 [31:17<01:26,  3.00it/s]
 96%|█████████▌| 5500/5758 [31:17<01:25,  3.02it/s]
                                                   

 96%|█████████▌| 5500/5758 [31:17<01:25,  3.02it/s]{'loss': '1.3', 'grad_norm': '11.38', 'learning_rate': '4.568e-07', 'num_tokens': '3.659e+06', 'mean_token_accuracy': '0.7111', 'epoch': '0.9545'}
+{'loss': '1.107', 'grad_norm': '13.44', 'learning_rate': '4.55e-07', 'num_tokens': '3.66e+06', 'mean_token_accuracy': '0.695', 'epoch': '0.9547'}
+{'loss': '1.808', 'grad_norm': '21.12', 'learning_rate': '4.533e-07', 'num_tokens': '3.66e+06', 'mean_token_accuracy': '0.6274', 'epoch': '0.9548'}
+{'loss': '2.177', 'grad_norm': '18.62', 'learning_rate': '4.515e-07', 'num_tokens': '3.66e+06', 'mean_token_accuracy': '0.5473', 'epoch': '0.955'}
+{'loss': '1.077', 'grad_norm': '33', 'learning_rate': '4.498e-07', 'num_tokens': '3.661e+06', 'mean_token_accuracy': '0.7363', 'epoch': '0.9552'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.56it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.56it/s]
+
 96%|█████████▌| 5501/5758 [31:19<03:49,  1.12it/s]
                                                   

 96%|█████████▌| 5501/5758 [31:19<03:49,  1.12it/s]
 96%|█████████▌| 5502/5758 [31:20<03:08,  1.36it/s]
                                                   

 96%|█████████▌| 5502/5758 [31:20<03:08,  1.36it/s]
 96%|█████████▌| 5503/5758 [31:20<02:34,  1.65it/s]
                                                   

 96%|█████████▌| 5503/5758 [31:20<02:34,  1.65it/s]
 96%|█████████▌| 5504/5758 [31:20<02:13,  1.90it/s]
                                                   

 96%|█████████▌| 5504/5758 [31:20<02:13,  1.90it/s]
 96%|█████████▌| 5505/5758 [31:21<01:56,  2.17it/s]
                                                   

 96%|█████████▌| 5505/5758 [31:21<01:56,  2.17it/s]
 96%|█████████▌| 5506/5758 [31:21<01:45,  2.38it/s]
                                                   

 96%|█████████▌| 5506/5758 [31:21<01:45,  2.38it/s]
 96%|█████████▌| 5507/5758 [31:21<01:36,  2.60it/s]
                                                   

 96%|█████████▌| 5507/5758 [31:21<01:36,  2.60it/s]
 96%|█████████▌| 5508/5758 [31:22<01:35,  2.63it/s]
                                                   

 96%|█████████▌| 5508/5758 [31:22<01:35,  2.63it/s]
 96%|█████████▌| 5509/5758 [31:22<01:35,  2.62it/s]
                                                   

 96%|█████████▌| 5509/5758 [31:22<01:35,  2.62it/s]
 96%|█████████▌| 5510/5758 [31:22<01:38,  2.51it/s]
                                                   

 96%|█████████▌| 5510/5758 [31:22<01:38,  2.51it/s]
 96%|█████████▌| 5511/5758 [31:23<01:36,  2.57it/s]
                                                   

 96%|���████████▌| 5511/5758 [31:23<01:36,  2.57it/s]
 96%|█████████▌| 5512/5758 [31:23<01:32,  2.65it/s]
                                                   

 96%|█████████▌| 5512/5758 [31:23<01:32,  2.65it/s]
 96%|█████████▌| 5513/5758 [31:23<01:27,  2.81it/s]
                                                   

 96%|█████████▌| 5513/5758 [31:23<01:27,  2.81it/s]
 96%|█████████▌| 5514/5758 [31:24<01:25,  2.87it/s]
                                                   

 96%|█████████▌| 5514/5758 [31:24<01:25,  2.87it/s]
 96%|█████████▌| 5515/5758 [31:24<01:26,  2.80it/s]
                                                   

 96%|█████████▌| 5515/5758 [31:24<01:26,  2.80it/s]
 96%|█████████▌| 5516/5758 [31:24<01:27,  2.77it/s]
                                                   

 96%|█████████▌| 5516/5758 [31:24<01:27,  2.77it/s]
 96%|█████████▌| 5517/5758 [31:25<01:29,  2.68it/s]
                                                   

 96%|█████████▌| 5517/5758 [31:25<01:29,  2.68it/s]
 96%|█████████▌| 5518/5758 [31:25<01:26,  2.76it/s]
                                                   

 96%|█████████▌| 5518/5758 [31:25<01:26,  2.76it/s]
 96%|█████████▌| 5519/5758 [31:26<01:23,  2.88it/s]
                                                   

 96%|█████████▌| 5519/5758 [31:26<01:23,  2.88it/s]
 96%|█████████▌| 5520/5758 [31:26<01:22,  2.88it/s]
                                                   

 96%|█████████▌| 5520/5758 [31:26<01:22,  2.88it/s]
 96%|█████████▌| 5521/5758 [31:26<01:25,  2.76it/s]
                                                   

 96%|█████████▌| 5521/5758 [31:26<01:25,  2.76it/s]
 96%|█████████▌| 5522/5758 [31:27<01:24,  2.81it/s]
                                                   

 96%|█████████▌| 5522/5758 [31:27<01:24,  2.81it/s]
 96%|█████████▌| 5523/5758 [31:27<01:24,  2.77it/s]
                                                   

 96%|█████████▌| 5523/5758 [31:27<01:24,  2.77it/s]
 96%|█████████▌| 5524/5758 [31:27<01:19,  2.93it/s]
                                                   

 96%|█████████▌| 5524/5758 [31:27<01:19,  2.93it/s]
 96%|█████████▌| 5525/5758 [31:28<01:16,  3.06it/s]
                                                   

 96%|█████████▌| 5525/5758 [31:28<01:16,  3.06it/s]
 96%|█████████▌| 5526/5758 [31:28<01:18,  2.97it/s]
                                                   

 96%|█████████▌| 5526/5758 [31:28<01:18,  2.97it/s]
 96%|█████████▌| 5527/5758 [31:28<01:16,  3.03it/s]
                                                   

 96%|█████████▌| 5527/5758 [31:28<01:16,  3.03it/s]
 96%|█████████▌| 5528/5758 [31:29<01:17,  2.98it/s]
                                                   

 96%|█████████▌| 5528/5758 [31:29<01:17,  2.98it/s]
 96%|█████████▌| 5529/5758 [31:29<01:18,  2.92it/s]
                                                   

 96%|█████████▌| 5529/5758 [31:29<01:18,  2.92it/s]
 96%|█████████▌| 5530/5758 [31:29<01:17,  2.93it/s]
                                                   

 96%|█████████▌| 5530/5758 [31:29<01:17,  2.93it/s]
 96%|█████████▌| 5531/5758 [31:30<01:14,  3.04it/s]
                                                   

 96%|█████████▌| 5531/5758 [31:30<01:14,  3.04it/s]
 96%|█████████▌| 5532/5758 [31:30<01:14,  3.05it/s]
                                                   

 96%|█████████▌| 5532/5758 [31:30<01:14,  3.05it/s]
 96%|█████████▌| 5533/5758 [31:30<01:13,  3.07it/s]
                                                   

 96%|█████████▌| 5533/5758 [31:30<01:13,  3.07it/s]
 96%|█████████▌| 5534/5758 [31:31<01:11,  3.13it/s]
                                                   

 96%|█████████▌| 5534/5758 [31:31<01:11,  3.13it/s]
 96%|█████████▌| 5535/5758 [31:31<01:13,  3.04it/s]
                                                   

 96%|█████████▌| 5535/5758 [31:31<01:13,  3.04it/s]
 96%|█████████▌| 5536/5758 [31:31<01:12,  3.05it/s]
                                                   

 96%|█████████▌| 5536/5758 [31:31<01:12,  3.05it/s]
 96%|█████████▌| 5537/5758 [31:32<01:10,  3.14it/s]
                                                   

 96%|█████████▌| 5537/5758 [31:32<01:10,  3.14it/s]
 96%|█████████▌| 5538/5758 [31:32<01:10,  3.12it/s]
                                                   

 96%|█████████▌| 5538/5758 [31:32<01:10,  3.12it/s]
 96%|█████████▌| 5539/5758 [31:32<01:22,  2.65it/s]
                                                   

 96%|█████████▌| 5539/5758 [31:32<01:22,  2.65it/s]
 96%|█████████▌| 5540/5758 [31:33<01:16,  2.84it/s]
                                                   

 96%|█████████▌| 5540/5758 [31:33<01:16,  2.84it/s]
 96%|█████████▌| 5541/5758 [31:33<01:14,  2.92it/s]
                                                   

 96%|█████████▌| 5541/5758 [31:33<01:14,  2.92it/s]
 96%|█████████▌| 5542/5758 [31:33<01:10,  3.08it/s]
                                                   

 96%|█████████▌| 5542/5758 [31:33<01:10,  3.08it/s]
 96%|█████████▋| 5543/5758 [31:34<01:13,  2.93it/s]
                                                   

 96%|█████████▋| 5543/5758 [31:34<01:13,  2.93it/s]
 96%|█████████▋| 5544/5758 [31:34<01:13,  2.89it/s]
                                                   

 96%|█████████▋| 5544/5758 [31:34<01:13,  2.89it/s]
 96%|█████████▋| 5545/5758 [31:34<01:15,  2.81it/s]
                                                   

 96%|█████████▋| 5545/5758 [31:34<01:15,  2.81it/s]
 96%|█████████▋| 5546/5758 [31:35<01:11,  2.99it/s]
                                                   

 96%|█████████▋| 5546/5758 [31:35<01:11,  2.99it/s]
 96%|█████████▋| 5547/5758 [31:35<01:07,  3.10it/s]
                                                   

 96%|█████████▋| 5547/5758 [31:35<01:07,  3.10it/s]
 96%|█████████▋| 5548/5758 [31:35<01:09,  3.02it/s]
                                                   

 96%|█████████▋| 5548/5758 [31:35<01:09,  3.02it/s]
 96%|█████████▋| 5549/5758 [31:36<01:11,  2.93it/s]
                                                   

 96%|█████████▋| 5549/5758 [31:36<01:11,  2.93it/s]
 96%|█████████▋| 5550/5758 [31:36<01:12,  2.86it/s]
                                                   

 96%|█████████▋| 5550/5758 [31:36<01:12,  2.86it/s]
 96%|█████████▋| 5551/5758 [31:36<01:09,  3.00it/s]
                                                   

 96%|█████████▋| 5551/5758 [31:36<01:09,  3.00it/s]
 96%|█████████▋| 5552/5758 [31:37<01:07,  3.04it/s]
                                                   

 96%|█████████▋| 5552/5758 [31:37<01:07,  3.04it/s]
 96%|█████████▋| 5553/5758 [31:37<01:05,  3.11it/s]
                                                   

 96%|█████████▋| 5553/5758 [31:37<01:05,  3.11it/s]
 96%|█████████▋| 5554/5758 [31:37<01:09,  2.95it/s]
                                                   

 96%|█████████▋| 5554/5758 [31:37<01:09,  2.95it/s]
 96%|█████████▋| 5555/5758 [31:38<01:05,  3.08it/s]
                                                   

 96%|█████████▋| 5555/5758 [31:38<01:05,  3.08it/s]
 96%|█████████▋| 5556/5758 [31:38<01:04,  3.14it/s]
                                                   
{'loss': '1.336', 'grad_norm': '17.5', 'learning_rate': '4.481e-07', 'num_tokens': '3.661e+06', 'mean_token_accuracy': '0.6537', 'epoch': '0.9554'}
+{'loss': '2.081', 'grad_norm': '21.5', 'learning_rate': '4.463e-07', 'num_tokens': '3.662e+06', 'mean_token_accuracy': '0.5803', 'epoch': '0.9555'}
+{'loss': '1.176', 'grad_norm': '12.5', 'learning_rate': '4.446e-07', 'num_tokens': '3.662e+06', 'mean_token_accuracy': '0.671', 'epoch': '0.9557'}
+{'loss': '0.8864', 'grad_norm': '18.88', 'learning_rate': '4.429e-07', 'num_tokens': '3.663e+06', 'mean_token_accuracy': '0.81', 'epoch': '0.9559'}
+{'loss': '1.24', 'grad_norm': '13.44', 'learning_rate': '4.411e-07', 'num_tokens': '3.663e+06', 'mean_token_accuracy': '0.6696', 'epoch': '0.9561'}
+{'loss': '1.09', 'grad_norm': '33.25', 'learning_rate': '4.394e-07', 'num_tokens': '3.663e+06', 'mean_token_accuracy': '0.7957', 'epoch': '0.9562'}
+{'loss': '1.929', 'grad_norm': '16.88', 'learning_rate': '4.377e-07', 'num_tokens': '3.664e+06', 'mean_token_accuracy': '0.5988', 'epoch': '0.9564'}
+{'loss': '0.9465', 'grad_norm': '12.81', 'learning_rate': '4.359e-07', 'num_tokens': '3.665e+06', 'mean_token_accuracy': '0.7761', 'epoch': '0.9566'}
+{'loss': '0.7357', 'grad_norm': '33.25', 'learning_rate': '4.342e-07', 'num_tokens': '3.665e+06', 'mean_token_accuracy': '0.7907', 'epoch': '0.9568'}
+{'loss': '1.248', 'grad_norm': '14', 'learning_rate': '4.324e-07', 'num_tokens': '3.665e+06', 'mean_token_accuracy': '0.6591', 'epoch': '0.9569'}
+{'loss': '1.487', 'grad_norm': '32.75', 'learning_rate': '4.307e-07', 'num_tokens': '3.665e+06', 'mean_token_accuracy': '0.6838', 'epoch': '0.9571'}
+{'loss': '1.301', 'grad_norm': '22.12', 'learning_rate': '4.29e-07', 'num_tokens': '3.666e+06', 'mean_token_accuracy': '0.7401', 'epoch': '0.9573'}
+{'loss': '1.048', 'grad_norm': '12.5', 'learning_rate': '4.272e-07', 'num_tokens': '3.666e+06', 'mean_token_accuracy': '0.7168', 'epoch': '0.9575'}
+{'loss': '0.5519', 'grad_norm': '14.31', 'learning_rate': '4.255e-07', 'num_tokens': '3.667e+06', 'mean_token_accuracy': '0.8617', 'epoch': '0.9576'}
+{'loss': '0.6245', 'grad_norm': '8.125', 'learning_rate': '4.238e-07', 'num_tokens': '3.668e+06', 'mean_token_accuracy': '0.8131', 'epoch': '0.9578'}
+{'loss': '1.749', 'grad_norm': '12.62', 'learning_rate': '4.22e-07', 'num_tokens': '3.669e+06', 'mean_token_accuracy': '0.5727', 'epoch': '0.958'}
+{'loss': '2.123', 'grad_norm': '15.62', 'learning_rate': '4.203e-07', 'num_tokens': '3.67e+06', 'mean_token_accuracy': '0.5826', 'epoch': '0.9581'}
+{'loss': '1.187', 'grad_norm': '15.31', 'learning_rate': '4.185e-07', 'num_tokens': '3.67e+06', 'mean_token_accuracy': '0.6821', 'epoch': '0.9583'}
+{'loss': '1.257', 'grad_norm': '16.62', 'learning_rate': '4.168e-07', 'num_tokens': '3.671e+06', 'mean_token_accuracy': '0.7395', 'epoch': '0.9585'}
+{'loss': '1.423', 'grad_norm': '12.81', 'learning_rate': '4.151e-07', 'num_tokens': '3.671e+06', 'mean_token_accuracy': '0.6419', 'epoch': '0.9587'}
+{'loss': '1.089', 'grad_norm': '14', 'learning_rate': '4.133e-07', 'num_tokens': '3.672e+06', 'mean_token_accuracy': '0.7073', 'epoch': '0.9588'}
+{'loss': '0.4438', 'grad_norm': '16.12', 'learning_rate': '4.116e-07', 'num_tokens': '3.672e+06', 'mean_token_accuracy': '0.9175', 'epoch': '0.959'}
+{'loss': '2.077', 'grad_norm': '40', 'learning_rate': '4.099e-07', 'num_tokens': '3.672e+06', 'mean_token_accuracy': '0.6694', 'epoch': '0.9592'}
+{'loss': '1.013', 'grad_norm': '12.31', 'learning_rate': '4.081e-07', 'num_tokens': '3.673e+06', 'mean_token_accuracy': '0.7062', 'epoch': '0.9594'}
+{'loss': '1.587', 'grad_norm': '17.5', 'learning_rate': '4.064e-07', 'num_tokens': '3.673e+06', 'mean_token_accuracy': '0.6809', 'epoch': '0.9595'}
+{'loss': '1.241', 'grad_norm': '19.12', 'learning_rate': '4.047e-07', 'num_tokens': '3.674e+06', 'mean_token_accuracy': '0.7104', 'epoch': '0.9597'}
+{'loss': '0.9957', 'grad_norm': '7.438', 'learning_rate': '4.029e-07', 'num_tokens': '3.675e+06', 'mean_token_accuracy': '0.7684', 'epoch': '0.9599'}
+{'loss': '2.333', 'grad_norm': '7.312', 'learning_rate': '4.012e-07', 'num_tokens': '3.677e+06', 'mean_token_accuracy': '0.5358', 'epoch': '0.9601'}
+{'loss': '1.234', 'grad_norm': '20', 'learning_rate': '3.994e-07', 'num_tokens': '3.677e+06', 'mean_token_accuracy': '0.754', 'epoch': '0.9602'}
+{'loss': '0.9914', 'grad_norm': '7.688', 'learning_rate': '3.977e-07', 'num_tokens': '3.679e+06', 'mean_token_accuracy': '0.7672', 'epoch': '0.9604'}
+{'loss': '1.404', 'grad_norm': '25.25', 'learning_rate': '3.96e-07', 'num_tokens': '3.679e+06', 'mean_token_accuracy': '0.681', 'epoch': '0.9606'}
+{'loss': '0.7144', 'grad_norm': '26', 'learning_rate': '3.942e-07', 'num_tokens': '3.679e+06', 'mean_token_accuracy': '0.871', 'epoch': '0.9608'}
+{'loss': '0.731', 'grad_norm': '16.25', 'learning_rate': '3.925e-07', 'num_tokens': '3.68e+06', 'mean_token_accuracy': '0.7887', 'epoch': '0.9609'}
+{'loss': '1.05', 'grad_norm': '12.06', 'learning_rate': '3.908e-07', 'num_tokens': '3.68e+06', 'mean_token_accuracy': '0.7192', 'epoch': '0.9611'}
+{'loss': '1.04', 'grad_norm': '9.25', 'learning_rate': '3.89e-07', 'num_tokens': '3.681e+06', 'mean_token_accuracy': '0.7596', 'epoch': '0.9613'}
+{'loss': '1.792', 'grad_norm': '9.875', 'learning_rate': '3.873e-07', 'num_tokens': '3.682e+06', 'mean_token_accuracy': '0.6422', 'epoch': '0.9614'}
+{'loss': '0.6757', 'grad_norm': '9.062', 'learning_rate': '3.856e-07', 'num_tokens': '3.683e+06', 'mean_token_accuracy': '0.7828', 'epoch': '0.9616'}
+{'loss': '1.545', 'grad_norm': '28', 'learning_rate': '3.838e-07', 'num_tokens': '3.684e+06', 'mean_token_accuracy': '0.65', 'epoch': '0.9618'}
+{'loss': '1.411', 'grad_norm': '17.38', 'learning_rate': '3.821e-07', 'num_tokens': '3.684e+06', 'mean_token_accuracy': '0.6736', 'epoch': '0.962'}
+{'loss': '1.449', 'grad_norm': '45', 'learning_rate': '3.803e-07', 'num_tokens': '3.684e+06', 'mean_token_accuracy': '0.6986', 'epoch': '0.9621'}
+{'loss': '1.365', 'grad_norm': '12.69', 'learning_rate': '3.786e-07', 'num_tokens': '3.685e+06', 'mean_token_accuracy': '0.6364', 'epoch': '0.9623'}
+{'loss': '1.468', 'grad_norm': '25.38', 'learning_rate': '3.769e-07', 'num_tokens': '3.685e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.9625'}
+{'loss': '0.5687', 'grad_norm': '8.875', 'learning_rate': '3.751e-07', 'num_tokens': '3.686e+06', 'mean_token_accuracy': '0.8183', 'epoch': '0.9627'}
+{'loss': '1.962', 'grad_norm': '19.88', 'learning_rate': '3.734e-07', 'num_tokens': '3.687e+06', 'mean_token_accuracy': '0.5968', 'epoch': '0.9628'}
+{'loss': '1.88', 'grad_norm': '19', 'learning_rate': '3.717e-07', 'num_tokens': '3.687e+06', 'mean_token_accuracy': '0.6261', 'epoch': '0.963'}
+{'loss': '1.457', 'grad_norm': '21.12', 'learning_rate': '3.699e-07', 'num_tokens': '3.688e+06', 'mean_token_accuracy': '0.6459', 'epoch': '0.9632'}
+{'loss': '2.359', 'grad_norm': '37.75', 'learning_rate': '3.682e-07', 'num_tokens': '3.688e+06', 'mean_token_accuracy': '0.5188', 'epoch': '0.9634'}
+{'loss': '1.371', 'grad_norm': '22', 'learning_rate': '3.664e-07', 'num_tokens': '3.688e+06', 'mean_token_accuracy': '0.7008', 'epoch': '0.9635'}
+{'loss': '1.188', 'grad_norm': '5.125', 'learning_rate': '3.647e-07', 'num_tokens': '3.692e+06', 'mean_token_accuracy': '0.6894', 'epoch': '0.9637'}
+{'loss': '1.303', 'grad_norm': '19', 'learning_rate': '3.63e-07', 'num_tokens': '3.692e+06', 'mean_token_accuracy': '0.7179', 'epoch': '0.9639'}
+{'loss': '1.211', 'grad_norm': '13.5', 'learning_rate': '3.612e-07', 'num_tokens': '3.693e+06', 'mean_token_accuracy': '0.6626', 'epoch': '0.9641'}
+{'loss': '0.6726', 'grad_norm': '29.25', 'learning_rate': '3.595e-07', 'num_tokens': '3.693e+06', 'mean_token_accuracy': '0.8238', 'epoch': '0.9642'}
+{'loss': '1.338', 'grad_norm': '46', 'learning_rate': '3.578e-07', 'num_tokens': '3.693e+06', 'mean_token_accuracy': '0.7037', 'epoch': '0.9644'}
+{'loss': '2.259', 'grad_norm': '9.125', 'learning_rate': '3.56e-07', 'num_tokens': '3.694e+06', 'mean_token_accuracy': '0.5299', 'epoch': '0.9646'}
+{'loss': '1.013', 'grad_norm': '13.75', 'learning_rate': '3.543e-07', 'num_tokens': '3.695e+06', 'mean_token_accuracy': '0.7118', 'epoch': '0.9647'}
+
 96%|█████████▋| 5556/5758 [31:38<01:04,  3.14it/s]
 97%|█████████▋| 5557/5758 [31:38<01:01,  3.25it/s]
                                                   

 97%|█████████▋| 5557/5758 [31:38<01:01,  3.25it/s]
 97%|█████████▋| 5558/5758 [31:39<01:02,  3.18it/s]
                                                   

 97%|█████████▋| 5558/5758 [31:39<01:02,  3.18it/s]
 97%|█████████▋| 5559/5758 [31:39<01:01,  3.25it/s]
                                                   

 97%|█████████▋| 5559/5758 [31:39<01:01,  3.25it/s]
 97%|█████████▋| 5560/5758 [31:39<01:00,  3.30it/s]
                                                   

 97%|█████████▋| 5560/5758 [31:39<01:00,  3.30it/s]
 97%|█████████▋| 5561/5758 [31:39<01:03,  3.09it/s]
                                                   

 97%|█████████▋| 5561/5758 [31:39<01:03,  3.09it/s]
 97%|█████████▋| 5562/5758 [31:40<01:06,  2.94it/s]
                                                   

 97%|█████████▋| 5562/5758 [31:40<01:06,  2.94it/s]
 97%|█████████▋| 5563/5758 [31:40<01:04,  3.02it/s]
                                                   

 97%|█████████▋| 5563/5758 [31:40<01:04,  3.02it/s]
 97%|█████████▋| 5564/5758 [31:41<01:05,  2.95it/s]
                                                   

 97%|█████████▋| 5564/5758 [31:41<01:05,  2.95it/s]
 97%|█████████▋| 5565/5758 [31:41<01:12,  2.65it/s]
                                                   

 97%|█████████▋| 5565/5758 [31:41<01:12,  2.65it/s]
 97%|█████████▋| 5566/5758 [31:41<01:07,  2.84it/s]
                                                   

 97%|█████████▋| 5566/5758 [31:41<01:07,  2.84it/s]
 97%|█████████▋| 5567/5758 [31:42<01:05,  2.90it/s]
                                                   

 97%|█████████▋| 5567/5758 [31:42<01:05,  2.90it/s]
 97%|█████████▋| 5568/5758 [31:42<01:04,  2.95it/s]
                                                   

 97%|█████████▋| 5568/5758 [31:42<01:04,  2.95it/s]
 97%|█████████▋| 5569/5758 [31:42<01:05,  2.89it/s]
                                                   

 97%|█████████▋| 5569/5758 [31:42<01:05,  2.89it/s]
 97%|█████████▋| 5570/5758 [31:43<01:04,  2.89it/s]
                                                   

 97%|█████████▋| 5570/5758 [31:43<01:04,  2.89it/s]
 97%|█████████▋| 5571/5758 [31:43<01:05,  2.85it/s]
                                                   

 97%|█████████▋| 5571/5758 [31:43<01:05,  2.85it/s]
 97%|█████████▋| 5572/5758 [31:43<01:03,  2.94it/s]
                                                   

 97%|█████████▋| 5572/5758 [31:43<01:03,  2.94it/s]
 97%|█████████▋| 5573/5758 [31:44<01:00,  3.04it/s]
                                                   

 97%|█████████▋| 5573/5758 [31:44<01:00,  3.04it/s]
 97%|��████████▋| 5574/5758 [31:44<00:59,  3.08it/s]
                                                   

 97%|█████████▋| 5574/5758 [31:44<00:59,  3.08it/s]
 97%|█████████▋| 5575/5758 [31:44<00:57,  3.18it/s]
                                                   

 97%|█████████▋| 5575/5758 [31:44<00:57,  3.18it/s]
 97%|█████████▋| 5576/5758 [31:45<01:00,  3.01it/s]
                                                   

 97%|█████████▋| 5576/5758 [31:45<01:00,  3.01it/s]
 97%|█████████▋| 5577/5758 [31:45<00:58,  3.08it/s]
                                                   

 97%|█████████▋| 5577/5758 [31:45<00:58,  3.08it/s]
 97%|█████████▋| 5578/5758 [31:45<00:59,  3.01it/s]
                                                   

 97%|█████████▋| 5578/5758 [31:45<00:59,  3.01it/s]
 97%|█████████▋| 5579/5758 [31:46<00:57,  3.12it/s]
                                                   

 97%|█████████▋| 5579/5758 [31:46<00:57,  3.12it/s]
 97%|█████████▋| 5580/5758 [31:46<00:59,  2.97it/s]
                                                   

 97%|█████████▋| 5580/5758 [31:46<00:59,  2.97it/s]
 97%|█████████▋| 5581/5758 [31:46<01:01,  2.86it/s]
                                                   

 97%|█████████▋| 5581/5758 [31:46<01:01,  2.86it/s]
 97%|█████████▋| 5582/5758 [31:47<01:03,  2.76it/s]
                                                   

 97%|█████████▋| 5582/5758 [31:47<01:03,  2.76it/s]
 97%|█████████▋| 5583/5758 [31:47<01:00,  2.89it/s]
                                                   

 97%|█████████▋| 5583/5758 [31:47<01:00,  2.89it/s]
 97%|█████████▋| 5584/5758 [31:47<00:57,  3.03it/s]
                                                   

 97%|█████████▋| 5584/5758 [31:47<00:57,  3.03it/s]
 97%|█████████▋| 5585/5758 [31:48<00:54,  3.16it/s]
                                                   

 97%|█████████▋| 5585/5758 [31:48<00:54,  3.16it/s]
 97%|█████████▋| 5586/5758 [31:48<00:53,  3.20it/s]
                                                   

 97%|█████████▋| 5586/5758 [31:48<00:53,  3.20it/s]
 97%|█████████▋| 5587/5758 [31:48<00:57,  2.99it/s]
                                                   

 97%|█████████▋| 5587/5758 [31:48<00:57,  2.99it/s]
 97%|█████████▋| 5588/5758 [31:49<00:58,  2.92it/s]
                                                   

 97%|█████████▋| 5588/5758 [31:49<00:58,  2.92it/s]
 97%|█████████▋| 5589/5758 [31:49<00:56,  3.00it/s]
                                                   

 97%|█████████▋| 5589/5758 [31:49<00:56,  3.00it/s]
 97%|█████████▋| 5590/5758 [31:49<00:53,  3.12it/s]
                                                   

 97%|█████████▋| 5590/5758 [31:49<00:53,  3.12it/s]
 97%|█████████▋| 5591/5758 [31:50<00:59,  2.80it/s]
                                                   

 97%|█████████▋| 5591/5758 [31:50<00:59,  2.80it/s]
 97%|█████████▋| 5592/5758 [31:50<00:57,  2.91it/s]
                                                   

 97%|█████████▋| 5592/5758 [31:50<00:57,  2.91it/s]
 97%|█████████▋| 5593/5758 [31:50<00:57,  2.86it/s]
                                                   

 97%|█████████▋| 5593/5758 [31:50<00:57,  2.86it/s]
 97%|█████████▋| 5594/5758 [31:51<00:58,  2.82it/s]
                                                   

 97%|█████████▋| 5594/5758 [31:51<00:58,  2.82it/s]
 97%|█████████▋| 5595/5758 [31:51<00:59,  2.73it/s]
                                                   

 97%|█████████▋| 5595/5758 [31:51<00:59,  2.73it/s]
 97%|█████████▋| 5596/5758 [31:51<00:59,  2.74it/s]
                                                   

 97%|█████████▋| 5596/5758 [31:51<00:59,  2.74it/s]
 97%|█████████▋| 5597/5758 [31:52<00:59,  2.69it/s]
                                                   

 97%|█████████▋| 5597/5758 [31:52<00:59,  2.69it/s]
 97%|█████████▋| 5598/5758 [31:52<00:57,  2.77it/s]
                                                   

 97%|█████████▋| 5598/5758 [31:52<00:57,  2.77it/s]
 97%|█████████▋| 5599/5758 [31:53<00:58,  2.72it/s]
                                                   

 97%|█████████▋| 5599/5758 [31:53<00:58,  2.72it/s]
 97%|█████████▋| 5600/5758 [31:53<00:53,  2.97it/s]
                                                   

 97%|█████████▋| 5600/5758 [31:53<00:53,  2.97it/s]
 97%|█████████▋| 5601/5758 [31:53<00:52,  3.01it/s]
                                                   

 97%|█████████▋| 5601/5758 [31:53<00:52,  3.01it/s]
 97%|█████████▋| 5602/5758 [31:54<00:53,  2.92it/s]
                                                   

 97%|█████████▋| 5602/5758 [31:54<00:53,  2.92it/s]
 97%|█████████▋| 5603/5758 [31:54<00:51,  3.02it/s]
                                                   

 97%|█████████▋| 5603/5758 [31:54<00:51,  3.02it/s]
 97%|█████████▋| 5604/5758 [31:54<00:49,  3.08it/s]
                                                   

 97%|█████████▋| 5604/5758 [31:54<00:49,  3.08it/s]
 97%|█████████▋| 5605/5758 [31:55<00:51,  2.99it/s]
                                                   

 97%|█████████▋| 5605/5758 [31:55<00:51,  2.99it/s]
 97%|█████████▋| 5606/5758 [31:55<00:49,  3.07it/s]
                                                   

 97%|█████████▋| 5606/5758 [31:55<00:49,  3.07it/s]
 97%|█████████▋| 5607/5758 [31:55<00:47,  3.15it/s]
                                                   

 97%|█████████▋| 5607/5758 [31:55<00:47,  3.15it/s]
 97%|█████████▋| 5608/5758 [31:55<00:47,  3.15it/s]
                                                   

 97%|█████████▋| 5608/5758 [31:55<00:47,  3.15it/s]
 97%|█████████▋| 5609/5758 [31:56<00:46,  3.18it/s]
                                                   

 97%|█████████▋| 5609/5758 [31:56<00:46,  3.18it/s]
 97%|█████████▋| 5610/5758 [31:56<00:45,  3.25it/s]
                                                   

 97%|█████████▋| 5610/5758 [31:56<00:45,  3.25it/s]
 97%|█████████▋| 5611/5758 [31:56<00:45,  3.24it/s]
                                                   
{'loss': '0.9797', 'grad_norm': '11.88', 'learning_rate': '3.526e-07', 'num_tokens': '3.696e+06', 'mean_token_accuracy': '0.734', 'epoch': '0.9649'}
+{'loss': '1.376', 'grad_norm': '37.75', 'learning_rate': '3.508e-07', 'num_tokens': '3.696e+06', 'mean_token_accuracy': '0.7237', 'epoch': '0.9651'}
+{'loss': '1.973', 'grad_norm': '15', 'learning_rate': '3.491e-07', 'num_tokens': '3.696e+06', 'mean_token_accuracy': '0.5804', 'epoch': '0.9653'}
+{'loss': '0.5862', 'grad_norm': '7.625', 'learning_rate': '3.473e-07', 'num_tokens': '3.698e+06', 'mean_token_accuracy': '0.8176', 'epoch': '0.9654'}
+{'loss': '1.356', 'grad_norm': '14.38', 'learning_rate': '3.456e-07', 'num_tokens': '3.698e+06', 'mean_token_accuracy': '0.6926', 'epoch': '0.9656'}
+{'loss': '1.41', 'grad_norm': '27.25', 'learning_rate': '3.439e-07', 'num_tokens': '3.698e+06', 'mean_token_accuracy': '0.664', 'epoch': '0.9658'}
+{'loss': '0.6566', 'grad_norm': '8.625', 'learning_rate': '3.421e-07', 'num_tokens': '3.699e+06', 'mean_token_accuracy': '0.8098', 'epoch': '0.966'}
+{'loss': '1.23', 'grad_norm': '11.75', 'learning_rate': '3.404e-07', 'num_tokens': '3.7e+06', 'mean_token_accuracy': '0.648', 'epoch': '0.9661'}
+{'loss': '1.216', 'grad_norm': '10.5', 'learning_rate': '3.387e-07', 'num_tokens': '3.701e+06', 'mean_token_accuracy': '0.6738', 'epoch': '0.9663'}
+{'loss': '1.478', 'grad_norm': '16.62', 'learning_rate': '3.369e-07', 'num_tokens': '3.702e+06', 'mean_token_accuracy': '0.6496', 'epoch': '0.9665'}
+{'loss': '1.643', 'grad_norm': '15.12', 'learning_rate': '3.352e-07', 'num_tokens': '3.702e+06', 'mean_token_accuracy': '0.6531', 'epoch': '0.9667'}
+{'loss': '1.35', 'grad_norm': '17.88', 'learning_rate': '3.334e-07', 'num_tokens': '3.703e+06', 'mean_token_accuracy': '0.7446', 'epoch': '0.9668'}
+{'loss': '1.422', 'grad_norm': '17.25', 'learning_rate': '3.317e-07', 'num_tokens': '3.703e+06', 'mean_token_accuracy': '0.6978', 'epoch': '0.967'}
+{'loss': '0.6411', 'grad_norm': '10.88', 'learning_rate': '3.3e-07', 'num_tokens': '3.704e+06', 'mean_token_accuracy': '0.8451', 'epoch': '0.9672'}
+{'loss': '1.37', 'grad_norm': '12.25', 'learning_rate': '3.282e-07', 'num_tokens': '3.705e+06', 'mean_token_accuracy': '0.6218', 'epoch': '0.9673'}
+{'loss': '1.764', 'grad_norm': '17.5', 'learning_rate': '3.265e-07', 'num_tokens': '3.705e+06', 'mean_token_accuracy': '0.605', 'epoch': '0.9675'}
+{'loss': '1.048', 'grad_norm': '11.44', 'learning_rate': '3.248e-07', 'num_tokens': '3.706e+06', 'mean_token_accuracy': '0.7201', 'epoch': '0.9677'}
+{'loss': '1.338', 'grad_norm': '6.781', 'learning_rate': '3.23e-07', 'num_tokens': '3.709e+06', 'mean_token_accuracy': '0.6883', 'epoch': '0.9679'}
+{'loss': '1.223', 'grad_norm': '11.94', 'learning_rate': '3.213e-07', 'num_tokens': '3.71e+06', 'mean_token_accuracy': '0.6633', 'epoch': '0.968'}
+{'loss': '1.061', 'grad_norm': '14.88', 'learning_rate': '3.196e-07', 'num_tokens': '3.71e+06', 'mean_token_accuracy': '0.7143', 'epoch': '0.9682'}
+{'loss': '1.926', 'grad_norm': '18.75', 'learning_rate': '3.178e-07', 'num_tokens': '3.711e+06', 'mean_token_accuracy': '0.5979', 'epoch': '0.9684'}
+{'loss': '1.13', 'grad_norm': '13.31', 'learning_rate': '3.161e-07', 'num_tokens': '3.711e+06', 'mean_token_accuracy': '0.68', 'epoch': '0.9686'}
+{'loss': '0.9052', 'grad_norm': '20.75', 'learning_rate': '3.143e-07', 'num_tokens': '3.712e+06', 'mean_token_accuracy': '0.8107', 'epoch': '0.9687'}
+{'loss': '2.279', 'grad_norm': '17.88', 'learning_rate': '3.126e-07', 'num_tokens': '3.712e+06', 'mean_token_accuracy': '0.5511', 'epoch': '0.9689'}
+{'loss': '2.191', 'grad_norm': '37.25', 'learning_rate': '3.109e-07', 'num_tokens': '3.712e+06', 'mean_token_accuracy': '0.5417', 'epoch': '0.9691'}
+{'loss': '0.6948', 'grad_norm': '12.62', 'learning_rate': '3.091e-07', 'num_tokens': '3.713e+06', 'mean_token_accuracy': '0.8369', 'epoch': '0.9693'}
+{'loss': '1.116', 'grad_norm': '9.062', 'learning_rate': '3.074e-07', 'num_tokens': '3.714e+06', 'mean_token_accuracy': '0.7246', 'epoch': '0.9694'}
+{'loss': '1.626', 'grad_norm': '13.25', 'learning_rate': '3.057e-07', 'num_tokens': '3.715e+06', 'mean_token_accuracy': '0.6403', 'epoch': '0.9696'}
+{'loss': '1.213', 'grad_norm': '13.25', 'learning_rate': '3.039e-07', 'num_tokens': '3.716e+06', 'mean_token_accuracy': '0.6362', 'epoch': '0.9698'}
+{'loss': '0.6378', 'grad_norm': '9.812', 'learning_rate': '3.022e-07', 'num_tokens': '3.717e+06', 'mean_token_accuracy': '0.797', 'epoch': '0.97'}
+{'loss': '1.125', 'grad_norm': '13', 'learning_rate': '3.005e-07', 'num_tokens': '3.717e+06', 'mean_token_accuracy': '0.6948', 'epoch': '0.9701'}
+{'loss': '2.123', 'grad_norm': '15.69', 'learning_rate': '2.987e-07', 'num_tokens': '3.718e+06', 'mean_token_accuracy': '0.5809', 'epoch': '0.9703'}
+{'loss': '1.883', 'grad_norm': '10.38', 'learning_rate': '2.97e-07', 'num_tokens': '3.719e+06', 'mean_token_accuracy': '0.6298', 'epoch': '0.9705'}
+{'loss': '1.218', 'grad_norm': '18.5', 'learning_rate': '2.952e-07', 'num_tokens': '3.719e+06', 'mean_token_accuracy': '0.7283', 'epoch': '0.9706'}
+{'loss': '0.8574', 'grad_norm': '32.75', 'learning_rate': '2.935e-07', 'num_tokens': '3.719e+06', 'mean_token_accuracy': '0.8382', 'epoch': '0.9708'}
+{'loss': '1.962', 'grad_norm': '41.5', 'learning_rate': '2.918e-07', 'num_tokens': '3.72e+06', 'mean_token_accuracy': '0.5556', 'epoch': '0.971'}
+{'loss': '1.036', 'grad_norm': '12.56', 'learning_rate': '2.9e-07', 'num_tokens': '3.72e+06', 'mean_token_accuracy': '0.7145', 'epoch': '0.9712'}
+{'loss': '1.289', 'grad_norm': '14.88', 'learning_rate': '2.883e-07', 'num_tokens': '3.721e+06', 'mean_token_accuracy': '0.6692', 'epoch': '0.9713'}
+{'loss': '1.518', 'grad_norm': '15.12', 'learning_rate': '2.866e-07', 'num_tokens': '3.721e+06', 'mean_token_accuracy': '0.6781', 'epoch': '0.9715'}
+{'loss': '2.1', 'grad_norm': '14', 'learning_rate': '2.848e-07', 'num_tokens': '3.722e+06', 'mean_token_accuracy': '0.6081', 'epoch': '0.9717'}
+{'loss': '1.731', 'grad_norm': '17.12', 'learning_rate': '2.831e-07', 'num_tokens': '3.722e+06', 'mean_token_accuracy': '0.6423', 'epoch': '0.9719'}
+{'loss': '1.008', 'grad_norm': '13.06', 'learning_rate': '2.813e-07', 'num_tokens': '3.723e+06', 'mean_token_accuracy': '0.7172', 'epoch': '0.972'}
+{'loss': '1.505', 'grad_norm': '36.5', 'learning_rate': '2.796e-07', 'num_tokens': '3.723e+06', 'mean_token_accuracy': '0.6458', 'epoch': '0.9722'}
+{'loss': '2.163', 'grad_norm': '20.75', 'learning_rate': '2.779e-07', 'num_tokens': '3.723e+06', 'mean_token_accuracy': '0.5543', 'epoch': '0.9724'}
+{'loss': '0.8822', 'grad_norm': '40.25', 'learning_rate': '2.761e-07', 'num_tokens': '3.723e+06', 'mean_token_accuracy': '0.8333', 'epoch': '0.9726'}
+{'loss': '1.002', 'grad_norm': '20.38', 'learning_rate': '2.744e-07', 'num_tokens': '3.724e+06', 'mean_token_accuracy': '0.7732', 'epoch': '0.9727'}
+{'loss': '0.7363', 'grad_norm': '20.25', 'learning_rate': '2.727e-07', 'num_tokens': '3.724e+06', 'mean_token_accuracy': '0.8154', 'epoch': '0.9729'}
+{'loss': '1.937', 'grad_norm': '15.94', 'learning_rate': '2.709e-07', 'num_tokens': '3.725e+06', 'mean_token_accuracy': '0.6085', 'epoch': '0.9731'}
+{'loss': '0.8683', 'grad_norm': '11.25', 'learning_rate': '2.692e-07', 'num_tokens': '3.725e+06', 'mean_token_accuracy': '0.7424', 'epoch': '0.9733'}
+{'loss': '1.995', 'grad_norm': '6.75', 'learning_rate': '2.675e-07', 'num_tokens': '3.727e+06', 'mean_token_accuracy': '0.5915', 'epoch': '0.9734'}
+{'loss': '1.252', 'grad_norm': '13.25', 'learning_rate': '2.657e-07', 'num_tokens': '3.728e+06', 'mean_token_accuracy': '0.6514', 'epoch': '0.9736'}
+{'loss': '1.217', 'grad_norm': '13.31', 'learning_rate': '2.64e-07', 'num_tokens': '3.729e+06', 'mean_token_accuracy': '0.678', 'epoch': '0.9738'}
+{'loss': '1.187', 'grad_norm': '6.281', 'learning_rate': '2.622e-07', 'num_tokens': '3.732e+06', 'mean_token_accuracy': '0.7251', 'epoch': '0.9739'}
+{'loss': '1.32', 'grad_norm': '11.88', 'learning_rate': '2.605e-07', 'num_tokens': '3.733e+06', 'mean_token_accuracy': '0.6426', 'epoch': '0.9741'}
+{'loss': '1.078', 'grad_norm': '12.56', 'learning_rate': '2.588e-07', 'num_tokens': '3.734e+06', 'mean_token_accuracy': '0.7151', 'epoch': '0.9743'}
+
 97%|█████████▋| 5611/5758 [31:56<00:45,  3.24it/s]
 97%|█████████▋| 5612/5758 [31:57<00:45,  3.21it/s]
                                                   

 97%|█████████▋| 5612/5758 [31:57<00:45,  3.21it/s]
 97%|█████████▋| 5613/5758 [31:57<00:43,  3.30it/s]
                                                   

 97%|█████████▋| 5613/5758 [31:57<00:43,  3.30it/s]
 97%|█████████▋| 5614/5758 [31:57<00:43,  3.34it/s]
                                                   

 97%|█████████▋| 5614/5758 [31:57<00:43,  3.34it/s]
 98%|█████████▊| 5615/5758 [31:58<00:45,  3.14it/s]
                                                   

 98%|█████████▊| 5615/5758 [31:58<00:45,  3.14it/s]
 98%|█████████▊| 5616/5758 [31:58<00:44,  3.20it/s]
                                                   

 98%|█████████▊| 5616/5758 [31:58<00:44,  3.20it/s]
 98%|█████████▊| 5617/5758 [31:58<00:48,  2.90it/s]
                                                   

 98%|█████████▊| 5617/5758 [31:58<00:48,  2.90it/s]
 98%|█████████▊| 5618/5758 [31:59<00:48,  2.87it/s]
                                                   

 98%|█████████▊| 5618/5758 [31:59<00:48,  2.87it/s]
 98%|█████████▊| 5619/5758 [31:59<00:49,  2.80it/s]
                                                   

 98%|█████████▊| 5619/5758 [31:59<00:49,  2.80it/s]
 98%|█████████▊| 5620/5758 [31:59<00:48,  2.87it/s]
                                                   

 98%|█████████▊| 5620/5758 [31:59<00:48,  2.87it/s]
 98%|█████████▊| 5621/5758 [32:00<00:45,  3.04it/s]
                                                   

 98%|█████████▊| 5621/5758 [32:00<00:45,  3.04it/s]
 98%|█████████▊| 5622/5758 [32:00<00:43,  3.12it/s]
                                                   

 98%|█████████▊| 5622/5758 [32:00<00:43,  3.12it/s]
 98%|█████████▊| 5623/5758 [32:00<00:43,  3.07it/s]
                                                   

 98%|█████████▊| 5623/5758 [32:00<00:43,  3.07it/s]
 98%|█████████▊| 5624/5758 [32:01<00:44,  3.01it/s]
                                                   

 98%|█████████▊| 5624/5758 [32:01<00:44,  3.01it/s]
 98%|█████████▊| 5625/5758 [32:01<00:43,  3.05it/s]
                                                   

 98%|█████████▊| 5625/5758 [32:01<00:43,  3.05it/s]
 98%|█████████▊| 5626/5758 [32:01<00:42,  3.08it/s]
                                                   

 98%|█████████▊| 5626/5758 [32:01<00:42,  3.08it/s]
 98%|█████████▊| 5627/5758 [32:02<00:43,  3.04it/s]
                                                   

 98%|█████████▊| 5627/5758 [32:02<00:43,  3.04it/s]
 98%|█████████▊| 5628/5758 [32:02<00:43,  3.01it/s]
                                                   

 98%|█████████▊| 5628/5758 [32:02<00:43,  3.01it/s]
 98%|█████████▊| 5629/5758 [32:02<00:41,  3.09it/s]
                                                   

 98%|█████████▊| 5629/5758 [32:02<00:41,  3.09it/s]
 98%|█████████▊| 5630/5758 [32:03<00:40,  3.17it/s]
                                                   

 98%|█████████▊| 5630/5758 [32:03<00:40,  3.17it/s]
 98%|█████████▊| 5631/5758 [32:03<00:42,  2.96it/s]
                                                   

 98%|█████████▊| 5631/5758 [32:03<00:42,  2.96it/s]
 98%|█████████▊| 5632/5758 [32:03<00:41,  3.03it/s]
                                                   

 98%|█████████▊| 5632/5758 [32:03<00:41,  3.03it/s]
 98%|█████████▊| 5633/5758 [32:04<00:40,  3.08it/s]
                                                   

 98%|█████████▊| 5633/5758 [32:04<00:40,  3.08it/s]
 98%|█████████▊| 5634/5758 [32:04<00:39,  3.18it/s]
                                                   

 98%|█████████▊| 5634/5758 [32:04<00:39,  3.18it/s]
 98%|█████████▊| 5635/5758 [32:04<00:38,  3.17it/s]
                                                   

 98%|█████████▊| 5635/5758 [32:04<00:38,  3.17it/s]
 98%|█████████▊| 5636/5758 [32:04<00:37,  3.22it/s]
                                                   

 98%|█████████▊| 5636/5758 [32:04<00:37,  3.22it/s]
 98%|█████████▊| 5637/5758 [32:05<00:36,  3.29it/s]
                                                   

 98%|█████████▊| 5637/5758 [32:05<00:36,  3.29it/s]
 98%|█████████▊| 5638/5758 [32:05<00:37,  3.17it/s]
                                                   

 98%|█████████▊| 5638/5758 [32:05<00:37,  3.17it/s]
 98%|█████████▊| 5639/5758 [32:05<00:36,  3.23it/s]
                                                   

 98%|█████████▊| 5639/5758 [32:05<00:36,  3.23it/s]
 98%|█████████▊| 5640/5758 [32:06<00:36,  3.22it/s]
                                                   

 98%|█████████▊| 5640/5758 [32:06<00:36,  3.22it/s]
 98%|█████████▊| 5641/5758 [32:06<00:36,  3.19it/s]
                                                   

 98%|█████████▊| 5641/5758 [32:06<00:36,  3.19it/s]
 98%|█████████▊| 5642/5758 [32:06<00:36,  3.14it/s]
                                                   

 98%|█████████▊| 5642/5758 [32:06<00:36,  3.14it/s]
 98%|█████████▊| 5643/5758 [32:07<00:43,  2.67it/s]
                                                   

 98%|█████████▊| 5643/5758 [32:07<00:43,  2.67it/s]
 98%|█████████▊| 5644/5758 [32:07<00:40,  2.82it/s]
                                                   

 98%|█████████▊| 5644/5758 [32:07<00:40,  2.82it/s]
 98%|█████████▊| 5645/5758 [32:08<00:40,  2.76it/s]
                                                   

 98%|█████████▊| 5645/5758 [32:08<00:40,  2.76it/s]
 98%|█████████▊| 5646/5758 [32:08<00:40,  2.79it/s]
                                                   

 98%|█████████▊| 5646/5758 [32:08<00:40,  2.79it/s]
 98%|█████████▊| 5647/5758 [32:08<00:37,  2.96it/s]
                                                   

 98%|█████████▊| 5647/5758 [32:08<00:37,  2.96it/s]
 98%|█████████▊| 5648/5758 [32:09<00:37,  2.90it/s]
                                                   

 98%|█████████▊| 5648/5758 [32:09<00:37,  2.90it/s]
 98%|█████████▊| 5649/5758 [32:09<00:35,  3.05it/s]
                                                   

 98%|█████████▊| 5649/5758 [32:09<00:35,  3.05it/s]
 98%|█████████▊| 5650/5758 [32:09<00:34,  3.14it/s]
                                                   

 98%|█████████▊| 5650/5758 [32:09<00:34,  3.14it/s]
 98%|█████████▊| 5651/5758 [32:09<00:33,  3.21it/s]
                                                   

 98%|█████████▊| 5651/5758 [32:09<00:33,  3.21it/s]
 98%|█████████▊| 5652/5758 [32:10<00:33,  3.14it/s]
                                                   

 98%|█████████▊| 5652/5758 [32:10<00:33,  3.14it/s]
 98%|█████████▊| 5653/5758 [32:10<00:34,  3.02it/s]
                                                   

 98%|█████████▊| 5653/5758 [32:10<00:34,  3.02it/s]
 98%|█████████▊| 5654/5758 [32:10<00:34,  3.05it/s]
                                                   

 98%|█████████▊| 5654/5758 [32:10<00:34,  3.05it/s]
 98%|█████████▊| 5655/5758 [32:11<00:33,  3.09it/s]
                                                   

 98%|█████████▊| 5655/5758 [32:11<00:33,  3.09it/s]
 98%|█████████▊| 5656/5758 [32:11<00:34,  2.97it/s]
                                                   

 98%|█████████▊| 5656/5758 [32:11<00:34,  2.97it/s]
 98%|█████████▊| 5657/5758 [32:11<00:32,  3.07it/s]
                                                   

 98%|█████████▊| 5657/5758 [32:11<00:32,  3.07it/s]
 98%|█████████▊| 5658/5758 [32:12<00:34,  2.94it/s]
                                                   

 98%|█████████▊| 5658/5758 [32:12<00:34,  2.94it/s]
 98%|█████████▊| 5659/5758 [32:12<00:34,  2.84it/s]
                                                   

 98%|█████████▊| 5659/5758 [32:12<00:34,  2.84it/s]
 98%|█████████▊| 5660/5758 [32:13<00:34,  2.83it/s]
                                                   

 98%|█████████▊| 5660/5758 [32:13<00:34,  2.83it/s]
 98%|█████████▊| 5661/5758 [32:13<00:35,  2.75it/s]
                                                   

 98%|█████████▊| 5661/5758 [32:13<00:35,  2.75it/s]
 98%|█████████▊| 5662/5758 [32:13<00:34,  2.77it/s]
                                                   

 98%|█████████▊| 5662/5758 [32:13<00:34,  2.77it/s]
 98%|█████████▊| 5663/5758 [32:14<00:32,  2.91it/s]
                                                   

 98%|█████████▊| 5663/5758 [32:14<00:32,  2.91it/s]
 98%|█████████▊| 5664/5758 [32:14<00:31,  2.97it/s]
                                                   

 98%|█████████▊| 5664/5758 [32:14<00:31,  2.97it/s]
 98%|█████████▊| 5665/5758 [32:14<00:30,  3.04it/s]
                                                   

 98%|█████████▊| 5665/5758 [32:14<00:30,  3.04it/s]
 98%|█████████▊| 5666/5758 [32:15<00:29,  3.08it/s]
                                                   
{'loss': '0.8542', 'grad_norm': '12.19', 'learning_rate': '2.57e-07', 'num_tokens': '3.735e+06', 'mean_token_accuracy': '0.7438', 'epoch': '0.9745'}
+{'loss': '1.596', 'grad_norm': '12.12', 'learning_rate': '2.553e-07', 'num_tokens': '3.736e+06', 'mean_token_accuracy': '0.635', 'epoch': '0.9746'}
+{'loss': '1.27', 'grad_norm': '18.5', 'learning_rate': '2.536e-07', 'num_tokens': '3.736e+06', 'mean_token_accuracy': '0.7096', 'epoch': '0.9748'}
+{'loss': '1.219', 'grad_norm': '12.62', 'learning_rate': '2.518e-07', 'num_tokens': '3.737e+06', 'mean_token_accuracy': '0.6794', 'epoch': '0.975'}
+{'loss': '1.035', 'grad_norm': '8.312', 'learning_rate': '2.501e-07', 'num_tokens': '3.738e+06', 'mean_token_accuracy': '0.7693', 'epoch': '0.9752'}
+{'loss': '1.489', 'grad_norm': '42.5', 'learning_rate': '2.484e-07', 'num_tokens': '3.738e+06', 'mean_token_accuracy': '0.7308', 'epoch': '0.9753'}
+{'loss': '0.9688', 'grad_norm': '17.5', 'learning_rate': '2.466e-07', 'num_tokens': '3.738e+06', 'mean_token_accuracy': '0.7756', 'epoch': '0.9755'}
+{'loss': '2.02', 'grad_norm': '21.38', 'learning_rate': '2.449e-07', 'num_tokens': '3.739e+06', 'mean_token_accuracy': '0.6195', 'epoch': '0.9757'}
+{'loss': '0.5598', 'grad_norm': '9.25', 'learning_rate': '2.431e-07', 'num_tokens': '3.74e+06', 'mean_token_accuracy': '0.8336', 'epoch': '0.9759'}
+{'loss': '1.323', 'grad_norm': '29.25', 'learning_rate': '2.414e-07', 'num_tokens': '3.74e+06', 'mean_token_accuracy': '0.7447', 'epoch': '0.976'}
+{'loss': '0.8934', 'grad_norm': '23.75', 'learning_rate': '2.397e-07', 'num_tokens': '3.74e+06', 'mean_token_accuracy': '0.8013', 'epoch': '0.9762'}
+{'loss': '1.049', 'grad_norm': '12.81', 'learning_rate': '2.379e-07', 'num_tokens': '3.741e+06', 'mean_token_accuracy': '0.7054', 'epoch': '0.9764'}
+{'loss': '1.951', 'grad_norm': '7.469', 'learning_rate': '2.362e-07', 'num_tokens': '3.742e+06', 'mean_token_accuracy': '0.6104', 'epoch': '0.9766'}
+{'loss': '0.7581', 'grad_norm': '11.06', 'learning_rate': '2.345e-07', 'num_tokens': '3.743e+06', 'mean_token_accuracy': '0.8266', 'epoch': '0.9767'}
+{'loss': '1.564', 'grad_norm': '15.12', 'learning_rate': '2.327e-07', 'num_tokens': '3.743e+06', 'mean_token_accuracy': '0.6772', 'epoch': '0.9769'}
+{'loss': '0.6285', 'grad_norm': '7.844', 'learning_rate': '2.31e-07', 'num_tokens': '3.745e+06', 'mean_token_accuracy': '0.8045', 'epoch': '0.9771'}
+{'loss': '1.057', 'grad_norm': '11.31', 'learning_rate': '2.292e-07', 'num_tokens': '3.746e+06', 'mean_token_accuracy': '0.7002', 'epoch': '0.9772'}
+{'loss': '1.473', 'grad_norm': '16.25', 'learning_rate': '2.275e-07', 'num_tokens': '3.746e+06', 'mean_token_accuracy': '0.6776', 'epoch': '0.9774'}
+{'loss': '0.9911', 'grad_norm': '12.38', 'learning_rate': '2.258e-07', 'num_tokens': '3.747e+06', 'mean_token_accuracy': '0.7124', 'epoch': '0.9776'}
+{'loss': '1.311', 'grad_norm': '34', 'learning_rate': '2.24e-07', 'num_tokens': '3.747e+06', 'mean_token_accuracy': '0.7467', 'epoch': '0.9778'}
+{'loss': '2.351', 'grad_norm': '38.25', 'learning_rate': '2.223e-07', 'num_tokens': '3.747e+06', 'mean_token_accuracy': '0.5368', 'epoch': '0.9779'}
+{'loss': '0.8842', 'grad_norm': '12.19', 'learning_rate': '2.206e-07', 'num_tokens': '3.748e+06', 'mean_token_accuracy': '0.7527', 'epoch': '0.9781'}
+{'loss': '0.7345', 'grad_norm': '8.938', 'learning_rate': '2.188e-07', 'num_tokens': '3.749e+06', 'mean_token_accuracy': '0.776', 'epoch': '0.9783'}
+{'loss': '0.9624', 'grad_norm': '45.25', 'learning_rate': '2.171e-07', 'num_tokens': '3.749e+06', 'mean_token_accuracy': '0.7397', 'epoch': '0.9785'}
+{'loss': '1.514', 'grad_norm': '13.44', 'learning_rate': '2.154e-07', 'num_tokens': '3.75e+06', 'mean_token_accuracy': '0.6216', 'epoch': '0.9786'}
+{'loss': '2.102', 'grad_norm': '17.12', 'learning_rate': '2.136e-07', 'num_tokens': '3.751e+06', 'mean_token_accuracy': '0.5812', 'epoch': '0.9788'}
+{'loss': '0.9372', 'grad_norm': '12.38', 'learning_rate': '2.119e-07', 'num_tokens': '3.751e+06', 'mean_token_accuracy': '0.7902', 'epoch': '0.979'}
+{'loss': '1.345', 'grad_norm': '18.5', 'learning_rate': '2.101e-07', 'num_tokens': '3.752e+06', 'mean_token_accuracy': '0.6543', 'epoch': '0.9792'}
+{'loss': '1.882', 'grad_norm': '16.25', 'learning_rate': '2.084e-07', 'num_tokens': '3.752e+06', 'mean_token_accuracy': '0.5972', 'epoch': '0.9793'}
+{'loss': '1.565', 'grad_norm': '25.12', 'learning_rate': '2.067e-07', 'num_tokens': '3.753e+06', 'mean_token_accuracy': '0.6573', 'epoch': '0.9795'}
+{'loss': '0.9186', 'grad_norm': '10.31', 'learning_rate': '2.049e-07', 'num_tokens': '3.753e+06', 'mean_token_accuracy': '0.7831', 'epoch': '0.9797'}
+{'loss': '1.944', 'grad_norm': '11.94', 'learning_rate': '2.032e-07', 'num_tokens': '3.754e+06', 'mean_token_accuracy': '0.5987', 'epoch': '0.9799'}
+{'loss': '1.714', 'grad_norm': '13.06', 'learning_rate': '2.015e-07', 'num_tokens': '3.755e+06', 'mean_token_accuracy': '0.6254', 'epoch': '0.98'}
+{'loss': '1.967', 'grad_norm': '11.19', 'learning_rate': '1.997e-07', 'num_tokens': '3.755e+06', 'mean_token_accuracy': '0.5669', 'epoch': '0.9802'}
+{'loss': '1.785', 'grad_norm': '18.88', 'learning_rate': '1.98e-07', 'num_tokens': '3.756e+06', 'mean_token_accuracy': '0.6094', 'epoch': '0.9804'}
+{'loss': '1.803', 'grad_norm': '13.94', 'learning_rate': '1.962e-07', 'num_tokens': '3.756e+06', 'mean_token_accuracy': '0.6034', 'epoch': '0.9805'}
+{'loss': '2.132', 'grad_norm': '16.75', 'learning_rate': '1.945e-07', 'num_tokens': '3.757e+06', 'mean_token_accuracy': '0.5611', 'epoch': '0.9807'}
+{'loss': '0.9186', 'grad_norm': '21.38', 'learning_rate': '1.928e-07', 'num_tokens': '3.757e+06', 'mean_token_accuracy': '0.8189', 'epoch': '0.9809'}
+{'loss': '0.9148', 'grad_norm': '7.688', 'learning_rate': '1.91e-07', 'num_tokens': '3.759e+06', 'mean_token_accuracy': '0.7681', 'epoch': '0.9811'}
+{'loss': '1.322', 'grad_norm': '20.12', 'learning_rate': '1.893e-07', 'num_tokens': '3.759e+06', 'mean_token_accuracy': '0.7375', 'epoch': '0.9812'}
+{'loss': '0.8849', 'grad_norm': '11.94', 'learning_rate': '1.876e-07', 'num_tokens': '3.76e+06', 'mean_token_accuracy': '0.7798', 'epoch': '0.9814'}
+{'loss': '1.335', 'grad_norm': '11.19', 'learning_rate': '1.858e-07', 'num_tokens': '3.76e+06', 'mean_token_accuracy': '0.703', 'epoch': '0.9816'}
+{'loss': '0.8758', 'grad_norm': '14.88', 'learning_rate': '1.841e-07', 'num_tokens': '3.761e+06', 'mean_token_accuracy': '0.7217', 'epoch': '0.9818'}
+{'loss': '1.177', 'grad_norm': '15.56', 'learning_rate': '1.824e-07', 'num_tokens': '3.761e+06', 'mean_token_accuracy': '0.7673', 'epoch': '0.9819'}
+{'loss': '1.291', 'grad_norm': '12.75', 'learning_rate': '1.806e-07', 'num_tokens': '3.762e+06', 'mean_token_accuracy': '0.6599', 'epoch': '0.9821'}
+{'loss': '1.836', 'grad_norm': '18.75', 'learning_rate': '1.789e-07', 'num_tokens': '3.763e+06', 'mean_token_accuracy': '0.6285', 'epoch': '0.9823'}
+{'loss': '2', 'grad_norm': '18.75', 'learning_rate': '1.771e-07', 'num_tokens': '3.763e+06', 'mean_token_accuracy': '0.6031', 'epoch': '0.9825'}
+{'loss': '1.293', 'grad_norm': '10.5', 'learning_rate': '1.754e-07', 'num_tokens': '3.764e+06', 'mean_token_accuracy': '0.6554', 'epoch': '0.9826'}
+{'loss': '1.992', 'grad_norm': '36', 'learning_rate': '1.737e-07', 'num_tokens': '3.764e+06', 'mean_token_accuracy': '0.627', 'epoch': '0.9828'}
+{'loss': '1.075', 'grad_norm': '14.5', 'learning_rate': '1.719e-07', 'num_tokens': '3.765e+06', 'mean_token_accuracy': '0.7122', 'epoch': '0.983'}
+{'loss': '0.8545', 'grad_norm': '13.38', 'learning_rate': '1.702e-07', 'num_tokens': '3.765e+06', 'mean_token_accuracy': '0.7295', 'epoch': '0.9832'}
+{'loss': '0.5244', 'grad_norm': '5.781', 'learning_rate': '1.685e-07', 'num_tokens': '3.767e+06', 'mean_token_accuracy': '0.8542', 'epoch': '0.9833'}
+{'loss': '2.315', 'grad_norm': '13.44', 'learning_rate': '1.667e-07', 'num_tokens': '3.768e+06', 'mean_token_accuracy': '0.541', 'epoch': '0.9835'}
+{'loss': '1.526', 'grad_norm': '5.781', 'learning_rate': '1.65e-07', 'num_tokens': '3.771e+06', 'mean_token_accuracy': '0.6623', 'epoch': '0.9837'}
+{'loss': '0.7735', 'grad_norm': '43.25', 'learning_rate': '1.633e-07', 'num_tokens': '3.771e+06', 'mean_token_accuracy': '0.8354', 'epoch': '0.9838'}
+
 98%|█████████▊| 5666/5758 [32:15<00:29,  3.08it/s]
 98%|█████████▊| 5667/5758 [32:15<00:29,  3.11it/s]
                                                   

 98%|█████████▊| 5667/5758 [32:15<00:29,  3.11it/s]
 98%|█████████▊| 5668/5758 [32:15<00:28,  3.15it/s]
                                                   

 98%|█████████▊| 5668/5758 [32:15<00:28,  3.15it/s]
 98%|█████████▊| 5669/5758 [32:16<00:31,  2.86it/s]
                                                   

 98%|█████████▊| 5669/5758 [32:16<00:31,  2.86it/s]
 98%|█████████▊| 5670/5758 [32:16<00:30,  2.89it/s]
                                                   

 98%|█████████▊| 5670/5758 [32:16<00:30,  2.89it/s]
 98%|█████████▊| 5671/5758 [32:16<00:28,  3.00it/s]
                                                   

 98%|█████████▊| 5671/5758 [32:16<00:28,  3.00it/s]
 99%|█████████▊| 5672/5758 [32:17<00:29,  2.93it/s]
                                                   

 99%|█████████▊| 5672/5758 [32:17<00:29,  2.93it/s]
 99%|█████████▊| 5673/5758 [32:17<00:27,  3.06it/s]
                                                   

 99%|█████████▊| 5673/5758 [32:17<00:27,  3.06it/s]
 99%|█████████▊| 5674/5758 [32:17<00:27,  3.01it/s]
                                                   

 99%|█████████▊| 5674/5758 [32:17<00:27,  3.01it/s]
 99%|█████████▊| 5675/5758 [32:18<00:26,  3.10it/s]
                                                   

 99%|█████████▊| 5675/5758 [32:18<00:26,  3.10it/s]
 99%|█████████▊| 5676/5758 [32:18<00:26,  3.11it/s]
                                                   

 99%|█████████▊| 5676/5758 [32:18<00:26,  3.11it/s]
 99%|█████████▊| 5677/5758 [32:18<00:27,  2.95it/s]
                                                   

 99%|█████████▊| 5677/5758 [32:18<00:27,  2.95it/s]
 99%|█████████▊| 5678/5758 [32:19<00:26,  2.98it/s]
                                                   

 99%|█████████▊| 5678/5758 [32:19<00:26,  2.98it/s]
 99%|█████████▊| 5679/5758 [32:19<00:27,  2.86it/s]
                                                   

 99%|█████████▊| 5679/5758 [32:19<00:27,  2.86it/s]
 99%|█████████▊| 5680/5758 [32:19<00:28,  2.73it/s]
                                                   

 99%|█████████▊| 5680/5758 [32:19<00:28,  2.73it/s]
 99%|█████████▊| 5681/5758 [32:20<00:27,  2.77it/s]
                                                   

 99%|█████████▊| 5681/5758 [32:20<00:27,  2.77it/s]
 99%|█████████▊| 5682/5758 [32:20<00:28,  2.66it/s]
                                                   

 99%|█████████▊| 5682/5758 [32:20<00:28,  2.66it/s]
 99%|█████████▊| 5683/5758 [32:20<00:26,  2.78it/s]
                                                   

 99%|█████████▊| 5683/5758 [32:20<00:26,  2.78it/s]
 99%|█████████▊| 5684/5758 [32:21<00:25,  2.87it/s]
                                                   

 99%|█████████▊| 5684/5758 [32:21<00:25,  2.87it/s]
 99%|█████████▊| 5685/5758 [32:21<00:25,  2.91it/s]
                                                   

 99%|█████████▊| 5685/5758 [32:21<00:25,  2.91it/s]
 99%|█████████▊| 5686/5758 [32:21<00:24,  2.94it/s]
                                                   

 99%|█████████▊| 5686/5758 [32:21<00:24,  2.94it/s]
 99%|█████████▉| 5687/5758 [32:22<00:25,  2.80it/s]
                                                   

 99%|█████████▉| 5687/5758 [32:22<00:25,  2.80it/s]
 99%|█████████▉| 5688/5758 [32:22<00:23,  2.93it/s]
                                                   

 99%|█████████▉| 5688/5758 [32:22<00:23,  2.93it/s]
 99%|█████████▉| 5689/5758 [32:22<00:22,  3.04it/s]
                                                   

 99%|█████████▉| 5689/5758 [32:22<00:22,  3.04it/s]
 99%|█████████▉| 5690/5758 [32:23<00:22,  2.96it/s]
                                                   

 99%|█████████▉| 5690/5758 [32:23<00:22,  2.96it/s]
 99%|█████████▉| 5691/5758 [32:23<00:22,  2.98it/s]
                                                   

 99%|█████████▉| 5691/5758 [32:23<00:22,  2.98it/s]
 99%|█████████▉| 5692/5758 [32:23<00:22,  2.97it/s]
                                                   

 99%|█████████▉| 5692/5758 [32:23<00:22,  2.97it/s]
 99%|█████████▉| 5693/5758 [32:24<00:21,  2.97it/s]
                                                   

 99%|█████████▉| 5693/5758 [32:24<00:21,  2.97it/s]
 99%|█████████▉| 5694/5758 [32:24<00:22,  2.83it/s]
                                                   

 99%|█████████▉| 5694/5758 [32:24<00:22,  2.83it/s]
 99%|█████████▉| 5695/5758 [32:25<00:23,  2.72it/s]
                                                   

 99%|█████████▉| 5695/5758 [32:25<00:23,  2.72it/s]
 99%|█████████▉| 5696/5758 [32:25<00:26,  2.36it/s]
                                                   

 99%|█████████▉| 5696/5758 [32:25<00:26,  2.36it/s]
 99%|█████████▉| 5697/5758 [32:25<00:23,  2.58it/s]
                                                   

 99%|█████████▉| 5697/5758 [32:25<00:23,  2.58it/s]
 99%|█████████▉| 5698/5758 [32:26<00:22,  2.66it/s]
                                                   

 99%|█████████▉| 5698/5758 [32:26<00:22,  2.66it/s]
 99%|█████████▉| 5699/5758 [32:26<00:20,  2.81it/s]
                                                   

 99%|█████████▉| 5699/5758 [32:26<00:20,  2.81it/s]
 99%|█████████▉| 5700/5758 [32:26<00:21,  2.75it/s]
                                                   

 99%|█████████▉| 5700/5758 [32:26<00:21,  2.75it/s]
 99%|█████████▉| 5701/5758 [32:27<00:19,  2.88it/s]
                                                   

 99%|█████████▉| 5701/5758 [32:27<00:19,  2.88it/s]
 99%|█████████▉| 5702/5758 [32:27<00:18,  2.97it/s]
                                                   

 99%|█████████▉| 5702/5758 [32:27<00:18,  2.97it/s]
 99%|█████████▉| 5703/5758 [32:27<00:18,  3.02it/s]
                                                   

 99%|█████████▉| 5703/5758 [32:27<00:18,  3.02it/s]
 99%|█████████▉| 5704/5758 [32:28<00:17,  3.06it/s]
                                                   

 99%|█████████▉| 5704/5758 [32:28<00:17,  3.06it/s]
 99%|█████████▉| 5705/5758 [32:28<00:17,  3.03it/s]
                                                   

 99%|█████████▉| 5705/5758 [32:28<00:17,  3.03it/s]
 99%|█████████▉| 5706/5758 [32:28<00:17,  3.00it/s]
                                                   

 99%|█████████▉| 5706/5758 [32:28<00:17,  3.00it/s]
 99%|█████████▉| 5707/5758 [32:29<00:17,  2.89it/s]
                                                   

 99%|█████████▉| 5707/5758 [32:29<00:17,  2.89it/s]
 99%|█████████▉| 5708/5758 [32:29<00:16,  3.01it/s]
                                                   

 99%|█████████▉| 5708/5758 [32:29<00:16,  3.01it/s]
 99%|█████████▉| 5709/5758 [32:29<00:16,  3.06it/s]
                                                   

 99%|█████████▉| 5709/5758 [32:29<00:16,  3.06it/s]
 99%|█████████▉| 5710/5758 [32:30<00:15,  3.08it/s]
                                                   

 99%|█████████▉| 5710/5758 [32:30<00:15,  3.08it/s]
 99%|█████████▉| 5711/5758 [32:30<00:14,  3.18it/s]
                                                   

 99%|█████████▉| 5711/5758 [32:30<00:14,  3.18it/s]
 99%|█████████▉| 5712/5758 [32:30<00:14,  3.24it/s]
                                                   

 99%|█████████▉| 5712/5758 [32:30<00:14,  3.24it/s]
 99%|█████████▉| 5713/5758 [32:31<00:14,  3.10it/s]
                                                   

 99%|█████████▉| 5713/5758 [32:31<00:14,  3.10it/s]
 99%|█████████▉| 5714/5758 [32:31<00:14,  3.12it/s]
                                                   

 99%|█████████▉| 5714/5758 [32:31<00:14,  3.12it/s]
 99%|█████████▉| 5715/5758 [32:31<00:13,  3.19it/s]
                                                   

 99%|█████████▉| 5715/5758 [32:31<00:13,  3.19it/s]
 99%|█████████▉| 5716/5758 [32:32<00:14,  2.93it/s]
                                                   

 99%|█████████▉| 5716/5758 [32:32<00:14,  2.93it/s]
 99%|█████████▉| 5717/5758 [32:32<00:13,  3.04it/s]
                                                   

 99%|█████████▉| 5717/5758 [32:32<00:13,  3.04it/s]
 99%|█████████▉| 5718/5758 [32:32<00:13,  3.01it/s]
                                                   

 99%|█████████▉| 5718/5758 [32:32<00:13,  3.01it/s]
 99%|█████████▉| 5719/5758 [32:33<00:12,  3.12it/s]
                                                   

 99%|█████████▉| 5719/5758 [32:33<00:12,  3.12it/s]
 99%|█████████▉| 5720/5758 [32:33<00:12,  3.13it/s]
                                                   

 99%|█████████▉| 5720/5758 [32:33<00:12,  3.13it/s]
 99%|█████████▉| 5721/5758 [32:33<00:11,  3.14it/s]
                                                   
{'loss': '1.805', 'grad_norm': '12.31', 'learning_rate': '1.615e-07', 'num_tokens': '3.772e+06', 'mean_token_accuracy': '0.6558', 'epoch': '0.984'}
+{'loss': '1.438', 'grad_norm': '18', 'learning_rate': '1.598e-07', 'num_tokens': '3.772e+06', 'mean_token_accuracy': '0.7213', 'epoch': '0.9842'}
+{'loss': '2.298', 'grad_norm': '17.12', 'learning_rate': '1.58e-07', 'num_tokens': '3.773e+06', 'mean_token_accuracy': '0.5424', 'epoch': '0.9844'}
+{'loss': '1.186', 'grad_norm': '12.38', 'learning_rate': '1.563e-07', 'num_tokens': '3.773e+06', 'mean_token_accuracy': '0.6915', 'epoch': '0.9845'}
+{'loss': '1.645', 'grad_norm': '16.12', 'learning_rate': '1.546e-07', 'num_tokens': '3.774e+06', 'mean_token_accuracy': '0.6683', 'epoch': '0.9847'}
+{'loss': '0.5858', 'grad_norm': '42.25', 'learning_rate': '1.528e-07', 'num_tokens': '3.774e+06', 'mean_token_accuracy': '0.8493', 'epoch': '0.9849'}
+{'loss': '1.058', 'grad_norm': '17.25', 'learning_rate': '1.511e-07', 'num_tokens': '3.774e+06', 'mean_token_accuracy': '0.7688', 'epoch': '0.9851'}
+{'loss': '1.396', 'grad_norm': '15', 'learning_rate': '1.494e-07', 'num_tokens': '3.775e+06', 'mean_token_accuracy': '0.65', 'epoch': '0.9852'}
+{'loss': '1.741', 'grad_norm': '23', 'learning_rate': '1.476e-07', 'num_tokens': '3.775e+06', 'mean_token_accuracy': '0.5938', 'epoch': '0.9854'}
+{'loss': '0.8913', 'grad_norm': '12.81', 'learning_rate': '1.459e-07', 'num_tokens': '3.776e+06', 'mean_token_accuracy': '0.7303', 'epoch': '0.9856'}
+{'loss': '1.38', 'grad_norm': '16.75', 'learning_rate': '1.441e-07', 'num_tokens': '3.776e+06', 'mean_token_accuracy': '0.7173', 'epoch': '0.9858'}
+{'loss': '1.228', 'grad_norm': '10.81', 'learning_rate': '1.424e-07', 'num_tokens': '3.777e+06', 'mean_token_accuracy': '0.6999', 'epoch': '0.9859'}
+{'loss': '1.093', 'grad_norm': '7.938', 'learning_rate': '1.407e-07', 'num_tokens': '3.779e+06', 'mean_token_accuracy': '0.7226', 'epoch': '0.9861'}
+{'loss': '1.947', 'grad_norm': '15.12', 'learning_rate': '1.389e-07', 'num_tokens': '3.779e+06', 'mean_token_accuracy': '0.5745', 'epoch': '0.9863'}
+{'loss': '1.005', 'grad_norm': '27.62', 'learning_rate': '1.372e-07', 'num_tokens': '3.779e+06', 'mean_token_accuracy': '0.7721', 'epoch': '0.9865'}
+{'loss': '1.353', 'grad_norm': '12.19', 'learning_rate': '1.355e-07', 'num_tokens': '3.78e+06', 'mean_token_accuracy': '0.6667', 'epoch': '0.9866'}
+{'loss': '1.729', 'grad_norm': '37.75', 'learning_rate': '1.337e-07', 'num_tokens': '3.78e+06', 'mean_token_accuracy': '0.6207', 'epoch': '0.9868'}
+{'loss': '2.52', 'grad_norm': '30.38', 'learning_rate': '1.32e-07', 'num_tokens': '3.781e+06', 'mean_token_accuracy': '0.5412', 'epoch': '0.987'}
+{'loss': '0.5566', 'grad_norm': '13.62', 'learning_rate': '1.303e-07', 'num_tokens': '3.781e+06', 'mean_token_accuracy': '0.8658', 'epoch': '0.9871'}
+{'loss': '0.498', 'grad_norm': '7', 'learning_rate': '1.285e-07', 'num_tokens': '3.783e+06', 'mean_token_accuracy': '0.8506', 'epoch': '0.9873'}
+{'loss': '1.976', 'grad_norm': '9.125', 'learning_rate': '1.268e-07', 'num_tokens': '3.784e+06', 'mean_token_accuracy': '0.5974', 'epoch': '0.9875'}
+{'loss': '1.048', 'grad_norm': '8.938', 'learning_rate': '1.25e-07', 'num_tokens': '3.786e+06', 'mean_token_accuracy': '0.7347', 'epoch': '0.9877'}
+{'loss': '1.675', 'grad_norm': '25.25', 'learning_rate': '1.233e-07', 'num_tokens': '3.786e+06', 'mean_token_accuracy': '0.6182', 'epoch': '0.9878'}
+{'loss': '1.022', 'grad_norm': '12.69', 'learning_rate': '1.216e-07', 'num_tokens': '3.786e+06', 'mean_token_accuracy': '0.6994', 'epoch': '0.988'}
+{'loss': '1.899', 'grad_norm': '14', 'learning_rate': '1.198e-07', 'num_tokens': '3.787e+06', 'mean_token_accuracy': '0.5869', 'epoch': '0.9882'}
+{'loss': '1.19', 'grad_norm': '12.56', 'learning_rate': '1.181e-07', 'num_tokens': '3.788e+06', 'mean_token_accuracy': '0.68', 'epoch': '0.9884'}
+{'loss': '0.2846', 'grad_norm': '23', 'learning_rate': '1.164e-07', 'num_tokens': '3.788e+06', 'mean_token_accuracy': '0.9355', 'epoch': '0.9885'}
+{'loss': '1.128', 'grad_norm': '10.06', 'learning_rate': '1.146e-07', 'num_tokens': '3.789e+06', 'mean_token_accuracy': '0.6969', 'epoch': '0.9887'}
+{'loss': '0.751', 'grad_norm': '19.75', 'learning_rate': '1.129e-07', 'num_tokens': '3.789e+06', 'mean_token_accuracy': '0.8259', 'epoch': '0.9889'}
+{'loss': '1.478', 'grad_norm': '17', 'learning_rate': '1.111e-07', 'num_tokens': '3.79e+06', 'mean_token_accuracy': '0.6634', 'epoch': '0.9891'}
+{'loss': '1.643', 'grad_norm': '13', 'learning_rate': '1.094e-07', 'num_tokens': '3.79e+06', 'mean_token_accuracy': '0.6596', 'epoch': '0.9892'}
+{'loss': '1.74', 'grad_norm': '20.88', 'learning_rate': '1.077e-07', 'num_tokens': '3.79e+06', 'mean_token_accuracy': '0.6185', 'epoch': '0.9894'}
+{'loss': '0.7301', 'grad_norm': '15.38', 'learning_rate': '1.059e-07', 'num_tokens': '3.791e+06', 'mean_token_accuracy': '0.8046', 'epoch': '0.9896'}
+{'loss': '1.135', 'grad_norm': '17.38', 'learning_rate': '1.042e-07', 'num_tokens': '3.791e+06', 'mean_token_accuracy': '0.7784', 'epoch': '0.9898'}
+{'loss': '1.125', 'grad_norm': '13.31', 'learning_rate': '1.025e-07', 'num_tokens': '3.792e+06', 'mean_token_accuracy': '0.7033', 'epoch': '0.9899'}
+{'loss': '1.094', 'grad_norm': '12.31', 'learning_rate': '1.007e-07', 'num_tokens': '3.793e+06', 'mean_token_accuracy': '0.7064', 'epoch': '0.9901'}
+{'loss': '1.817', 'grad_norm': '9.312', 'learning_rate': '9.899e-08', 'num_tokens': '3.794e+06', 'mean_token_accuracy': '0.6156', 'epoch': '0.9903'}
+{'loss': '1.36', 'grad_norm': '20.12', 'learning_rate': '9.726e-08', 'num_tokens': '3.794e+06', 'mean_token_accuracy': '0.7077', 'epoch': '0.9904'}
+{'loss': '0.8151', 'grad_norm': '11.44', 'learning_rate': '9.552e-08', 'num_tokens': '3.795e+06', 'mean_token_accuracy': '0.75', 'epoch': '0.9906'}
+{'loss': '0.9335', 'grad_norm': '18.88', 'learning_rate': '9.378e-08', 'num_tokens': '3.795e+06', 'mean_token_accuracy': '0.8', 'epoch': '0.9908'}
+{'loss': '1.639', 'grad_norm': '17.62', 'learning_rate': '9.205e-08', 'num_tokens': '3.796e+06', 'mean_token_accuracy': '0.6374', 'epoch': '0.991'}
+{'loss': '1.109', 'grad_norm': '17.75', 'learning_rate': '9.031e-08', 'num_tokens': '3.796e+06', 'mean_token_accuracy': '0.7546', 'epoch': '0.9911'}
+{'loss': '1.261', 'grad_norm': '14.5', 'learning_rate': '8.857e-08', 'num_tokens': '3.797e+06', 'mean_token_accuracy': '0.6646', 'epoch': '0.9913'}
+{'loss': '0.5045', 'grad_norm': '7.344', 'learning_rate': '8.684e-08', 'num_tokens': '3.798e+06', 'mean_token_accuracy': '0.8476', 'epoch': '0.9915'}
+{'loss': '0.9085', 'grad_norm': '11.62', 'learning_rate': '8.51e-08', 'num_tokens': '3.799e+06', 'mean_token_accuracy': '0.7765', 'epoch': '0.9917'}
+{'loss': '0.5173', 'grad_norm': '16.88', 'learning_rate': '8.336e-08', 'num_tokens': '3.799e+06', 'mean_token_accuracy': '0.875', 'epoch': '0.9918'}
+{'loss': '0.7203', 'grad_norm': '20.38', 'learning_rate': '8.163e-08', 'num_tokens': '3.8e+06', 'mean_token_accuracy': '0.8063', 'epoch': '0.992'}
+{'loss': '0.3276', 'grad_norm': '14.38', 'learning_rate': '7.989e-08', 'num_tokens': '3.8e+06', 'mean_token_accuracy': '0.9208', 'epoch': '0.9922'}
+{'loss': '0.7376', 'grad_norm': '23.12', 'learning_rate': '7.815e-08', 'num_tokens': '3.8e+06', 'mean_token_accuracy': '0.7914', 'epoch': '0.9924'}
+{'loss': '0.4578', 'grad_norm': '16.12', 'learning_rate': '7.642e-08', 'num_tokens': '3.8e+06', 'mean_token_accuracy': '0.8844', 'epoch': '0.9925'}
+{'loss': '1.086', 'grad_norm': '20.25', 'learning_rate': '7.468e-08', 'num_tokens': '3.8e+06', 'mean_token_accuracy': '0.8', 'epoch': '0.9927'}
+{'loss': '1.567', 'grad_norm': '18.5', 'learning_rate': '7.294e-08', 'num_tokens': '3.801e+06', 'mean_token_accuracy': '0.6709', 'epoch': '0.9929'}
+{'loss': '1.258', 'grad_norm': '10.62', 'learning_rate': '7.121e-08', 'num_tokens': '3.802e+06', 'mean_token_accuracy': '0.6999', 'epoch': '0.9931'}
+{'loss': '1.075', 'grad_norm': '12.5', 'learning_rate': '6.947e-08', 'num_tokens': '3.802e+06', 'mean_token_accuracy': '0.6865', 'epoch': '0.9932'}
+{'loss': '1.424', 'grad_norm': '22.38', 'learning_rate': '6.773e-08', 'num_tokens': '3.803e+06', 'mean_token_accuracy': '0.672', 'epoch': '0.9934'}
+
 99%|█████████▉| 5721/5758 [32:33<00:11,  3.14it/s]
 99%|█████████▉| 5722/5758 [32:34<00:11,  3.15it/s]
                                                   

 99%|█████████▉| 5722/5758 [32:34<00:11,  3.15it/s]
 99%|█████████▉| 5723/5758 [32:34<00:11,  3.12it/s]
                                                   

 99%|█████████▉| 5723/5758 [32:34<00:11,  3.12it/s]
 99%|█████████▉| 5724/5758 [32:34<00:11,  2.84it/s]
                                                   

 99%|█████████▉| 5724/5758 [32:34<00:11,  2.84it/s]
 99%|█████████▉| 5725/5758 [32:35<00:11,  2.96it/s]
                                                   

 99%|█████████▉| 5725/5758 [32:35<00:11,  2.96it/s]
 99%|█████████▉| 5726/5758 [32:35<00:10,  3.09it/s]
                                                   

 99%|█████████▉| 5726/5758 [32:35<00:10,  3.09it/s]
 99%|█████████▉| 5727/5758 [32:35<00:10,  3.07it/s]
                                                   

 99%|█████████▉| 5727/5758 [32:35<00:10,  3.07it/s]
 99%|█████████▉| 5728/5758 [32:36<00:09,  3.06it/s]
                                                   

 99%|█████████▉| 5728/5758 [32:36<00:09,  3.06it/s]
 99%|█████████▉| 5729/5758 [32:36<00:09,  2.99it/s]
                                                   

 99%|█████████▉| 5729/5758 [32:36<00:09,  2.99it/s]
100%|█████████▉| 5730/5758 [32:36<00:09,  3.05it/s]
                                                   

100%|█████████▉| 5730/5758 [32:36<00:09,  3.05it/s]
100%|█████████▉| 5731/5758 [32:37<00:08,  3.07it/s]
                                                   

100%|█████████▉| 5731/5758 [32:37<00:08,  3.07it/s]
100%|█████████▉| 5732/5758 [32:37<00:08,  3.14it/s]
                                                   

100%|█████████▉| 5732/5758 [32:37<00:08,  3.14it/s]
100%|█████████▉| 5733/5758 [32:37<00:08,  3.09it/s]
                                                   

100%|█████████▉| 5733/5758 [32:37<00:08,  3.09it/s]
100%|█████████▉| 5734/5758 [32:38<00:07,  3.09it/s]
                                                   

100%|█████████▉| 5734/5758 [32:38<00:07,  3.09it/s]
100%|█████████▉| 5735/5758 [32:38<00:08,  2.83it/s]
                                                   

100%|█████████▉| 5735/5758 [32:38<00:08,  2.83it/s]
100%|█████████▉| 5736/5758 [32:38<00:08,  2.62it/s]
                                                   

100%|█████████▉| 5736/5758 [32:38<00:08,  2.62it/s]
100%|█████████▉| 5737/5758 [32:39<00:08,  2.46it/s]
                                                   

100%|█████████▉| 5737/5758 [32:39<00:08,  2.46it/s]
100%|█████████▉| 5738/5758 [32:39<00:08,  2.46it/s]
                                                   

100%|█████████▉| 5738/5758 [32:39<00:08,  2.46it/s]
100%|█████████▉| 5739/5758 [32:40<00:07,  2.48it/s]
                                                   

100%|█████████▉| 5739/5758 [32:40<00:07,  2.48it/s]
100%|█████████▉| 5740/5758 [32:40<00:06,  2.59it/s]
                                                   

100%|█████████▉| 5740/5758 [32:40<00:06,  2.59it/s]
100%|█████████▉| 5741/5758 [32:40<00:06,  2.59it/s]
                                                   

100%|█████████▉| 5741/5758 [32:40<00:06,  2.59it/s]
100%|█████████▉| 5742/5758 [32:41<00:06,  2.43it/s]
                                                   

100%|█████████▉| 5742/5758 [32:41<00:06,  2.43it/s]
100%|█████████▉| 5743/5758 [32:41<00:05,  2.52it/s]
                                                   

100%|█████████▉| 5743/5758 [32:41<00:05,  2.52it/s]
100%|█████████▉| 5744/5758 [32:42<00:05,  2.52it/s]
                                                   

100%|█████████▉| 5744/5758 [32:42<00:05,  2.52it/s]
100%|█████████▉| 5745/5758 [32:42<00:05,  2.45it/s]
                                                   

100%|█████████▉| 5745/5758 [32:42<00:05,  2.45it/s]
100%|█████████▉| 5746/5758 [32:42<00:04,  2.41it/s]
                                                   

100%|█████████▉| 5746/5758 [32:42<00:04,  2.41it/s]
100%|█████████▉| 5747/5758 [32:43<00:04,  2.28it/s]
                                                   

100%|█████████▉| 5747/5758 [32:43<00:04,  2.28it/s]
100%|█████████▉| 5748/5758 [32:43<00:04,  2.43it/s]
                                                   

100%|█████████▉| 5748/5758 [32:43<00:04,  2.43it/s]
100%|█████████▉| 5749/5758 [32:44<00:03,  2.48it/s]
                                                   

100%|█████████▉| 5749/5758 [32:44<00:03,  2.48it/s]
100%|█████████▉| 5750/5758 [32:44<00:03,  2.41it/s]
                                                   

100%|█████████▉| 5750/5758 [32:44<00:03,  2.41it/s]
100%|█████████▉| 5751/5758 [32:44<00:02,  2.61it/s]
                                                   

100%|█████████▉| 5751/5758 [32:44<00:02,  2.61it/s]
100%|█████████▉| 5752/5758 [32:45<00:02,  2.47it/s]
                                                   

100%|█████████▉| 5752/5758 [32:45<00:02,  2.47it/s]
100%|█████████▉| 5753/5758 [32:45<00:01,  2.65it/s]
                                                   

100%|█████████▉| 5753/5758 [32:45<00:01,  2.65it/s]
100%|█████████▉| 5754/5758 [32:46<00:01,  2.64it/s]
                                                   

100%|█████████▉| 5754/5758 [32:46<00:01,  2.64it/s]
100%|█████████▉| 5755/5758 [32:46<00:01,  2.68it/s]
                                                   

100%|█████████▉| 5755/5758 [32:46<00:01,  2.68it/s]
100%|█████████▉| 5756/5758 [32:46<00:00,  2.73it/s]
                                                   

100%|█████████▉| 5756/5758 [32:46<00:00,  2.73it/s]
100%|█████████▉| 5757/5758 [32:47<00:00,  2.77it/s]
                                                   

100%|█████████▉| 5757/5758 [32:47<00:00,  2.77it/s]
100%|██████████| 5758/5758 [32:47<00:00,  2.74it/s]
                                                   

100%|██████████| 5758/5758 [32:47<00:00,  2.74it/s]{'loss': '1.255', 'grad_norm': '13.31', 'learning_rate': '6.6e-08', 'num_tokens': '3.803e+06', 'mean_token_accuracy': '0.6481', 'epoch': '0.9936'}
+{'loss': '0.9666', 'grad_norm': '33.25', 'learning_rate': '6.426e-08', 'num_tokens': '3.804e+06', 'mean_token_accuracy': '0.7068', 'epoch': '0.9937'}
+{'loss': '0.9283', 'grad_norm': '18.62', 'learning_rate': '6.252e-08', 'num_tokens': '3.804e+06', 'mean_token_accuracy': '0.7463', 'epoch': '0.9939'}
+{'loss': '1.192', 'grad_norm': '15', 'learning_rate': '6.078e-08', 'num_tokens': '3.805e+06', 'mean_token_accuracy': '0.6577', 'epoch': '0.9941'}
+{'loss': '1.305', 'grad_norm': '13.38', 'learning_rate': '5.905e-08', 'num_tokens': '3.805e+06', 'mean_token_accuracy': '0.6425', 'epoch': '0.9943'}
+{'loss': '0.8205', 'grad_norm': '13', 'learning_rate': '5.731e-08', 'num_tokens': '3.806e+06', 'mean_token_accuracy': '0.8127', 'epoch': '0.9944'}
+{'loss': '0.9843', 'grad_norm': '33.5', 'learning_rate': '5.557e-08', 'num_tokens': '3.806e+06', 'mean_token_accuracy': '0.8211', 'epoch': '0.9946'}
+{'loss': '1.29', 'grad_norm': '7.219', 'learning_rate': '5.384e-08', 'num_tokens': '3.808e+06', 'mean_token_accuracy': '0.7151', 'epoch': '0.9948'}
+{'loss': '1.317', 'grad_norm': '11', 'learning_rate': '5.21e-08', 'num_tokens': '3.809e+06', 'mean_token_accuracy': '0.6454', 'epoch': '0.995'}
+{'loss': '0.6731', 'grad_norm': '15.5', 'learning_rate': '5.036e-08', 'num_tokens': '3.809e+06', 'mean_token_accuracy': '0.8238', 'epoch': '0.9951'}
+{'loss': '1.389', 'grad_norm': '15.62', 'learning_rate': '4.863e-08', 'num_tokens': '3.809e+06', 'mean_token_accuracy': '0.7494', 'epoch': '0.9953'}
+{'loss': '1.971', 'grad_norm': '16.12', 'learning_rate': '4.689e-08', 'num_tokens': '3.81e+06', 'mean_token_accuracy': '0.5988', 'epoch': '0.9955'}
+{'loss': '1.328', 'grad_norm': '12.06', 'learning_rate': '4.515e-08', 'num_tokens': '3.811e+06', 'mean_token_accuracy': '0.6548', 'epoch': '0.9957'}
+{'loss': '1.055', 'grad_norm': '27.25', 'learning_rate': '4.342e-08', 'num_tokens': '3.811e+06', 'mean_token_accuracy': '0.7816', 'epoch': '0.9958'}
+{'loss': '0.964', 'grad_norm': '13.5', 'learning_rate': '4.168e-08', 'num_tokens': '3.812e+06', 'mean_token_accuracy': '0.7076', 'epoch': '0.996'}
+{'loss': '1.491', 'grad_norm': '27', 'learning_rate': '3.994e-08', 'num_tokens': '3.812e+06', 'mean_token_accuracy': '0.6554', 'epoch': '0.9962'}
+{'loss': '1.418', 'grad_norm': '15.75', 'learning_rate': '3.821e-08', 'num_tokens': '3.812e+06', 'mean_token_accuracy': '0.6661', 'epoch': '0.9964'}
+{'loss': '1.464', 'grad_norm': '16.5', 'learning_rate': '3.647e-08', 'num_tokens': '3.813e+06', 'mean_token_accuracy': '0.6984', 'epoch': '0.9965'}
+{'loss': '1.193', 'grad_norm': '10.69', 'learning_rate': '3.473e-08', 'num_tokens': '3.814e+06', 'mean_token_accuracy': '0.674', 'epoch': '0.9967'}
+{'loss': '2.154', 'grad_norm': '9.938', 'learning_rate': '3.3e-08', 'num_tokens': '3.815e+06', 'mean_token_accuracy': '0.5543', 'epoch': '0.9969'}
+{'loss': '1.795', 'grad_norm': '23.88', 'learning_rate': '3.126e-08', 'num_tokens': '3.815e+06', 'mean_token_accuracy': '0.6368', 'epoch': '0.997'}
+{'loss': '1.156', 'grad_norm': '13.5', 'learning_rate': '2.952e-08', 'num_tokens': '3.816e+06', 'mean_token_accuracy': '0.6781', 'epoch': '0.9972'}
+{'loss': '1.576', 'grad_norm': '12.31', 'learning_rate': '2.779e-08', 'num_tokens': '3.816e+06', 'mean_token_accuracy': '0.6561', 'epoch': '0.9974'}
+{'loss': '0.9568', 'grad_norm': '21', 'learning_rate': '2.605e-08', 'num_tokens': '3.817e+06', 'mean_token_accuracy': '0.786', 'epoch': '0.9976'}
+{'loss': '2.211', 'grad_norm': '10.81', 'learning_rate': '2.431e-08', 'num_tokens': '3.818e+06', 'mean_token_accuracy': '0.5363', 'epoch': '0.9977'}
+{'loss': '1.034', 'grad_norm': '13.44', 'learning_rate': '2.258e-08', 'num_tokens': '3.818e+06', 'mean_token_accuracy': '0.7605', 'epoch': '0.9979'}
+{'loss': '1.222', 'grad_norm': '43', 'learning_rate': '2.084e-08', 'num_tokens': '3.818e+06', 'mean_token_accuracy': '0.6835', 'epoch': '0.9981'}
+{'loss': '0.8326', 'grad_norm': '23.62', 'learning_rate': '1.91e-08', 'num_tokens': '3.818e+06', 'mean_token_accuracy': '0.8544', 'epoch': '0.9983'}
+{'loss': '1.175', 'grad_norm': '14', 'learning_rate': '1.737e-08', 'num_tokens': '3.819e+06', 'mean_token_accuracy': '0.6654', 'epoch': '0.9984'}
+{'loss': '2.343', 'grad_norm': '19.62', 'learning_rate': '1.563e-08', 'num_tokens': '3.82e+06', 'mean_token_accuracy': '0.546', 'epoch': '0.9986'}
+{'loss': '1.047', 'grad_norm': '11.5', 'learning_rate': '1.389e-08', 'num_tokens': '3.82e+06', 'mean_token_accuracy': '0.7297', 'epoch': '0.9988'}
+{'loss': '1.324', 'grad_norm': '23.5', 'learning_rate': '1.216e-08', 'num_tokens': '3.821e+06', 'mean_token_accuracy': '0.6698', 'epoch': '0.999'}
+{'loss': '1.922', 'grad_norm': '17.25', 'learning_rate': '1.042e-08', 'num_tokens': '3.821e+06', 'mean_token_accuracy': '0.5884', 'epoch': '0.9991'}
+{'loss': '1.073', 'grad_norm': '13.81', 'learning_rate': '8.684e-09', 'num_tokens': '3.822e+06', 'mean_token_accuracy': '0.7021', 'epoch': '0.9993'}
+{'loss': '0.7747', 'grad_norm': '7.25', 'learning_rate': '6.947e-09', 'num_tokens': '3.824e+06', 'mean_token_accuracy': '0.7689', 'epoch': '0.9995'}
+{'loss': '1.251', 'grad_norm': '10.62', 'learning_rate': '5.21e-09', 'num_tokens': '3.825e+06', 'mean_token_accuracy': '0.6335', 'epoch': '0.9997'}
+{'loss': '2.093', 'grad_norm': '19.75', 'learning_rate': '3.473e-09', 'num_tokens': '3.825e+06', 'mean_token_accuracy': '0.5366', 'epoch': '0.9998'}
+{'loss': '1.018', 'grad_norm': '14.69', 'learning_rate': '1.737e-09', 'num_tokens': '3.826e+06', 'mean_token_accuracy': '0.7214', 'epoch': '1'}
+
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
+
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  1.45it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  1.45it/s]
+
                                                   

100%|██████████| 5758/5758 [32:49<00:00,  2.74it/s]
100%|██████████| 5758/5758 [32:49<00:00,  2.92it/s]
+{'train_runtime': '1970', 'train_samples_per_second': '2.923', 'train_steps_per_second': '2.923', 'train_loss': '1.359', 'epoch': '1'}
+
Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.31it/s]
Writing model shards: 100%|██████████| 1/1 [00:00<00:00,  2.31it/s]