SFT checkpoint step 32

checkpoints/checkpoint-32/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7ca1f6f7d5be6aa02d62c533e798347df7d8cbe28dde19a10b2279b55644322e
+oid sha256:ac4185cd050c8ec04df1a7c0b5cfff97f1434f2febe75749a9af9b10f9170d3f
 size 204500912

checkpoints/checkpoint-32/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b8806c4e8215cc1261a5dcbc0f390465d115e52f8c315386fd1b579af97c8b16
+oid sha256:17f3289f84682bc9a5ae016c1f3057446893c447b0c491592f1c94003cf463fe
 size 620569547

checkpoints/checkpoint-32/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4d4554f823bdad60cb7285b04abc5c634f3a2998e13f5dae9f5b55c2226bb622
+oid sha256:d6c72dbf2bb42cbd57c53b8628dd5b5abb38c5c44175ac9f4c0a898279d87359
 size 1383

checkpoints/checkpoint-32/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e4512d0083440109748e439e94524c7e78fe07765f616adfc020fe86a3deba34
+oid sha256:ce3ab2dc5a54221473fefd9014c9838afd55c2c20a5337fbeed20c5a935ff456
 size 1465

checkpoints/checkpoint-32/time_state.json

@@ -1 +1 @@
-{"consumed_seconds": 332.8946771621704, "global_step": 32}
+{"consumed_seconds": 344.63673210144043, "global_step": 32}

checkpoints/checkpoint-32/trainer_state.json

@@ -17,15 +17,15 @@
       "num_input_tokens_seen": 1494,
       "num_tokens": 1494.0,
       "step": 1,
-      "train_runtime": 44.9864,
-      "train_tokens_per_second": 33.21
+      "train_runtime": 42.4619,
+      "train_tokens_per_second": 35.185
     },
     {
       "epoch": 0.06956521739130435,
       "eval_loss": 5.31978178024292,
       "eval_num_tokens": 4980.0,
-      "eval_runtime": 28.6663,
-      "eval_samples_per_second": 1.047,
+      "eval_runtime": 28.7144,
+      "eval_samples_per_second": 1.045,
       "eval_steps_per_second": 0.279,
       "num_input_tokens_seen": 4980,
       "step": 4
@@ -34,102 +34,102 @@
       "epoch": 0.1391304347826087,
       "eval_loss": 5.31978178024292,
       "eval_num_tokens": 9004.0,
-      "eval_runtime": 3.5559,
-      "eval_samples_per_second": 8.437,
-      "eval_steps_per_second": 2.25,
+      "eval_runtime": 3.4006,
+      "eval_samples_per_second": 8.822,
+      "eval_steps_per_second": 2.353,
       "num_input_tokens_seen": 9004,
       "step": 8
     },
     {
       "epoch": 0.17391304347826086,
-      "grad_norm": 28.138591766357422,
+      "grad_norm": 28.1390438079834,
       "learning_rate": 0.0,
       "loss": 2.5477417839898004,
       "num_input_tokens_seen": 11145,
       "num_tokens": 11145.0,
       "step": 10,
-      "train_runtime": 145.0656,
-      "train_tokens_per_second": 76.827
+      "train_runtime": 146.4448,
+      "train_tokens_per_second": 76.104
     },
     {
       "epoch": 0.20869565217391303,
-      "eval_loss": 1.795594573020935,
+      "eval_loss": 1.7850792407989502,
       "eval_num_tokens": 13293.0,
-      "eval_runtime": 3.5484,
-      "eval_samples_per_second": 8.455,
-      "eval_steps_per_second": 2.255,
+      "eval_runtime": 3.4816,
+      "eval_samples_per_second": 8.617,
+      "eval_steps_per_second": 2.298,
       "num_input_tokens_seen": 13293,
       "step": 12
     },
     {
       "epoch": 0.2782608695652174,
-      "eval_loss": 0.5860968232154846,
+      "eval_loss": 0.5965744256973267,
       "eval_num_tokens": 17742.0,
-      "eval_runtime": 3.4658,
-      "eval_samples_per_second": 8.656,
-      "eval_steps_per_second": 2.308,
+      "eval_runtime": 3.4391,
+      "eval_samples_per_second": 8.723,
+      "eval_steps_per_second": 2.326,
       "num_input_tokens_seen": 17742,
       "step": 16
     },
     {
       "epoch": 0.34782608695652173,
-      "grad_norm": 7.353201389312744,
-      "learning_rate": 0.0001442909649383465,
-      "loss": 0.8008382797241211,
+      "grad_norm": 6.518080711364746,
+      "learning_rate": 0.0001425726650926814,
+      "loss": 0.7919160842895507,
       "num_input_tokens_seen": 20264,
       "num_tokens": 20264.0,
       "step": 20,
-      "train_runtime": 227.664,
-      "train_tokens_per_second": 89.008
+      "train_runtime": 241.9804,
+      "train_tokens_per_second": 83.742
     },
     {
       "epoch": 0.34782608695652173,
-      "eval_loss": 0.39688172936439514,
+      "eval_loss": 0.6307523250579834,
       "eval_num_tokens": 20264.0,
-      "eval_runtime": 3.494,
-      "eval_samples_per_second": 8.586,
-      "eval_steps_per_second": 2.29,
+      "eval_runtime": 3.512,
+      "eval_samples_per_second": 8.542,
+      "eval_steps_per_second": 2.278,
       "num_input_tokens_seen": 20264,
       "step": 20
     },
     {
       "epoch": 0.41739130434782606,
-      "eval_loss": 0.5452550053596497,
+      "eval_loss": 0.3135659396648407,
       "eval_num_tokens": 24990.0,
-      "eval_runtime": 3.5856,
-      "eval_samples_per_second": 8.367,
-      "eval_steps_per_second": 2.231,
+      "eval_runtime": 3.4341,
+      "eval_samples_per_second": 8.736,
+      "eval_steps_per_second": 2.33,
       "num_input_tokens_seen": 24990,
       "step": 24
     },
     {
       "epoch": 0.48695652173913045,
-      "eval_loss": 0.31608301401138306,
+      "eval_loss": 0.4761301875114441,
       "eval_num_tokens": 28628.0,
-      "eval_runtime": 3.5529,
-      "eval_samples_per_second": 8.444,
-      "eval_steps_per_second": 2.252,
+      "eval_runtime": 3.4198,
+      "eval_samples_per_second": 8.772,
+      "eval_steps_per_second": 2.339,
       "num_input_tokens_seen": 28628,
       "step": 28
     },
     {
       "epoch": 0.5217391304347826,
-      "grad_norm": 2.936347007751465,
-      "learning_rate": 0.00011840034721348544,
-      "loss": 0.20213119983673095,
+      "grad_norm": 3.650630235671997,
+      "learning_rate": 0.00011490240573865023,
+      "loss": 0.22239422798156738,
       "num_input_tokens_seen": 30604,
       "num_tokens": 30604.0,
       "step": 30,
-      "train_runtime": 314.4011,
-      "train_tokens_per_second": 97.341
+      "train_runtime": 326.8185,
+      "train_tokens_per_second": 93.642
     },
     {
       "epoch": 0.5565217391304348,
-      "eval_loss": 0.3312884569168091,
+      "eval_loss": 0.2599254846572876,
       "eval_num_tokens": 33017.0,
-      "eval_runtime": 3.5533,
-      "eval_samples_per_second": 8.443,
-      "eval_steps_per_second": 2.251,
+      "eval_runtime": 3.409,
+      "eval_samples_per_second": 8.8,
+      "eval_steps_per_second": 2.347,
       "num_input_tokens_seen": 33017,
       "step": 32
     }