Upload folder using huggingface_hub

added_tokens.json

@@ -0,0 +1,5 @@
+{
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644
+}

all_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.9985842378480414,
+    "total_flos": 5.875342281120154e+16,
+    "train_loss": 0.11385704865424108,
+    "train_runtime": 162519.6807,
+    "train_samples_per_second": 3.441,
+    "train_steps_per_second": 0.02
+}

config.json

@@ -0,0 +1,28 @@
+{
+  "_name_or_path": "/apdcephfs_sh4/share_300001476/trenousqiu/models/models--Qwen--Qwen2-72B/snapshots/87993795c78576318087f70b43fbf530eb7789e7",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 8192,
+  "initializer_range": 0.02,
+  "intermediate_size": 29568,
+  "max_position_embeddings": 131072,
+  "max_window_layers": 80,
+  "model_type": "qwen2",
+  "num_attention_heads": 64,
+  "num_hidden_layers": 80,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.44.1",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.44.1"
+}

model-00001-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b479e533f8d91e4a71e75e922c244fad8d95827a917914e792c67ca127d76393
+size 4548798728

model-00002-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:694d4302a4093dc8d8bf024e2f83665d205b8d6be6bfc1b3bd8abb88e58d8480
+size 4964101384

model-00003-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:549de6fe185337a9b31338818494c1f8cc0816a4cde4f86861bfd6adb78a639f
+size 4781637328

model-00004-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bfc432840c4f7a7748aa38f5e64e059dac8c562530a1531c7100e21560130fcb
+size 4781670320

model-00005-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2894954565c53db836060f018f76c540c7b36c6c5f544fea131040ddf96f50c1
+size 4781670360

model-00006-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:251fa12119f970c5c08560007351dfae6283f06ab490502e281b6856a11d5911
+size 4964101416

model-00007-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0cd0c7a2b0c406b6f6f53e45308068300e3d17e89e3450f945ea7dc4ed3eecf
+size 4781637360

model-00008-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53550200da186f89dc848dd40b0f16900bae0e5a01e89da69a632833d9482090
+size 4781670360

model-00009-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e732b2099464dab5ff7594c6f3fb41fecf9fe5104b1e498783fdcd9cd2674302
+size 4781670360

model-00010-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ae4c50f0e307bfaab39cbcee5694c037668fb6ce87150c8a61c35811a29ba81
+size 4964101416

model-00011-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46d6436c54a504e6c4f5451d62def2de50044236db0f862944ef24bd9640f3fe
+size 4781637360

model-00012-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71dde60db7538ea3bab274457e87b90c7b1301f1666ee8f22e450917904ebf9e
+size 4781670360

model-00013-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f28377f2d2634b9c24d1187077dfebcd66f885bead671772eab2dc9ff7d5dd2
+size 4781670360

model-00014-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40cba5215be8b67a4a78eef96c0ccbc314cff53f180ed534864ec56dbacdc63b
+size 4964101416

model-00015-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd1cf43cfa9726e402bf7dd1a18169f74a8fa4fa06d75b6194ce9130d9e04777
+size 4781637360

model-00016-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32bfb4b048ce284c7184a15516e38bffa004539adc63d2c679af750281326768
+size 4781670360

model-00017-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cabb3fa7986f57462f83928cd5fb6f519314ca7dad3c8d42e6b484b7a6a004b2
+size 4781670360

model-00018-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b58e8e0c26ac0c4b240f0a8a44dc2309524fe6a4657ce7ca9597e629340126e8
+size 4964101416

model-00019-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6d324aff1294116358c7399181d1a837f869abf5a0e327898871a6117946b666
+size 4781637360

model-00020-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64290f5b7793269cd1325a548b77f37796483a86dcff907463d8c7b652597c65
+size 4781670360

model-00021-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49affcaef7bb0946fe778232f1f862bb0e3d030623867a8045f114479436d443
+size 4781670360

model-00022-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18913acd7085462945976cf75f25fd05a038c799264b5825a2a7810494d9807b
+size 4964101416

model-00023-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:133929b381f23a47b1de884865864fbc1e4566bd12ca874189a9722f6198151c
+size 4781637360

model-00024-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f20b86472ed8f32e23b731c9ffc44f844377d61875c135c21452614e8dadffba
+size 4781670360

model-00025-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae9c31c57843fd433a238aa8a57ba3692a5c9556482784f59d8a41db5911dee0
+size 4781670360

model-00026-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b0f4af92c7cd395b375d023419baf949608ed70a106f1698707a584fc6a8d8d
+size 4964101416

model-00027-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fed11336d12598707669190e9185e89b33c7b4063975be6833892cb0ea38d39
+size 4781637360

model-00028-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e69cbde30087789f1a7addce6fb882f18d53b911853a067b3cf7b325c4af48ed
+size 4781670360

model-00029-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbad54db8f2e0ad4d97b5b03fcd838da782fdd9a3de23cbe85b96c5d22ff54e4
+size 4781670360

model-00030-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3b578fa6adaf230c7c2b9e9799ac81aba9c90c6523d9db4dcf26be451d711a54
+size 3208747032

model-00031-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed6d1144ca75e01b4ce7898e69157f87e801fa60592b30fd64b6bdf8fa7fb67b
+size 2491416704

model.safetensors.index.json

@@ -0,0 +1,970 @@
+{
+  "metadata": {
+    "total_size": 145412407296
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00031-of-00031.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00031.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00031.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00031.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00031.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00031.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00031.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00031.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.k_proj.bias": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.q_proj.bias": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.v_proj.bias": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.11.self_attn.k_proj.bias": "model-00005-of-00031.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.11.self_attn.q_proj.bias": "model-00005-of-00031.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.11.self_attn.v_proj.bias": "model-00005-of-00031.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.self_attn.k_proj.bias": "model-00005-of-00031.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.12.self_attn.q_proj.bias": "model-00005-of-00031.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.12.self_attn.v_proj.bias": "model-00005-of-00031.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.self_attn.k_proj.bias": "model-00006-of-00031.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.self_attn.q_proj.bias": "model-00006-of-00031.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.self_attn.v_proj.bias": "model-00006-of-00031.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.14.self_attn.k_proj.bias": "model-00006-of-00031.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.14.self_attn.q_proj.bias": "model-00006-of-00031.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.14.self_attn.v_proj.bias": "model-00006-of-00031.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.k_proj.bias": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.q_proj.bias": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.v_proj.bias": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.k_proj.bias": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.q_proj.bias": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.v_proj.bias": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.self_attn.k_proj.bias": "model-00007-of-00031.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.17.self_attn.q_proj.bias": "model-00007-of-00031.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.17.self_attn.v_proj.bias": "model-00007-of-00031.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.k_proj.bias": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.q_proj.bias": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.v_proj.bias": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.self_attn.k_proj.bias": "model-00008-of-00031.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "model-00008-of-00031.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.self_attn.v_proj.bias": "model-00008-of-00031.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.self_attn.k_proj.bias": "model-00002-of-00031.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.self_attn.q_proj.bias": "model-00002-of-00031.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.self_attn.v_proj.bias": "model-00002-of-00031.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00008-of-00031.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00008-of-00031.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00008-of-00031.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.k_proj.bias": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.q_proj.bias": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.v_proj.bias": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.self_attn.k_proj.bias": "model-00009-of-00031.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.self_attn.q_proj.bias": "model-00009-of-00031.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.self_attn.v_proj.bias": "model-00009-of-00031.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.self_attn.k_proj.bias": "model-00009-of-00031.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.23.self_attn.q_proj.bias": "model-00009-of-00031.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.23.self_attn.v_proj.bias": "model-00009-of-00031.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.self_attn.k_proj.bias": "model-00010-of-00031.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.self_attn.q_proj.bias": "model-00010-of-00031.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.self_attn.v_proj.bias": "model-00010-of-00031.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.25.self_attn.k_proj.bias": "model-00010-of-00031.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.25.self_attn.q_proj.bias": "model-00010-of-00031.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.25.self_attn.v_proj.bias": "model-00010-of-00031.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.k_proj.bias": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.q_proj.bias": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.v_proj.bias": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.self_attn.k_proj.bias": "model-00011-of-00031.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.self_attn.q_proj.bias": "model-00011-of-00031.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.self_attn.v_proj.bias": "model-00011-of-00031.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.self_attn.k_proj.bias": "model-00011-of-00031.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.28.self_attn.q_proj.bias": "model-00011-of-00031.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.28.self_attn.v_proj.bias": "model-00011-of-00031.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.k_proj.bias": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.q_proj.bias": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.v_proj.bias": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.3.self_attn.k_proj.bias": "model-00002-of-00031.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.3.self_attn.q_proj.bias": "model-00002-of-00031.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.3.self_attn.v_proj.bias": "model-00002-of-00031.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.self_attn.k_proj.bias": "model-00012-of-00031.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.self_attn.q_proj.bias": "model-00012-of-00031.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.self_attn.v_proj.bias": "model-00012-of-00031.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.self_attn.k_proj.bias": "model-00012-of-00031.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.31.self_attn.q_proj.bias": "model-00012-of-00031.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.31.self_attn.v_proj.bias": "model-00012-of-00031.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.32.input_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.mlp.down_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.mlp.gate_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.mlp.up_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.post_attention_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.k_proj.bias": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.k_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.o_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.q_proj.bias": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.q_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.v_proj.bias": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.v_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.input_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.mlp.down_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.mlp.gate_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.mlp.up_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.post_attention_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.self_attn.k_proj.bias": "model-00013-of-00031.safetensors",
+    "model.layers.33.self_attn.k_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.self_attn.o_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.self_attn.q_proj.bias": "model-00013-of-00031.safetensors",
+    "model.layers.33.self_attn.q_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.self_attn.v_proj.bias": "model-00013-of-00031.safetensors",
+    "model.layers.33.self_attn.v_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.34.input_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.mlp.down_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.mlp.gate_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.mlp.up_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.post_attention_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.self_attn.k_proj.bias": "model-00013-of-00031.safetensors",
+    "model.layers.34.self_attn.k_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.34.self_attn.o_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.34.self_attn.q_proj.bias": "model-00013-of-00031.safetensors",
+    "model.layers.34.self_attn.q_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.34.self_attn.v_proj.bias": "model-00013-of-00031.safetensors",
+    "model.layers.34.self_attn.v_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.35.input_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.mlp.down_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.mlp.gate_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.mlp.up_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.post_attention_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.self_attn.k_proj.bias": "model-00014-of-00031.safetensors",
+    "model.layers.35.self_attn.k_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.self_attn.o_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.self_attn.q_proj.bias": "model-00014-of-00031.safetensors",
+    "model.layers.35.self_attn.q_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.self_attn.v_proj.bias": "model-00014-of-00031.safetensors",
+    "model.layers.35.self_attn.v_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.36.input_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.36.mlp.down_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.36.mlp.gate_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.36.mlp.up_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.36.post_attention_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.36.self_attn.k_proj.bias": "model-00014-of-00031.safetensors",
+    "model.layers.36.self_attn.k_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.36.self_attn.o_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.36.self_attn.q_proj.bias": "model-00014-of-00031.safetensors",
+    "model.layers.36.self_attn.q_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.36.self_attn.v_proj.bias": "model-00014-of-00031.safetensors",
+    "model.layers.36.self_attn.v_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.37.input_layernorm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.mlp.down_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.mlp.gate_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.mlp.up_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.post_attention_layernorm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.k_proj.bias": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.k_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.o_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.q_proj.bias": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.q_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.v_proj.bias": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.v_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.input_layernorm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.mlp.down_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.mlp.gate_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.mlp.up_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.post_attention_layernorm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.self_attn.k_proj.bias": "model-00015-of-00031.safetensors",
+    "model.layers.38.self_attn.k_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.self_attn.o_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.self_attn.q_proj.bias": "model-00015-of-00031.safetensors",
+    "model.layers.38.self_attn.q_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.self_attn.v_proj.bias": "model-00015-of-00031.safetensors",
+    "model.layers.38.self_attn.v_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.39.input_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.mlp.down_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.mlp.gate_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.39.mlp.up_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.39.post_attention_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.self_attn.k_proj.bias": "model-00015-of-00031.safetensors",
+    "model.layers.39.self_attn.k_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.39.self_attn.o_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.39.self_attn.q_proj.bias": "model-00015-of-00031.safetensors",
+    "model.layers.39.self_attn.q_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.39.self_attn.v_proj.bias": "model-00015-of-00031.safetensors",
+    "model.layers.39.self_attn.v_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.k_proj.bias": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.q_proj.bias": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.v_proj.bias": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.40.input_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.mlp.down_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.mlp.gate_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.mlp.up_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.post_attention_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.k_proj.bias": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.k_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.o_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.q_proj.bias": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.q_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.v_proj.bias": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.v_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.input_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.mlp.down_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.mlp.gate_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.mlp.up_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.post_attention_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.self_attn.k_proj.bias": "model-00016-of-00031.safetensors",
+    "model.layers.41.self_attn.k_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.self_attn.o_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.self_attn.q_proj.bias": "model-00016-of-00031.safetensors",
+    "model.layers.41.self_attn.q_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.self_attn.v_proj.bias": "model-00016-of-00031.safetensors",
+    "model.layers.41.self_attn.v_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.42.input_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.mlp.down_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.mlp.gate_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.42.mlp.up_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.post_attention_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.self_attn.k_proj.bias": "model-00016-of-00031.safetensors",
+    "model.layers.42.self_attn.k_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.42.self_attn.o_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.42.self_attn.q_proj.bias": "model-00016-of-00031.safetensors",
+    "model.layers.42.self_attn.q_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.42.self_attn.v_proj.bias": "model-00016-of-00031.safetensors",
+    "model.layers.42.self_attn.v_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.43.input_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.mlp.down_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.mlp.gate_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.mlp.up_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.post_attention_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.k_proj.bias": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.k_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.o_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.q_proj.bias": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.q_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.v_proj.bias": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.v_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.input_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.mlp.down_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.mlp.gate_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.mlp.up_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.post_attention_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.self_attn.k_proj.bias": "model-00017-of-00031.safetensors",
+    "model.layers.44.self_attn.k_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.self_attn.o_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.self_attn.q_proj.bias": "model-00017-of-00031.safetensors",
+    "model.layers.44.self_attn.q_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.self_attn.v_proj.bias": "model-00017-of-00031.safetensors",
+    "model.layers.44.self_attn.v_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.45.input_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.mlp.down_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.mlp.gate_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.mlp.up_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.post_attention_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.self_attn.k_proj.bias": "model-00017-of-00031.safetensors",
+    "model.layers.45.self_attn.k_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.45.self_attn.o_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.45.self_attn.q_proj.bias": "model-00017-of-00031.safetensors",
+    "model.layers.45.self_attn.q_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.45.self_attn.v_proj.bias": "model-00017-of-00031.safetensors",
+    "model.layers.45.self_attn.v_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.46.input_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.mlp.down_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.mlp.gate_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.mlp.up_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.post_attention_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.self_attn.k_proj.bias": "model-00018-of-00031.safetensors",
+    "model.layers.46.self_attn.k_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.self_attn.o_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.self_attn.q_proj.bias": "model-00018-of-00031.safetensors",
+    "model.layers.46.self_attn.q_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.self_attn.v_proj.bias": "model-00018-of-00031.safetensors",
+    "model.layers.46.self_attn.v_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.47.input_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.47.mlp.down_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.47.mlp.gate_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.47.mlp.up_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.47.post_attention_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.47.self_attn.k_proj.bias": "model-00018-of-00031.safetensors",
+    "model.layers.47.self_attn.k_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.47.self_attn.o_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.47.self_attn.q_proj.bias": "model-00018-of-00031.safetensors",
+    "model.layers.47.self_attn.q_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.47.self_attn.v_proj.bias": "model-00018-of-00031.safetensors",
+    "model.layers.47.self_attn.v_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.48.input_layernorm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.mlp.down_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.mlp.gate_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.mlp.up_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.post_attention_layernorm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.k_proj.bias": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.k_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.o_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.q_proj.bias": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.q_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.v_proj.bias": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.v_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.input_layernorm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.mlp.down_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.mlp.gate_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.mlp.up_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.post_attention_layernorm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.self_attn.k_proj.bias": "model-00019-of-00031.safetensors",
+    "model.layers.49.self_attn.k_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.self_attn.o_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.self_attn.q_proj.bias": "model-00019-of-00031.safetensors",
+    "model.layers.49.self_attn.q_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.self_attn.v_proj.bias": "model-00019-of-00031.safetensors",
+    "model.layers.49.self_attn.v_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.k_proj.bias": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.q_proj.bias": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.v_proj.bias": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.50.input_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.mlp.down_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.mlp.gate_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.50.mlp.up_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.50.post_attention_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.self_attn.k_proj.bias": "model-00019-of-00031.safetensors",
+    "model.layers.50.self_attn.k_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.50.self_attn.o_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.50.self_attn.q_proj.bias": "model-00019-of-00031.safetensors",
+    "model.layers.50.self_attn.q_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.50.self_attn.v_proj.bias": "model-00019-of-00031.safetensors",
+    "model.layers.50.self_attn.v_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.51.input_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.mlp.down_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.mlp.gate_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.mlp.up_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.post_attention_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.k_proj.bias": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.k_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.o_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.q_proj.bias": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.q_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.v_proj.bias": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.v_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.input_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.mlp.down_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.mlp.gate_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.mlp.up_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.post_attention_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.k_proj.bias": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.k_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.o_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.q_proj.bias": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.q_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.v_proj.bias": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.v_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.53.input_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.mlp.down_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.mlp.gate_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.53.mlp.up_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.post_attention_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.self_attn.k_proj.bias": "model-00020-of-00031.safetensors",
+    "model.layers.53.self_attn.k_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.53.self_attn.o_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.53.self_attn.q_proj.bias": "model-00020-of-00031.safetensors",
+    "model.layers.53.self_attn.q_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.53.self_attn.v_proj.bias": "model-00020-of-00031.safetensors",
+    "model.layers.53.self_attn.v_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.54.input_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.mlp.down_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.mlp.gate_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.mlp.up_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.post_attention_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.k_proj.bias": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.k_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.o_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.q_proj.bias": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.q_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.v_proj.bias": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.v_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.input_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.mlp.down_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.mlp.gate_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.mlp.up_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.post_attention_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.self_attn.k_proj.bias": "model-00021-of-00031.safetensors",
+    "model.layers.55.self_attn.k_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.self_attn.o_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.self_attn.q_proj.bias": "model-00021-of-00031.safetensors",
+    "model.layers.55.self_attn.q_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.self_attn.v_proj.bias": "model-00021-of-00031.safetensors",
+    "model.layers.55.self_attn.v_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.56.input_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.mlp.down_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.mlp.gate_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.mlp.up_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.post_attention_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.self_attn.k_proj.bias": "model-00021-of-00031.safetensors",
+    "model.layers.56.self_attn.k_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.56.self_attn.o_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.56.self_attn.q_proj.bias": "model-00021-of-00031.safetensors",
+    "model.layers.56.self_attn.q_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.56.self_attn.v_proj.bias": "model-00021-of-00031.safetensors",
+    "model.layers.56.self_attn.v_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.57.input_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.mlp.down_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.mlp.gate_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.mlp.up_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.post_attention_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.k_proj.bias": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.k_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.o_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.q_proj.bias": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.q_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.v_proj.bias": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.v_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.58.input_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.58.mlp.down_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.58.mlp.gate_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.58.mlp.up_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.58.post_attention_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.58.self_attn.k_proj.bias": "model-00022-of-00031.safetensors",
+    "model.layers.58.self_attn.k_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.58.self_attn.o_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.58.self_attn.q_proj.bias": "model-00022-of-00031.safetensors",
+    "model.layers.58.self_attn.q_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.58.self_attn.v_proj.bias": "model-00022-of-00031.safetensors",
+    "model.layers.58.self_attn.v_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.59.input_layernorm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.mlp.down_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.mlp.gate_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.mlp.up_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.post_attention_layernorm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.k_proj.bias": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.k_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.o_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.q_proj.bias": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.q_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.v_proj.bias": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.v_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.self_attn.k_proj.bias": "model-00003-of-00031.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.6.self_attn.q_proj.bias": "model-00003-of-00031.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.6.self_attn.v_proj.bias": "model-00003-of-00031.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.60.input_layernorm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.mlp.down_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.mlp.gate_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.mlp.up_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.post_attention_layernorm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.self_attn.k_proj.bias": "model-00023-of-00031.safetensors",
+    "model.layers.60.self_attn.k_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.self_attn.o_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.self_attn.q_proj.bias": "model-00023-of-00031.safetensors",
+    "model.layers.60.self_attn.q_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.self_attn.v_proj.bias": "model-00023-of-00031.safetensors",
+    "model.layers.60.self_attn.v_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.61.input_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.mlp.down_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.mlp.gate_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.61.mlp.up_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.61.post_attention_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.self_attn.k_proj.bias": "model-00023-of-00031.safetensors",
+    "model.layers.61.self_attn.k_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.61.self_attn.o_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.61.self_attn.q_proj.bias": "model-00023-of-00031.safetensors",
+    "model.layers.61.self_attn.q_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.61.self_attn.v_proj.bias": "model-00023-of-00031.safetensors",
+    "model.layers.61.self_attn.v_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.62.input_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.mlp.down_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.mlp.gate_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.mlp.up_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.post_attention_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.k_proj.bias": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.k_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.o_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.q_proj.bias": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.q_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.v_proj.bias": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.v_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.input_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.mlp.down_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.mlp.gate_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.mlp.up_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.post_attention_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.self_attn.k_proj.bias": "model-00024-of-00031.safetensors",
+    "model.layers.63.self_attn.k_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.self_attn.o_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.self_attn.q_proj.bias": "model-00024-of-00031.safetensors",
+    "model.layers.63.self_attn.q_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.self_attn.v_proj.bias": "model-00024-of-00031.safetensors",
+    "model.layers.63.self_attn.v_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.64.input_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.mlp.down_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.mlp.gate_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.64.mlp.up_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.post_attention_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.self_attn.k_proj.bias": "model-00024-of-00031.safetensors",
+    "model.layers.64.self_attn.k_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.64.self_attn.o_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.64.self_attn.q_proj.bias": "model-00024-of-00031.safetensors",
+    "model.layers.64.self_attn.q_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.64.self_attn.v_proj.bias": "model-00024-of-00031.safetensors",
+    "model.layers.64.self_attn.v_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.65.input_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.mlp.down_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.mlp.gate_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.mlp.up_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.post_attention_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.k_proj.bias": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.k_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.o_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.q_proj.bias": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.q_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.v_proj.bias": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.v_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.input_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.mlp.down_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.mlp.gate_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.mlp.up_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.post_attention_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.self_attn.k_proj.bias": "model-00025-of-00031.safetensors",
+    "model.layers.66.self_attn.k_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.self_attn.o_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.self_attn.q_proj.bias": "model-00025-of-00031.safetensors",
+    "model.layers.66.self_attn.q_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.self_attn.v_proj.bias": "model-00025-of-00031.safetensors",
+    "model.layers.66.self_attn.v_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.67.input_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.mlp.down_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.mlp.gate_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.mlp.up_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.post_attention_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.self_attn.k_proj.bias": "model-00025-of-00031.safetensors",
+    "model.layers.67.self_attn.k_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.67.self_attn.o_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.67.self_attn.q_proj.bias": "model-00025-of-00031.safetensors",
+    "model.layers.67.self_attn.q_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.67.self_attn.v_proj.bias": "model-00025-of-00031.safetensors",
+    "model.layers.67.self_attn.v_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.68.input_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.mlp.down_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.mlp.gate_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.mlp.up_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.post_attention_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.self_attn.k_proj.bias": "model-00026-of-00031.safetensors",
+    "model.layers.68.self_attn.k_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.self_attn.o_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.self_attn.q_proj.bias": "model-00026-of-00031.safetensors",
+    "model.layers.68.self_attn.q_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.self_attn.v_proj.bias": "model-00026-of-00031.safetensors",
+    "model.layers.68.self_attn.v_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.69.input_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.69.mlp.down_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.69.mlp.gate_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.69.mlp.up_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.69.post_attention_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.69.self_attn.k_proj.bias": "model-00026-of-00031.safetensors",
+    "model.layers.69.self_attn.k_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.69.self_attn.o_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.69.self_attn.q_proj.bias": "model-00026-of-00031.safetensors",
+    "model.layers.69.self_attn.q_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.69.self_attn.v_proj.bias": "model-00026-of-00031.safetensors",
+    "model.layers.69.self_attn.v_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.k_proj.bias": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.q_proj.bias": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.v_proj.bias": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.70.input_layernorm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.mlp.down_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.mlp.gate_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.mlp.up_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.post_attention_layernorm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.k_proj.bias": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.k_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.o_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.q_proj.bias": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.q_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.v_proj.bias": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.v_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.input_layernorm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.mlp.down_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.mlp.gate_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.mlp.up_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.post_attention_layernorm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.self_attn.k_proj.bias": "model-00027-of-00031.safetensors",
+    "model.layers.71.self_attn.k_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.self_attn.o_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.self_attn.q_proj.bias": "model-00027-of-00031.safetensors",
+    "model.layers.71.self_attn.q_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.self_attn.v_proj.bias": "model-00027-of-00031.safetensors",
+    "model.layers.71.self_attn.v_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.72.input_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.mlp.down_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.mlp.gate_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.72.mlp.up_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.72.post_attention_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.self_attn.k_proj.bias": "model-00027-of-00031.safetensors",
+    "model.layers.72.self_attn.k_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.72.self_attn.o_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.72.self_attn.q_proj.bias": "model-00027-of-00031.safetensors",
+    "model.layers.72.self_attn.q_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.72.self_attn.v_proj.bias": "model-00027-of-00031.safetensors",
+    "model.layers.72.self_attn.v_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.73.input_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.mlp.down_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.mlp.gate_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.mlp.up_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.post_attention_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.k_proj.bias": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.k_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.o_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.q_proj.bias": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.q_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.v_proj.bias": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.v_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.input_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.mlp.down_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.mlp.gate_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.mlp.up_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.post_attention_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.self_attn.k_proj.bias": "model-00028-of-00031.safetensors",
+    "model.layers.74.self_attn.k_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.self_attn.o_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.self_attn.q_proj.bias": "model-00028-of-00031.safetensors",
+    "model.layers.74.self_attn.q_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.self_attn.v_proj.bias": "model-00028-of-00031.safetensors",
+    "model.layers.74.self_attn.v_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.75.input_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.mlp.down_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.mlp.gate_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.75.mlp.up_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.post_attention_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.self_attn.k_proj.bias": "model-00028-of-00031.safetensors",
+    "model.layers.75.self_attn.k_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.75.self_attn.o_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.75.self_attn.q_proj.bias": "model-00028-of-00031.safetensors",
+    "model.layers.75.self_attn.q_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.75.self_attn.v_proj.bias": "model-00028-of-00031.safetensors",
+    "model.layers.75.self_attn.v_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.76.input_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.mlp.down_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.mlp.gate_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.mlp.up_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.post_attention_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.k_proj.bias": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.k_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.o_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.q_proj.bias": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.q_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.v_proj.bias": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.v_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.input_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.mlp.down_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.mlp.gate_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.mlp.up_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.post_attention_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.self_attn.k_proj.bias": "model-00029-of-00031.safetensors",
+    "model.layers.77.self_attn.k_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.self_attn.o_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.self_attn.q_proj.bias": "model-00029-of-00031.safetensors",
+    "model.layers.77.self_attn.q_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.self_attn.v_proj.bias": "model-00029-of-00031.safetensors",
+    "model.layers.77.self_attn.v_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.78.input_layernorm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.mlp.down_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.mlp.gate_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.mlp.up_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.post_attention_layernorm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.self_attn.k_proj.bias": "model-00029-of-00031.safetensors",
+    "model.layers.78.self_attn.k_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.78.self_attn.o_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.78.self_attn.q_proj.bias": "model-00029-of-00031.safetensors",
+    "model.layers.78.self_attn.q_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.78.self_attn.v_proj.bias": "model-00029-of-00031.safetensors",
+    "model.layers.78.self_attn.v_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.79.input_layernorm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.mlp.down_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.mlp.gate_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.mlp.up_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.post_attention_layernorm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.k_proj.bias": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.k_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.o_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.q_proj.bias": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.q_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.v_proj.bias": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.v_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-00004-of-00031.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.self_attn.q_proj.bias": "model-00004-of-00031.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.self_attn.v_proj.bias": "model-00004-of-00031.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.self_attn.k_proj.bias": "model-00004-of-00031.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.9.self_attn.q_proj.bias": "model-00004-of-00031.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.9.self_attn.v_proj.bias": "model-00004-of-00031.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00004-of-00031.safetensors",
+    "model.norm.weight": "model-00030-of-00031.safetensors"
+  }
+}

special_tokens_map.json

@@ -0,0 +1,20 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,44 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{% set system_message = 'Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n' %}{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '### Instruction:\n' + content + '\n\n### Response:\n' }}{% elif message['role'] == 'assistant' %}{{ content + '<|endoftext|>' + '\n\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.9985842378480414,
+    "total_flos": 5.875342281120154e+16,
+    "train_loss": 0.11385704865424108,
+    "train_runtime": 162519.6807,
+    "train_samples_per_second": 3.441,
+    "train_steps_per_second": 0.02
+}

trainer_log.jsonl

@@ -0,0 +1,291 @@
+{"current_steps": 10, "total_steps": 3177, "loss": 0.3368, "learning_rate": 2.0833333333333334e-06, "epoch": 0.009438414346389807, "percentage": 0.31, "elapsed_time": "0:08:46", "remaining_time": "1 day, 22:18:37", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 20, "total_steps": 3177, "loss": 0.2696, "learning_rate": 4.166666666666667e-06, "epoch": 0.018876828692779613, "percentage": 0.63, "elapsed_time": "0:17:15", "remaining_time": "1 day, 21:24:29", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 30, "total_steps": 3177, "loss": 0.2345, "learning_rate": 6.25e-06, "epoch": 0.028315243039169418, "percentage": 0.94, "elapsed_time": "0:25:44", "remaining_time": "1 day, 21:00:53", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 40, "total_steps": 3177, "loss": 0.213, "learning_rate": 8.333333333333334e-06, "epoch": 0.037753657385559226, "percentage": 1.26, "elapsed_time": "0:34:14", "remaining_time": "1 day, 20:45:20", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 50, "total_steps": 3177, "loss": 0.194, "learning_rate": 1.0416666666666668e-05, "epoch": 0.04719207173194903, "percentage": 1.57, "elapsed_time": "0:42:43", "remaining_time": "1 day, 20:32:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 60, "total_steps": 3177, "loss": 0.1871, "learning_rate": 1.25e-05, "epoch": 0.056630486078338836, "percentage": 1.89, "elapsed_time": "0:51:13", "remaining_time": "1 day, 20:20:59", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 70, "total_steps": 3177, "loss": 0.1791, "learning_rate": 1.4583333333333333e-05, "epoch": 0.06606890042472864, "percentage": 2.2, "elapsed_time": "0:59:42", "remaining_time": "1 day, 20:09:59", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 80, "total_steps": 3177, "loss": 0.1777, "learning_rate": 1.6666666666666667e-05, "epoch": 0.07550731477111845, "percentage": 2.52, "elapsed_time": "1:08:12", "remaining_time": "1 day, 20:00:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 90, "total_steps": 3177, "loss": 0.1743, "learning_rate": 1.8750000000000002e-05, "epoch": 0.08494572911750826, "percentage": 2.83, "elapsed_time": "1:16:41", "remaining_time": "1 day, 19:50:39", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 100, "total_steps": 3177, "loss": 0.173, "learning_rate": 1.9999916822524766e-05, "epoch": 0.09438414346389806, "percentage": 3.15, "elapsed_time": "1:25:11", "remaining_time": "1 day, 19:41:29", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 110, "total_steps": 3177, "loss": 0.1697, "learning_rate": 1.999898109181919e-05, "epoch": 0.10382255781028787, "percentage": 3.46, "elapsed_time": "1:33:41", "remaining_time": "1 day, 19:32:20", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 120, "total_steps": 3177, "loss": 0.1669, "learning_rate": 1.9997005756177228e-05, "epoch": 0.11326097215667767, "percentage": 3.78, "elapsed_time": "1:42:11", "remaining_time": "1 day, 19:23:10", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 130, "total_steps": 3177, "loss": 0.1664, "learning_rate": 1.999399102097668e-05, "epoch": 0.12269938650306748, "percentage": 4.09, "elapsed_time": "1:50:40", "remaining_time": "1 day, 19:14:02", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 140, "total_steps": 3177, "loss": 0.1652, "learning_rate": 1.9989937199662845e-05, "epoch": 0.13213780084945728, "percentage": 4.41, "elapsed_time": "1:59:09", "remaining_time": "1 day, 19:04:55", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 150, "total_steps": 3177, "loss": 0.1619, "learning_rate": 1.998484471371593e-05, "epoch": 0.1415762151958471, "percentage": 4.72, "elapsed_time": "2:07:39", "remaining_time": "1 day, 18:56:00", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 160, "total_steps": 3177, "loss": 0.1606, "learning_rate": 1.9978714092607234e-05, "epoch": 0.1510146295422369, "percentage": 5.04, "elapsed_time": "2:16:08", "remaining_time": "1 day, 18:47:04", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 170, "total_steps": 3177, "loss": 0.16, "learning_rate": 1.9971545973744102e-05, "epoch": 0.16045304388862672, "percentage": 5.35, "elapsed_time": "2:24:37", "remaining_time": "1 day, 18:38:09", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 180, "total_steps": 3177, "loss": 0.1582, "learning_rate": 1.9963341102403652e-05, "epoch": 0.16989145823501653, "percentage": 5.67, "elapsed_time": "2:33:06", "remaining_time": "1 day, 18:29:20", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 190, "total_steps": 3177, "loss": 0.1551, "learning_rate": 1.9954100331655265e-05, "epoch": 0.1793298725814063, "percentage": 5.98, "elapsed_time": "2:41:36", "remaining_time": "1 day, 18:20:34", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 200, "total_steps": 3177, "loss": 0.1559, "learning_rate": 1.9943824622271934e-05, "epoch": 0.18876828692779613, "percentage": 6.3, "elapsed_time": "2:50:05", "remaining_time": "1 day, 18:11:50", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 210, "total_steps": 3177, "loss": 0.1534, "learning_rate": 1.9932515042630335e-05, "epoch": 0.19820670127418594, "percentage": 6.61, "elapsed_time": "2:58:34", "remaining_time": "1 day, 18:03:07", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 220, "total_steps": 3177, "loss": 0.1545, "learning_rate": 1.9920172768599763e-05, "epoch": 0.20764511562057575, "percentage": 6.92, "elapsed_time": "3:07:04", "remaining_time": "1 day, 17:54:26", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 230, "total_steps": 3177, "loss": 0.1543, "learning_rate": 1.9906799083419865e-05, "epoch": 0.21708352996696556, "percentage": 7.24, "elapsed_time": "3:15:33", "remaining_time": "1 day, 17:45:43", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 240, "total_steps": 3177, "loss": 0.1538, "learning_rate": 1.989239537756723e-05, "epoch": 0.22652194431335534, "percentage": 7.55, "elapsed_time": "3:24:03", "remaining_time": "1 day, 17:37:04", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 250, "total_steps": 3177, "loss": 0.1541, "learning_rate": 1.987696314861082e-05, "epoch": 0.23596035865974516, "percentage": 7.87, "elapsed_time": "3:32:32", "remaining_time": "1 day, 17:28:22", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 260, "total_steps": 3177, "loss": 0.1518, "learning_rate": 1.986050400105626e-05, "epoch": 0.24539877300613497, "percentage": 8.18, "elapsed_time": "3:41:01", "remaining_time": "1 day, 17:19:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 270, "total_steps": 3177, "loss": 0.1501, "learning_rate": 1.9843019646179014e-05, "epoch": 0.25483718735252475, "percentage": 8.5, "elapsed_time": "3:49:31", "remaining_time": "1 day, 17:11:11", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 280, "total_steps": 3177, "loss": 0.1483, "learning_rate": 1.9824511901846475e-05, "epoch": 0.26427560169891456, "percentage": 8.81, "elapsed_time": "3:58:00", "remaining_time": "1 day, 17:02:33", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 290, "total_steps": 3177, "loss": 0.1514, "learning_rate": 1.9804982692328944e-05, "epoch": 0.2737140160453044, "percentage": 9.13, "elapsed_time": "4:06:30", "remaining_time": "1 day, 16:53:58", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 300, "total_steps": 3177, "loss": 0.151, "learning_rate": 1.9784434048099565e-05, "epoch": 0.2831524303916942, "percentage": 9.44, "elapsed_time": "4:14:59", "remaining_time": "1 day, 16:45:25", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 310, "total_steps": 3177, "loss": 0.1498, "learning_rate": 1.976286810562323e-05, "epoch": 0.292590844738084, "percentage": 9.76, "elapsed_time": "4:23:29", "remaining_time": "1 day, 16:36:48", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 320, "total_steps": 3177, "loss": 0.1513, "learning_rate": 1.9740287107134417e-05, "epoch": 0.3020292590844738, "percentage": 10.07, "elapsed_time": "4:31:58", "remaining_time": "1 day, 16:28:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 330, "total_steps": 3177, "loss": 0.1489, "learning_rate": 1.97166934004041e-05, "epoch": 0.3114676734308636, "percentage": 10.39, "elapsed_time": "4:40:28", "remaining_time": "1 day, 16:19:43", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 340, "total_steps": 3177, "loss": 0.1449, "learning_rate": 1.9692089438495622e-05, "epoch": 0.32090608777725343, "percentage": 10.7, "elapsed_time": "4:48:58", "remaining_time": "1 day, 16:11:10", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 350, "total_steps": 3177, "loss": 0.1469, "learning_rate": 1.9666477779509655e-05, "epoch": 0.33034450212364325, "percentage": 11.02, "elapsed_time": "4:57:28", "remaining_time": "1 day, 16:02:44", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 360, "total_steps": 3177, "loss": 0.1468, "learning_rate": 1.963986108631823e-05, "epoch": 0.33978291647003306, "percentage": 11.33, "elapsed_time": "5:05:58", "remaining_time": "1 day, 15:54:11", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 370, "total_steps": 3177, "loss": 0.1483, "learning_rate": 1.9612242126287876e-05, "epoch": 0.3492213308164228, "percentage": 11.65, "elapsed_time": "5:14:27", "remaining_time": "1 day, 15:45:36", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 380, "total_steps": 3177, "loss": 0.1443, "learning_rate": 1.958362377099191e-05, "epoch": 0.3586597451628126, "percentage": 11.96, "elapsed_time": "5:22:56", "remaining_time": "1 day, 15:37:01", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 390, "total_steps": 3177, "loss": 0.1463, "learning_rate": 1.9554008995911837e-05, "epoch": 0.36809815950920244, "percentage": 12.28, "elapsed_time": "5:31:25", "remaining_time": "1 day, 15:28:26", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 400, "total_steps": 3177, "loss": 0.1471, "learning_rate": 1.9523400880128032e-05, "epoch": 0.37753657385559225, "percentage": 12.59, "elapsed_time": "5:39:54", "remaining_time": "1 day, 15:19:51", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 410, "total_steps": 3177, "loss": 0.1452, "learning_rate": 1.949180260599957e-05, "epoch": 0.38697498820198206, "percentage": 12.91, "elapsed_time": "5:48:24", "remaining_time": "1 day, 15:11:17", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 420, "total_steps": 3177, "loss": 0.1455, "learning_rate": 1.945921745883337e-05, "epoch": 0.3964134025483719, "percentage": 13.22, "elapsed_time": "5:56:53", "remaining_time": "1 day, 15:02:44", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 430, "total_steps": 3177, "loss": 0.1435, "learning_rate": 1.9425648826542618e-05, "epoch": 0.4058518168947617, "percentage": 13.53, "elapsed_time": "6:05:22", "remaining_time": "1 day, 14:54:09", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 440, "total_steps": 3177, "loss": 0.1436, "learning_rate": 1.939110019929451e-05, "epoch": 0.4152902312411515, "percentage": 13.85, "elapsed_time": "6:13:52", "remaining_time": "1 day, 14:45:37", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 450, "total_steps": 3177, "loss": 0.1451, "learning_rate": 1.935557516914739e-05, "epoch": 0.4247286455875413, "percentage": 14.16, "elapsed_time": "6:22:21", "remaining_time": "1 day, 14:37:03", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 460, "total_steps": 3177, "loss": 0.1444, "learning_rate": 1.931907742967727e-05, "epoch": 0.4341670599339311, "percentage": 14.48, "elapsed_time": "6:30:50", "remaining_time": "1 day, 14:28:32", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 470, "total_steps": 3177, "loss": 0.142, "learning_rate": 1.92816107755938e-05, "epoch": 0.44360547428032093, "percentage": 14.79, "elapsed_time": "6:39:19", "remaining_time": "1 day, 14:19:58", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 480, "total_steps": 3177, "loss": 0.1406, "learning_rate": 1.9243179102345753e-05, "epoch": 0.4530438886267107, "percentage": 15.11, "elapsed_time": "6:47:49", "remaining_time": "1 day, 14:11:26", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 490, "total_steps": 3177, "loss": 0.144, "learning_rate": 1.9203786405715984e-05, "epoch": 0.4624823029731005, "percentage": 15.42, "elapsed_time": "6:56:18", "remaining_time": "1 day, 14:02:54", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 500, "total_steps": 3177, "loss": 0.1428, "learning_rate": 1.9163436781405992e-05, "epoch": 0.4719207173194903, "percentage": 15.74, "elapsed_time": "7:04:48", "remaining_time": "1 day, 13:54:24", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 510, "total_steps": 3177, "loss": 0.1399, "learning_rate": 1.912213442461009e-05, "epoch": 0.4813591316658801, "percentage": 16.05, "elapsed_time": "7:13:17", "remaining_time": "1 day, 13:45:53", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 520, "total_steps": 3177, "loss": 0.1396, "learning_rate": 1.9079883629579224e-05, "epoch": 0.49079754601226994, "percentage": 16.37, "elapsed_time": "7:21:47", "remaining_time": "1 day, 13:37:24", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 530, "total_steps": 3177, "loss": 0.1403, "learning_rate": 1.9036688789174496e-05, "epoch": 0.5002359603586597, "percentage": 16.68, "elapsed_time": "7:30:17", "remaining_time": "1 day, 13:28:54", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 540, "total_steps": 3177, "loss": 0.1399, "learning_rate": 1.899255439441043e-05, "epoch": 0.5096743747050495, "percentage": 17.0, "elapsed_time": "7:38:46", "remaining_time": "1 day, 13:20:23", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 550, "total_steps": 3177, "loss": 0.1376, "learning_rate": 1.8947485033988034e-05, "epoch": 0.5191127890514393, "percentage": 17.31, "elapsed_time": "7:47:16", "remaining_time": "1 day, 13:11:51", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 560, "total_steps": 3177, "loss": 0.1415, "learning_rate": 1.8901485393817724e-05, "epoch": 0.5285512033978291, "percentage": 17.63, "elapsed_time": "7:55:46", "remaining_time": "1 day, 13:03:23", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 570, "total_steps": 3177, "loss": 0.1423, "learning_rate": 1.8854560256532098e-05, "epoch": 0.5379896177442189, "percentage": 17.94, "elapsed_time": "8:04:15", "remaining_time": "1 day, 12:54:51", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 580, "total_steps": 3177, "loss": 0.139, "learning_rate": 1.880671450098871e-05, "epoch": 0.5474280320906088, "percentage": 18.26, "elapsed_time": "8:12:45", "remaining_time": "1 day, 12:46:20", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 590, "total_steps": 3177, "loss": 0.1396, "learning_rate": 1.8757953101762786e-05, "epoch": 0.5568664464369986, "percentage": 18.57, "elapsed_time": "8:21:14", "remaining_time": "1 day, 12:37:48", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 600, "total_steps": 3177, "loss": 0.138, "learning_rate": 1.8708281128630023e-05, "epoch": 0.5663048607833884, "percentage": 18.89, "elapsed_time": "8:29:44", "remaining_time": "1 day, 12:29:20", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 610, "total_steps": 3177, "loss": 0.1406, "learning_rate": 1.865770374603948e-05, "epoch": 0.5757432751297782, "percentage": 19.2, "elapsed_time": "8:38:14", "remaining_time": "1 day, 12:20:52", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 620, "total_steps": 3177, "loss": 0.138, "learning_rate": 1.8606226212576612e-05, "epoch": 0.585181689476168, "percentage": 19.52, "elapsed_time": "8:46:44", "remaining_time": "1 day, 12:12:21", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 630, "total_steps": 3177, "loss": 0.1408, "learning_rate": 1.8553853880416555e-05, "epoch": 0.5946201038225578, "percentage": 19.83, "elapsed_time": "8:55:13", "remaining_time": "1 day, 12:03:50", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 640, "total_steps": 3177, "loss": 0.1394, "learning_rate": 1.8500592194767625e-05, "epoch": 0.6040585181689476, "percentage": 20.14, "elapsed_time": "9:03:42", "remaining_time": "1 day, 11:55:18", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 650, "total_steps": 3177, "loss": 0.1384, "learning_rate": 1.8446446693305194e-05, "epoch": 0.6134969325153374, "percentage": 20.46, "elapsed_time": "9:12:11", "remaining_time": "1 day, 11:46:46", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 660, "total_steps": 3177, "loss": 0.1393, "learning_rate": 1.8391423005595928e-05, "epoch": 0.6229353468617272, "percentage": 20.77, "elapsed_time": "9:20:41", "remaining_time": "1 day, 11:38:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 670, "total_steps": 3177, "loss": 0.1398, "learning_rate": 1.833552685251246e-05, "epoch": 0.6323737612081171, "percentage": 21.09, "elapsed_time": "9:29:10", "remaining_time": "1 day, 11:29:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 680, "total_steps": 3177, "loss": 0.1369, "learning_rate": 1.827876404563861e-05, "epoch": 0.6418121755545069, "percentage": 21.4, "elapsed_time": "9:37:40", "remaining_time": "1 day, 11:21:14", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 690, "total_steps": 3177, "loss": 0.1346, "learning_rate": 1.8221140486665125e-05, "epoch": 0.6512505899008967, "percentage": 21.72, "elapsed_time": "9:46:09", "remaining_time": "1 day, 11:12:42", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 700, "total_steps": 3177, "loss": 0.1357, "learning_rate": 1.8162662166776085e-05, "epoch": 0.6606890042472865, "percentage": 22.03, "elapsed_time": "9:54:39", "remaining_time": "1 day, 11:04:12", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 710, "total_steps": 3177, "loss": 0.1389, "learning_rate": 1.8103335166026002e-05, "epoch": 0.6701274185936763, "percentage": 22.35, "elapsed_time": "10:03:08", "remaining_time": "1 day, 10:55:42", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 720, "total_steps": 3177, "loss": 0.1375, "learning_rate": 1.804316565270765e-05, "epoch": 0.6795658329400661, "percentage": 22.66, "elapsed_time": "10:11:38", "remaining_time": "1 day, 10:47:13", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 730, "total_steps": 3177, "loss": 0.1364, "learning_rate": 1.798215988271075e-05, "epoch": 0.6890042472864559, "percentage": 22.98, "elapsed_time": "10:20:07", "remaining_time": "1 day, 10:38:42", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 740, "total_steps": 3177, "loss": 0.138, "learning_rate": 1.7920324198871546e-05, "epoch": 0.6984426616328456, "percentage": 23.29, "elapsed_time": "10:28:37", "remaining_time": "1 day, 10:30:13", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 750, "total_steps": 3177, "loss": 0.1346, "learning_rate": 1.785766503031332e-05, "epoch": 0.7078810759792354, "percentage": 23.61, "elapsed_time": "10:37:07", "remaining_time": "1 day, 10:21:43", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 760, "total_steps": 3177, "loss": 0.1352, "learning_rate": 1.7794188891777964e-05, "epoch": 0.7173194903256253, "percentage": 23.92, "elapsed_time": "10:45:37", "remaining_time": "1 day, 10:13:14", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 770, "total_steps": 3177, "loss": 0.1353, "learning_rate": 1.7729902382948617e-05, "epoch": 0.7267579046720151, "percentage": 24.24, "elapsed_time": "10:54:07", "remaining_time": "1 day, 10:04:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 780, "total_steps": 3177, "loss": 0.1352, "learning_rate": 1.76648121877635e-05, "epoch": 0.7361963190184049, "percentage": 24.55, "elapsed_time": "11:02:36", "remaining_time": "1 day, 9:56:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 790, "total_steps": 3177, "loss": 0.1341, "learning_rate": 1.759892507372099e-05, "epoch": 0.7456347333647947, "percentage": 24.87, "elapsed_time": "11:11:06", "remaining_time": "1 day, 9:47:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 800, "total_steps": 3177, "loss": 0.1333, "learning_rate": 1.7532247891175968e-05, "epoch": 0.7550731477111845, "percentage": 25.18, "elapsed_time": "11:19:36", "remaining_time": "1 day, 9:39:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 810, "total_steps": 3177, "loss": 0.136, "learning_rate": 1.746478757262761e-05, "epoch": 0.7645115620575743, "percentage": 25.5, "elapsed_time": "11:28:05", "remaining_time": "1 day, 9:30:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 820, "total_steps": 3177, "loss": 0.1336, "learning_rate": 1.739655113199858e-05, "epoch": 0.7739499764039641, "percentage": 25.81, "elapsed_time": "11:36:35", "remaining_time": "1 day, 9:22:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 830, "total_steps": 3177, "loss": 0.1331, "learning_rate": 1.7327545663905813e-05, "epoch": 0.7833883907503539, "percentage": 26.13, "elapsed_time": "11:45:05", "remaining_time": "1 day, 9:13:46", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 840, "total_steps": 3177, "loss": 0.1328, "learning_rate": 1.7257778342922853e-05, "epoch": 0.7928268050967437, "percentage": 26.44, "elapsed_time": "11:53:34", "remaining_time": "1 day, 9:05:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 850, "total_steps": 3177, "loss": 0.1319, "learning_rate": 1.7187256422833928e-05, "epoch": 0.8022652194431336, "percentage": 26.75, "elapsed_time": "12:02:03", "remaining_time": "1 day, 8:56:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 860, "total_steps": 3177, "loss": 0.1324, "learning_rate": 1.711598723587975e-05, "epoch": 0.8117036337895234, "percentage": 27.07, "elapsed_time": "12:10:33", "remaining_time": "1 day, 8:48:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 870, "total_steps": 3177, "loss": 0.1325, "learning_rate": 1.7043978191995177e-05, "epoch": 0.8211420481359132, "percentage": 27.38, "elapsed_time": "12:19:03", "remaining_time": "1 day, 8:39:46", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 880, "total_steps": 3177, "loss": 0.1315, "learning_rate": 1.6971236778038806e-05, "epoch": 0.830580462482303, "percentage": 27.7, "elapsed_time": "12:27:32", "remaining_time": "1 day, 8:31:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 890, "total_steps": 3177, "loss": 0.1328, "learning_rate": 1.6897770557014535e-05, "epoch": 0.8400188768286928, "percentage": 28.01, "elapsed_time": "12:36:02", "remaining_time": "1 day, 8:22:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 900, "total_steps": 3177, "loss": 0.1351, "learning_rate": 1.682358716728525e-05, "epoch": 0.8494572911750826, "percentage": 28.33, "elapsed_time": "12:44:32", "remaining_time": "1 day, 8:14:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 910, "total_steps": 3177, "loss": 0.1325, "learning_rate": 1.674869432177864e-05, "epoch": 0.8588957055214724, "percentage": 28.64, "elapsed_time": "12:53:01", "remaining_time": "1 day, 8:05:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 920, "total_steps": 3177, "loss": 0.1312, "learning_rate": 1.667309980718529e-05, "epoch": 0.8683341198678622, "percentage": 28.96, "elapsed_time": "13:01:30", "remaining_time": "1 day, 7:57:14", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 930, "total_steps": 3177, "loss": 0.1317, "learning_rate": 1.6596811483149077e-05, "epoch": 0.877772534214252, "percentage": 29.27, "elapsed_time": "13:10:00", "remaining_time": "1 day, 7:48:44", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 940, "total_steps": 3177, "loss": 0.1355, "learning_rate": 1.651983728145e-05, "epoch": 0.8872109485606419, "percentage": 29.59, "elapsed_time": "13:18:29", "remaining_time": "1 day, 7:40:14", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 950, "total_steps": 3177, "loss": 0.1309, "learning_rate": 1.6442185205179507e-05, "epoch": 0.8966493629070316, "percentage": 29.9, "elapsed_time": "13:26:59", "remaining_time": "1 day, 7:31:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 960, "total_steps": 3177, "loss": 0.1339, "learning_rate": 1.6363863327908405e-05, "epoch": 0.9060877772534214, "percentage": 30.22, "elapsed_time": "13:35:29", "remaining_time": "1 day, 7:23:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 970, "total_steps": 3177, "loss": 0.1299, "learning_rate": 1.6284879792847433e-05, "epoch": 0.9155261915998112, "percentage": 30.53, "elapsed_time": "13:43:58", "remaining_time": "1 day, 7:14:46", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 980, "total_steps": 3177, "loss": 0.1305, "learning_rate": 1.620524281200062e-05, "epoch": 0.924964605946201, "percentage": 30.85, "elapsed_time": "13:52:28", "remaining_time": "1 day, 7:06:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 990, "total_steps": 3177, "loss": 0.1322, "learning_rate": 1.6124960665311447e-05, "epoch": 0.9344030202925908, "percentage": 31.16, "elapsed_time": "14:00:57", "remaining_time": "1 day, 6:57:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1000, "total_steps": 3177, "loss": 0.129, "learning_rate": 1.6044041699802005e-05, "epoch": 0.9438414346389806, "percentage": 31.48, "elapsed_time": "14:09:27", "remaining_time": "1 day, 6:49:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1010, "total_steps": 3177, "loss": 0.1321, "learning_rate": 1.5962494328705123e-05, "epoch": 0.9532798489853704, "percentage": 31.79, "elapsed_time": "14:17:56", "remaining_time": "1 day, 6:40:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1020, "total_steps": 3177, "loss": 0.1334, "learning_rate": 1.588032703058964e-05, "epoch": 0.9627182633317602, "percentage": 32.11, "elapsed_time": "14:26:26", "remaining_time": "1 day, 6:32:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1030, "total_steps": 3177, "loss": 0.1325, "learning_rate": 1.5797548348478893e-05, "epoch": 0.9721566776781501, "percentage": 32.42, "elapsed_time": "14:34:56", "remaining_time": "1 day, 6:23:47", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1040, "total_steps": 3177, "loss": 0.132, "learning_rate": 1.571416688896246e-05, "epoch": 0.9815950920245399, "percentage": 32.74, "elapsed_time": "14:43:26", "remaining_time": "1 day, 6:15:17", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1050, "total_steps": 3177, "loss": 0.1301, "learning_rate": 1.563019132130136e-05, "epoch": 0.9910335063709297, "percentage": 33.05, "elapsed_time": "14:51:55", "remaining_time": "1 day, 6:06:47", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1060, "total_steps": 3177, "loss": 0.1282, "learning_rate": 1.5545630376526665e-05, "epoch": 1.0004719207173194, "percentage": 33.36, "elapsed_time": "15:00:32", "remaining_time": "1 day, 5:58:31", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1070, "total_steps": 3177, "loss": 0.1142, "learning_rate": 1.5460492846531748e-05, "epoch": 1.0099103350637093, "percentage": 33.68, "elapsed_time": "15:09:01", "remaining_time": "1 day, 5:50:01", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1080, "total_steps": 3177, "loss": 0.1157, "learning_rate": 1.5374787583158188e-05, "epoch": 1.019348749410099, "percentage": 33.99, "elapsed_time": "15:17:31", "remaining_time": "1 day, 5:41:31", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1090, "total_steps": 3177, "loss": 0.1143, "learning_rate": 1.5288523497275392e-05, "epoch": 1.028787163756489, "percentage": 34.31, "elapsed_time": "15:26:01", "remaining_time": "1 day, 5:33:01", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1100, "total_steps": 3177, "loss": 0.1128, "learning_rate": 1.5201709557854178e-05, "epoch": 1.0382255781028786, "percentage": 34.62, "elapsed_time": "15:34:31", "remaining_time": "1 day, 5:24:32", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1110, "total_steps": 3177, "loss": 0.1161, "learning_rate": 1.5114354791034225e-05, "epoch": 1.0476639924492686, "percentage": 34.94, "elapsed_time": "15:43:01", "remaining_time": "1 day, 5:16:03", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1120, "total_steps": 3177, "loss": 0.1159, "learning_rate": 1.5026468279185615e-05, "epoch": 1.0571024067956583, "percentage": 35.25, "elapsed_time": "15:51:32", "remaining_time": "1 day, 5:07:35", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1130, "total_steps": 3177, "loss": 0.1161, "learning_rate": 1.4938059159964555e-05, "epoch": 1.0665408211420482, "percentage": 35.57, "elapsed_time": "16:00:01", "remaining_time": "1 day, 4:59:05", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1140, "total_steps": 3177, "loss": 0.1141, "learning_rate": 1.4849136625363297e-05, "epoch": 1.0759792354884379, "percentage": 35.88, "elapsed_time": "16:13:14", "remaining_time": "1 day, 4:59:01", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1150, "total_steps": 3177, "loss": 0.1125, "learning_rate": 1.4759709920754453e-05, "epoch": 1.0854176498348278, "percentage": 36.2, "elapsed_time": "16:22:01", "remaining_time": "1 day, 4:50:56", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1160, "total_steps": 3177, "loss": 0.1141, "learning_rate": 1.4669788343929736e-05, "epoch": 1.0948560641812175, "percentage": 36.51, "elapsed_time": "16:30:32", "remaining_time": "1 day, 4:42:20", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1170, "total_steps": 3177, "loss": 0.1128, "learning_rate": 1.4579381244133265e-05, "epoch": 1.1042944785276074, "percentage": 36.83, "elapsed_time": "16:39:01", "remaining_time": "1 day, 4:33:42", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1180, "total_steps": 3177, "loss": 0.1137, "learning_rate": 1.4488498021089514e-05, "epoch": 1.1137328928739971, "percentage": 37.14, "elapsed_time": "16:47:30", "remaining_time": "1 day, 4:25:05", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1190, "total_steps": 3177, "loss": 0.1143, "learning_rate": 1.4397148124025997e-05, "epoch": 1.123171307220387, "percentage": 37.46, "elapsed_time": "16:56:01", "remaining_time": "1 day, 4:16:30", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1200, "total_steps": 3177, "loss": 0.117, "learning_rate": 1.4305341050690845e-05, "epoch": 1.1326097215667768, "percentage": 37.77, "elapsed_time": "17:04:30", "remaining_time": "1 day, 4:07:53", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1210, "total_steps": 3177, "loss": 0.1137, "learning_rate": 1.421308634636529e-05, "epoch": 1.1420481359131667, "percentage": 38.09, "elapsed_time": "17:13:01", "remaining_time": "1 day, 3:59:17", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1220, "total_steps": 3177, "loss": 0.1145, "learning_rate": 1.412039360287126e-05, "epoch": 1.1514865502595564, "percentage": 38.4, "elapsed_time": "17:21:31", "remaining_time": "1 day, 3:50:41", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1230, "total_steps": 3177, "loss": 0.1138, "learning_rate": 1.4027272457574082e-05, "epoch": 1.1609249646059463, "percentage": 38.72, "elapsed_time": "17:30:00", "remaining_time": "1 day, 3:42:05", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1240, "total_steps": 3177, "loss": 0.1135, "learning_rate": 1.3933732592380485e-05, "epoch": 1.170363378952336, "percentage": 39.03, "elapsed_time": "17:38:31", "remaining_time": "1 day, 3:33:30", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1250, "total_steps": 3177, "loss": 0.1134, "learning_rate": 1.3839783732731966e-05, "epoch": 1.1798017932987257, "percentage": 39.35, "elapsed_time": "17:47:00", "remaining_time": "1 day, 3:24:53", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1260, "total_steps": 3177, "loss": 0.1136, "learning_rate": 1.3745435646593613e-05, "epoch": 1.1892402076451156, "percentage": 39.66, "elapsed_time": "17:55:30", "remaining_time": "1 day, 3:16:18", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1270, "total_steps": 3177, "loss": 0.113, "learning_rate": 1.3650698143438534e-05, "epoch": 1.1986786219915055, "percentage": 39.97, "elapsed_time": "18:04:00", "remaining_time": "1 day, 3:07:43", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1280, "total_steps": 3177, "loss": 0.1167, "learning_rate": 1.3555581073227942e-05, "epoch": 1.2081170363378952, "percentage": 40.29, "elapsed_time": "18:12:30", "remaining_time": "1 day, 2:59:08", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1290, "total_steps": 3177, "loss": 0.1147, "learning_rate": 1.346009432538705e-05, "epoch": 1.217555450684285, "percentage": 40.6, "elapsed_time": "18:21:00", "remaining_time": "1 day, 2:50:32", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1300, "total_steps": 3177, "loss": 0.1145, "learning_rate": 1.3364247827776854e-05, "epoch": 1.2269938650306749, "percentage": 40.92, "elapsed_time": "18:29:29", "remaining_time": "1 day, 2:41:56", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1310, "total_steps": 3177, "loss": 0.1141, "learning_rate": 1.3268051545661937e-05, "epoch": 1.2364322793770646, "percentage": 41.23, "elapsed_time": "18:37:59", "remaining_time": "1 day, 2:33:21", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1320, "total_steps": 3177, "loss": 0.1122, "learning_rate": 1.3171515480674342e-05, "epoch": 1.2458706937234545, "percentage": 41.55, "elapsed_time": "18:46:29", "remaining_time": "1 day, 2:24:46", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1330, "total_steps": 3177, "loss": 0.1173, "learning_rate": 1.3074649669773716e-05, "epoch": 1.2553091080698442, "percentage": 41.86, "elapsed_time": "18:54:59", "remaining_time": "1 day, 2:16:11", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1340, "total_steps": 3177, "loss": 0.1103, "learning_rate": 1.297746418420374e-05, "epoch": 1.2647475224162341, "percentage": 42.18, "elapsed_time": "19:03:29", "remaining_time": "1 day, 2:07:36", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1350, "total_steps": 3177, "loss": 0.1122, "learning_rate": 1.2879969128445025e-05, "epoch": 1.2741859367626238, "percentage": 42.49, "elapsed_time": "19:11:59", "remaining_time": "1 day, 1:59:01", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1360, "total_steps": 3177, "loss": 0.1112, "learning_rate": 1.2782174639164528e-05, "epoch": 1.2836243511090137, "percentage": 42.81, "elapsed_time": "19:20:28", "remaining_time": "1 day, 1:50:25", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1370, "total_steps": 3177, "loss": 0.1125, "learning_rate": 1.2684090884161636e-05, "epoch": 1.2930627654554034, "percentage": 43.12, "elapsed_time": "19:28:58", "remaining_time": "1 day, 1:41:51", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1380, "total_steps": 3177, "loss": 0.1107, "learning_rate": 1.2585728061311003e-05, "epoch": 1.3025011798017934, "percentage": 43.44, "elapsed_time": "19:37:28", "remaining_time": "1 day, 1:33:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1390, "total_steps": 3177, "loss": 0.1122, "learning_rate": 1.248709639750228e-05, "epoch": 1.311939594148183, "percentage": 43.75, "elapsed_time": "19:45:58", "remaining_time": "1 day, 1:24:41", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1400, "total_steps": 3177, "loss": 0.1124, "learning_rate": 1.2388206147576796e-05, "epoch": 1.321378008494573, "percentage": 44.07, "elapsed_time": "19:54:28", "remaining_time": "1 day, 1:16:07", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1410, "total_steps": 3177, "loss": 0.1151, "learning_rate": 1.2289067593261358e-05, "epoch": 1.3308164228409627, "percentage": 44.38, "elapsed_time": "20:02:57", "remaining_time": "1 day, 1:07:32", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1420, "total_steps": 3177, "loss": 0.1124, "learning_rate": 1.2189691042099265e-05, "epoch": 1.3402548371873526, "percentage": 44.7, "elapsed_time": "20:11:27", "remaining_time": "1 day, 0:58:58", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1430, "total_steps": 3177, "loss": 0.1154, "learning_rate": 1.209008682637859e-05, "epoch": 1.3496932515337423, "percentage": 45.01, "elapsed_time": "20:19:57", "remaining_time": "1 day, 0:50:23", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1440, "total_steps": 3177, "loss": 0.1127, "learning_rate": 1.1990265302057948e-05, "epoch": 1.359131665880132, "percentage": 45.33, "elapsed_time": "20:28:26", "remaining_time": "1 day, 0:41:48", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1450, "total_steps": 3177, "loss": 0.1134, "learning_rate": 1.1890236847689762e-05, "epoch": 1.368570080226522, "percentage": 45.64, "elapsed_time": "20:36:56", "remaining_time": "1 day, 0:33:14", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1460, "total_steps": 3177, "loss": 0.1145, "learning_rate": 1.1790011863341197e-05, "epoch": 1.3780084945729119, "percentage": 45.96, "elapsed_time": "20:45:26", "remaining_time": "1 day, 0:24:40", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1470, "total_steps": 3177, "loss": 0.1128, "learning_rate": 1.1689600769512855e-05, "epoch": 1.3874469089193016, "percentage": 46.27, "elapsed_time": "20:53:56", "remaining_time": "1 day, 0:16:06", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1480, "total_steps": 3177, "loss": 0.1158, "learning_rate": 1.1589014006055337e-05, "epoch": 1.3968853232656913, "percentage": 46.58, "elapsed_time": "21:02:26", "remaining_time": "1 day, 0:07:32", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1490, "total_steps": 3177, "loss": 0.1107, "learning_rate": 1.1488262031083816e-05, "epoch": 1.4063237376120812, "percentage": 46.9, "elapsed_time": "21:10:55", "remaining_time": "23:58:58", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1500, "total_steps": 3177, "loss": 0.1138, "learning_rate": 1.1387355319890685e-05, "epoch": 1.415762151958471, "percentage": 47.21, "elapsed_time": "21:19:26", "remaining_time": "23:50:24", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1510, "total_steps": 3177, "loss": 0.112, "learning_rate": 1.1286304363856418e-05, "epoch": 1.4252005663048608, "percentage": 47.53, "elapsed_time": "21:27:55", "remaining_time": "23:41:50", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1520, "total_steps": 3177, "loss": 0.1137, "learning_rate": 1.1185119669358792e-05, "epoch": 1.4346389806512505, "percentage": 47.84, "elapsed_time": "21:36:25", "remaining_time": "23:33:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1530, "total_steps": 3177, "loss": 0.1093, "learning_rate": 1.1083811756680523e-05, "epoch": 1.4440773949976404, "percentage": 48.16, "elapsed_time": "21:44:55", "remaining_time": "23:24:42", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1540, "total_steps": 3177, "loss": 0.1138, "learning_rate": 1.0982391158915441e-05, "epoch": 1.4535158093440301, "percentage": 48.47, "elapsed_time": "21:53:24", "remaining_time": "23:16:08", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1550, "total_steps": 3177, "loss": 0.1135, "learning_rate": 1.0880868420873375e-05, "epoch": 1.46295422369042, "percentage": 48.79, "elapsed_time": "22:01:54", "remaining_time": "23:07:34", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1560, "total_steps": 3177, "loss": 0.1104, "learning_rate": 1.0779254097983788e-05, "epoch": 1.4723926380368098, "percentage": 49.1, "elapsed_time": "22:10:24", "remaining_time": "22:59:00", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1570, "total_steps": 3177, "loss": 0.114, "learning_rate": 1.0677558755198327e-05, "epoch": 1.4818310523831997, "percentage": 49.42, "elapsed_time": "22:18:54", "remaining_time": "22:50:27", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1580, "total_steps": 3177, "loss": 0.1112, "learning_rate": 1.0575792965892349e-05, "epoch": 1.4912694667295894, "percentage": 49.73, "elapsed_time": "22:27:24", "remaining_time": "22:41:54", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1590, "total_steps": 3177, "loss": 0.1099, "learning_rate": 1.0473967310765629e-05, "epoch": 1.500707881075979, "percentage": 50.05, "elapsed_time": "22:35:54", "remaining_time": "22:33:21", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1600, "total_steps": 3177, "loss": 0.1109, "learning_rate": 1.0372092376742247e-05, "epoch": 1.510146295422369, "percentage": 50.36, "elapsed_time": "22:44:24", "remaining_time": "22:24:47", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1610, "total_steps": 3177, "loss": 0.1123, "learning_rate": 1.0270178755869861e-05, "epoch": 1.519584709768759, "percentage": 50.68, "elapsed_time": "22:52:54", "remaining_time": "22:16:14", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1620, "total_steps": 3177, "loss": 0.1088, "learning_rate": 1.0168237044218452e-05, "epoch": 1.5290231241151486, "percentage": 50.99, "elapsed_time": "23:01:23", "remaining_time": "22:07:40", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1630, "total_steps": 3177, "loss": 0.1125, "learning_rate": 1.0066277840778626e-05, "epoch": 1.5384615384615383, "percentage": 51.31, "elapsed_time": "23:09:53", "remaining_time": "21:59:07", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1640, "total_steps": 3177, "loss": 0.1078, "learning_rate": 9.964311746359631e-06, "epoch": 1.5478999528079282, "percentage": 51.62, "elapsed_time": "23:18:23", "remaining_time": "21:50:34", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1650, "total_steps": 3177, "loss": 0.108, "learning_rate": 9.862349362487172e-06, "epoch": 1.5573383671543182, "percentage": 51.94, "elapsed_time": "23:26:53", "remaining_time": "21:42:01", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1660, "total_steps": 3177, "loss": 0.1073, "learning_rate": 9.760401290301164e-06, "epoch": 1.5667767815007079, "percentage": 52.25, "elapsed_time": "23:35:23", "remaining_time": "21:33:28", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1670, "total_steps": 3177, "loss": 0.1095, "learning_rate": 9.658478129453532e-06, "epoch": 1.5762151958470976, "percentage": 52.57, "elapsed_time": "23:43:53", "remaining_time": "21:24:54", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1680, "total_steps": 3177, "loss": 0.109, "learning_rate": 9.556590477006123e-06, "epoch": 1.5856536101934875, "percentage": 52.88, "elapsed_time": "23:52:23", "remaining_time": "21:16:21", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1690, "total_steps": 3177, "loss": 0.1111, "learning_rate": 9.454748926328962e-06, "epoch": 1.5950920245398774, "percentage": 53.19, "elapsed_time": "1 day, 0:00:53", "remaining_time": "21:07:48", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1700, "total_steps": 3177, "loss": 0.1091, "learning_rate": 9.352964065998801e-06, "epoch": 1.6045304388862671, "percentage": 53.51, "elapsed_time": "1 day, 0:09:22", "remaining_time": "20:59:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1710, "total_steps": 3177, "loss": 0.1124, "learning_rate": 9.251246478698242e-06, "epoch": 1.6139688532326568, "percentage": 53.82, "elapsed_time": "1 day, 0:17:52", "remaining_time": "20:50:42", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1720, "total_steps": 3177, "loss": 0.1091, "learning_rate": 9.149606740115444e-06, "epoch": 1.6234072675790467, "percentage": 54.14, "elapsed_time": "1 day, 0:26:22", "remaining_time": "20:42:09", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1730, "total_steps": 3177, "loss": 0.1084, "learning_rate": 9.04805541784454e-06, "epoch": 1.6328456819254367, "percentage": 54.45, "elapsed_time": "1 day, 0:34:52", "remaining_time": "20:33:36", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1740, "total_steps": 3177, "loss": 0.1071, "learning_rate": 8.946603070286926e-06, "epoch": 1.6422840962718264, "percentage": 54.77, "elapsed_time": "1 day, 0:43:22", "remaining_time": "20:25:03", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1750, "total_steps": 3177, "loss": 0.1106, "learning_rate": 8.845260245553493e-06, "epoch": 1.651722510618216, "percentage": 55.08, "elapsed_time": "1 day, 0:51:52", "remaining_time": "20:16:31", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1760, "total_steps": 3177, "loss": 0.1095, "learning_rate": 8.744037480367922e-06, "epoch": 1.661160924964606, "percentage": 55.4, "elapsed_time": "1 day, 1:00:23", "remaining_time": "20:07:58", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1770, "total_steps": 3177, "loss": 0.1086, "learning_rate": 8.642945298971168e-06, "epoch": 1.670599339310996, "percentage": 55.71, "elapsed_time": "1 day, 1:08:53", "remaining_time": "19:59:26", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1780, "total_steps": 3177, "loss": 0.1096, "learning_rate": 8.54199421202726e-06, "epoch": 1.6800377536573856, "percentage": 56.03, "elapsed_time": "1 day, 1:17:23", "remaining_time": "19:50:53", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1790, "total_steps": 3177, "loss": 0.111, "learning_rate": 8.441194715530472e-06, "epoch": 1.6894761680037753, "percentage": 56.34, "elapsed_time": "1 day, 1:25:53", "remaining_time": "19:42:20", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1810, "total_steps": 3177, "loss": 0.1077, "learning_rate": 8.240092397960601e-06, "epoch": 1.708352996696555, "percentage": 56.97, "elapsed_time": "1 day, 1:42:52", "remaining_time": "19:25:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1820, "total_steps": 3177, "loss": 0.109, "learning_rate": 8.139810485714142e-06, "epoch": 1.7177914110429446, "percentage": 57.29, "elapsed_time": "1 day, 1:51:22", "remaining_time": "19:16:42", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1830, "total_steps": 3177, "loss": 0.1103, "learning_rate": 8.03972197939414e-06, "epoch": 1.7272298253893346, "percentage": 57.6, "elapsed_time": "1 day, 1:59:52", "remaining_time": "19:08:10", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1840, "total_steps": 3177, "loss": 0.106, "learning_rate": 7.939837285311425e-06, "epoch": 1.7366682397357245, "percentage": 57.92, "elapsed_time": "1 day, 2:08:22", "remaining_time": "18:59:37", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1850, "total_steps": 3177, "loss": 0.1111, "learning_rate": 7.840166788586244e-06, "epoch": 1.7461066540821142, "percentage": 58.23, "elapsed_time": "1 day, 2:16:52", "remaining_time": "18:51:05", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1860, "total_steps": 3177, "loss": 0.1107, "learning_rate": 7.740720852068524e-06, "epoch": 1.7555450684285039, "percentage": 58.55, "elapsed_time": "1 day, 2:25:22", "remaining_time": "18:42:33", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1880, "total_steps": 3177, "loss": 0.1092, "learning_rate": 7.542543993241278e-06, "epoch": 1.7744218971212837, "percentage": 59.18, "elapsed_time": "1 day, 2:42:22", "remaining_time": "18:25:27", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1890, "total_steps": 3177, "loss": 0.1033, "learning_rate": 7.443833675595254e-06, "epoch": 1.7838603114676734, "percentage": 59.49, "elapsed_time": "1 day, 2:50:52", "remaining_time": "18:16:55", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1900, "total_steps": 3177, "loss": 0.1088, "learning_rate": 7.3453891253413935e-06, "epoch": 1.7932987258140631, "percentage": 59.8, "elapsed_time": "1 day, 2:59:22", "remaining_time": "18:08:23", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1910, "total_steps": 3177, "loss": 0.1074, "learning_rate": 7.247220577866625e-06, "epoch": 1.802737140160453, "percentage": 60.12, "elapsed_time": "1 day, 3:07:51", "remaining_time": "17:59:50", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1920, "total_steps": 3177, "loss": 0.1069, "learning_rate": 7.149338239861579e-06, "epoch": 1.812175554506843, "percentage": 60.43, "elapsed_time": "1 day, 3:16:21", "remaining_time": "17:51:18", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1930, "total_steps": 3177, "loss": 0.1051, "learning_rate": 7.051752288259366e-06, "epoch": 1.8216139688532327, "percentage": 60.75, "elapsed_time": "1 day, 3:24:51", "remaining_time": "17:42:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1940, "total_steps": 3177, "loss": 0.1071, "learning_rate": 6.954472869177479e-06, "epoch": 1.8310523831996224, "percentage": 61.06, "elapsed_time": "1 day, 3:33:20", "remaining_time": "17:34:13", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1950, "total_steps": 3177, "loss": 0.108, "learning_rate": 6.857510096862901e-06, "epoch": 1.8404907975460123, "percentage": 61.38, "elapsed_time": "1 day, 3:41:50", "remaining_time": "17:25:41", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1960, "total_steps": 3177, "loss": 0.1081, "learning_rate": 6.760874052640494e-06, "epoch": 1.8499292118924022, "percentage": 61.69, "elapsed_time": "1 day, 3:50:20", "remaining_time": "17:17:08", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1970, "total_steps": 3177, "loss": 0.1079, "learning_rate": 6.664574783864862e-06, "epoch": 1.859367626238792, "percentage": 62.01, "elapsed_time": "1 day, 3:58:50", "remaining_time": "17:08:36", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1980, "total_steps": 3177, "loss": 0.1068, "learning_rate": 6.568622302875682e-06, "epoch": 1.8688060405851816, "percentage": 62.32, "elapsed_time": "1 day, 4:07:20", "remaining_time": "17:00:04", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 1990, "total_steps": 3177, "loss": 0.106, "learning_rate": 6.473026585956736e-06, "epoch": 1.8782444549315715, "percentage": 62.64, "elapsed_time": "1 day, 4:15:50", "remaining_time": "16:51:32", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2000, "total_steps": 3177, "loss": 0.1076, "learning_rate": 6.377797572298661e-06, "epoch": 1.8876828692779613, "percentage": 62.95, "elapsed_time": "1 day, 4:24:20", "remaining_time": "16:43:00", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2010, "total_steps": 3177, "loss": 0.1079, "learning_rate": 6.282945162965548e-06, "epoch": 1.897121283624351, "percentage": 63.27, "elapsed_time": "1 day, 4:32:49", "remaining_time": "16:34:27", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2020, "total_steps": 3177, "loss": 0.1064, "learning_rate": 6.188479219865529e-06, "epoch": 1.9065596979707409, "percentage": 63.58, "elapsed_time": "1 day, 4:41:20", "remaining_time": "16:25:56", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2030, "total_steps": 3177, "loss": 0.1054, "learning_rate": 6.094409564725435e-06, "epoch": 1.9159981123171308, "percentage": 63.9, "elapsed_time": "1 day, 4:49:50", "remaining_time": "16:17:24", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2040, "total_steps": 3177, "loss": 0.1082, "learning_rate": 6.0007459780695885e-06, "epoch": 1.9254365266635205, "percentage": 64.21, "elapsed_time": "1 day, 4:58:21", "remaining_time": "16:08:52", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2050, "total_steps": 3177, "loss": 0.1081, "learning_rate": 5.907498198202939e-06, "epoch": 1.9348749410099102, "percentage": 64.53, "elapsed_time": "1 day, 5:06:50", "remaining_time": "16:00:20", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2060, "total_steps": 3177, "loss": 0.1069, "learning_rate": 5.8146759201985525e-06, "epoch": 1.9443133553563001, "percentage": 64.84, "elapsed_time": "1 day, 5:15:21", "remaining_time": "15:51:48", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2070, "total_steps": 3177, "loss": 0.1064, "learning_rate": 5.722288794889603e-06, "epoch": 1.95375176970269, "percentage": 65.16, "elapsed_time": "1 day, 5:23:51", "remaining_time": "15:43:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2080, "total_steps": 3177, "loss": 0.1045, "learning_rate": 5.630346427865965e-06, "epoch": 1.9631901840490797, "percentage": 65.47, "elapsed_time": "1 day, 5:32:21", "remaining_time": "15:34:44", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2090, "total_steps": 3177, "loss": 0.1066, "learning_rate": 5.538858378475508e-06, "epoch": 1.9726285983954694, "percentage": 65.79, "elapsed_time": "1 day, 5:40:51", "remaining_time": "15:26:12", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2100, "total_steps": 3177, "loss": 0.1037, "learning_rate": 5.447834158830202e-06, "epoch": 1.9820670127418594, "percentage": 66.1, "elapsed_time": "1 day, 5:49:20", "remaining_time": "15:17:40", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2110, "total_steps": 3177, "loss": 0.1054, "learning_rate": 5.357283232817147e-06, "epoch": 1.9915054270882493, "percentage": 66.41, "elapsed_time": "1 day, 5:57:50", "remaining_time": "15:09:08", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2120, "total_steps": 3177, "loss": 0.1031, "learning_rate": 5.267215015114574e-06, "epoch": 2.0009438414346388, "percentage": 66.73, "elapsed_time": "1 day, 6:06:26", "remaining_time": "15:00:39", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2130, "total_steps": 3177, "loss": 0.0868, "learning_rate": 5.177638870213008e-06, "epoch": 2.0103822557810287, "percentage": 67.04, "elapsed_time": "1 day, 6:14:56", "remaining_time": "14:52:08", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2140, "total_steps": 3177, "loss": 0.0834, "learning_rate": 5.088564111441645e-06, "epoch": 2.0198206701274186, "percentage": 67.36, "elapsed_time": "1 day, 6:23:27", "remaining_time": "14:43:36", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2150, "total_steps": 3177, "loss": 0.0852, "learning_rate": 5.000000000000003e-06, "epoch": 2.0292590844738085, "percentage": 67.67, "elapsed_time": "1 day, 6:31:57", "remaining_time": "14:35:04", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2160, "total_steps": 3177, "loss": 0.0845, "learning_rate": 4.911955743995042e-06, "epoch": 2.038697498820198, "percentage": 67.99, "elapsed_time": "1 day, 6:40:26", "remaining_time": "14:26:32", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2170, "total_steps": 3177, "loss": 0.0847, "learning_rate": 4.824440497483802e-06, "epoch": 2.048135913166588, "percentage": 68.3, "elapsed_time": "1 day, 6:48:56", "remaining_time": "14:18:00", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2180, "total_steps": 3177, "loss": 0.0845, "learning_rate": 4.737463359521618e-06, "epoch": 2.057574327512978, "percentage": 68.62, "elapsed_time": "1 day, 6:57:26", "remaining_time": "14:09:28", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2190, "total_steps": 3177, "loss": 0.085, "learning_rate": 4.6510333732160915e-06, "epoch": 2.067012741859368, "percentage": 68.93, "elapsed_time": "1 day, 7:05:56", "remaining_time": "14:00:57", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2200, "total_steps": 3177, "loss": 0.0867, "learning_rate": 4.565159524786888e-06, "epoch": 2.0764511562057573, "percentage": 69.25, "elapsed_time": "1 day, 7:14:26", "remaining_time": "13:52:25", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2210, "total_steps": 3177, "loss": 0.0834, "learning_rate": 4.479850742631396e-06, "epoch": 2.085889570552147, "percentage": 69.56, "elapsed_time": "1 day, 7:22:56", "remaining_time": "13:43:53", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2220, "total_steps": 3177, "loss": 0.0849, "learning_rate": 4.395115896396457e-06, "epoch": 2.095327984898537, "percentage": 69.88, "elapsed_time": "1 day, 7:31:26", "remaining_time": "13:35:22", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2230, "total_steps": 3177, "loss": 0.084, "learning_rate": 4.310963796056168e-06, "epoch": 2.104766399244927, "percentage": 70.19, "elapsed_time": "1 day, 7:39:56", "remaining_time": "13:26:50", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2240, "total_steps": 3177, "loss": 0.0875, "learning_rate": 4.227403190995901e-06, "epoch": 2.1142048135913165, "percentage": 70.51, "elapsed_time": "1 day, 7:48:26", "remaining_time": "13:18:18", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2250, "total_steps": 3177, "loss": 0.0853, "learning_rate": 4.14444276910263e-06, "epoch": 2.1236432279377064, "percentage": 70.82, "elapsed_time": "1 day, 7:56:57", "remaining_time": "13:09:47", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2260, "total_steps": 3177, "loss": 0.0867, "learning_rate": 4.06209115586162e-06, "epoch": 2.1330816422840964, "percentage": 71.14, "elapsed_time": "1 day, 8:05:27", "remaining_time": "13:01:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2270, "total_steps": 3177, "loss": 0.0865, "learning_rate": 3.980356913459642e-06, "epoch": 2.1425200566304863, "percentage": 71.45, "elapsed_time": "1 day, 8:13:58", "remaining_time": "12:52:44", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2280, "total_steps": 3177, "loss": 0.0848, "learning_rate": 3.899248539894756e-06, "epoch": 2.1519584709768758, "percentage": 71.77, "elapsed_time": "1 day, 8:22:28", "remaining_time": "12:44:12", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2290, "total_steps": 3177, "loss": 0.0843, "learning_rate": 3.818774468092754e-06, "epoch": 2.1613968853232657, "percentage": 72.08, "elapsed_time": "1 day, 8:30:58", "remaining_time": "12:35:40", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2300, "total_steps": 3177, "loss": 0.0811, "learning_rate": 3.738943065030376e-06, "epoch": 2.1708352996696556, "percentage": 72.4, "elapsed_time": "1 day, 8:39:28", "remaining_time": "12:27:09", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2310, "total_steps": 3177, "loss": 0.083, "learning_rate": 3.659762630865411e-06, "epoch": 2.180273714016045, "percentage": 72.71, "elapsed_time": "1 day, 8:47:57", "remaining_time": "12:18:37", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2320, "total_steps": 3177, "loss": 0.0827, "learning_rate": 3.5812413980736916e-06, "epoch": 2.189712128362435, "percentage": 73.02, "elapsed_time": "1 day, 8:56:27", "remaining_time": "12:10:05", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2330, "total_steps": 3177, "loss": 0.0849, "learning_rate": 3.5033875305931662e-06, "epoch": 2.199150542708825, "percentage": 73.34, "elapsed_time": "1 day, 9:04:57", "remaining_time": "12:01:34", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2340, "total_steps": 3177, "loss": 0.0822, "learning_rate": 3.4262091229750973e-06, "epoch": 2.208588957055215, "percentage": 73.65, "elapsed_time": "1 day, 9:13:26", "remaining_time": "11:53:02", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2350, "total_steps": 3177, "loss": 0.0835, "learning_rate": 3.3497141995424397e-06, "epoch": 2.2180273714016043, "percentage": 73.97, "elapsed_time": "1 day, 9:21:56", "remaining_time": "11:44:30", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2360, "total_steps": 3177, "loss": 0.0841, "learning_rate": 3.2739107135555603e-06, "epoch": 2.2274657857479943, "percentage": 74.28, "elapsed_time": "1 day, 9:30:26", "remaining_time": "11:35:59", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2620, "total_steps": 3177, "loss": 0.0826, "learning_rate": 1.5699684619909983e-06, "epoch": 2.472864558754129, "percentage": 82.47, "elapsed_time": "1 day, 13:11:25", "remaining_time": "7:54:23", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2630, "total_steps": 3177, "loss": 0.0802, "learning_rate": 1.5155589093841939e-06, "epoch": 2.482302973100519, "percentage": 82.78, "elapsed_time": "1 day, 13:19:56", "remaining_time": "7:45:52", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2640, "total_steps": 3177, "loss": 0.081, "learning_rate": 1.4620314933430269e-06, "epoch": 2.491741387446909, "percentage": 83.1, "elapsed_time": "1 day, 13:28:26", "remaining_time": "7:37:21", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2650, "total_steps": 3177, "loss": 0.0808, "learning_rate": 1.4093917791711497e-06, "epoch": 2.501179801793299, "percentage": 83.41, "elapsed_time": "1 day, 13:36:55", "remaining_time": "7:28:49", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2660, "total_steps": 3177, "loss": 0.08, "learning_rate": 1.357645239876879e-06, "epoch": 2.5106182161396884, "percentage": 83.73, "elapsed_time": "1 day, 13:45:25", "remaining_time": "7:20:18", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2670, "total_steps": 3177, "loss": 0.0818, "learning_rate": 1.3067972556041753e-06, "epoch": 2.5200566304860783, "percentage": 84.04, "elapsed_time": "1 day, 13:53:55", "remaining_time": "7:11:47", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2680, "total_steps": 3177, "loss": 0.0807, "learning_rate": 1.2568531130732498e-06, "epoch": 2.5294950448324682, "percentage": 84.36, "elapsed_time": "1 day, 14:02:25", "remaining_time": "7:03:16", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2690, "total_steps": 3177, "loss": 0.0802, "learning_rate": 1.207818005030904e-06, "epoch": 2.5389334591788577, "percentage": 84.67, "elapsed_time": "1 day, 14:10:55", "remaining_time": "6:54:45", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2700, "total_steps": 3177, "loss": 0.0818, "learning_rate": 1.1596970297106458e-06, "epoch": 2.5483718735252476, "percentage": 84.99, "elapsed_time": "1 day, 14:19:24", "remaining_time": "6:46:13", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2710, "total_steps": 3177, "loss": 0.0806, "learning_rate": 1.1124951903025981e-06, "epoch": 2.5578102878716376, "percentage": 85.3, "elapsed_time": "1 day, 14:27:54", "remaining_time": "6:37:42", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2720, "total_steps": 3177, "loss": 0.081, "learning_rate": 1.0662173944333288e-06, "epoch": 2.5672487022180275, "percentage": 85.62, "elapsed_time": "1 day, 14:36:24", "remaining_time": "6:29:11", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2730, "total_steps": 3177, "loss": 0.081, "learning_rate": 1.0208684536555968e-06, "epoch": 2.5766871165644174, "percentage": 85.93, "elapsed_time": "1 day, 14:44:54", "remaining_time": "6:20:40", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2740, "total_steps": 3177, "loss": 0.0832, "learning_rate": 9.764530829480822e-07, "epoch": 2.586125530910807, "percentage": 86.24, "elapsed_time": "1 day, 14:53:24", "remaining_time": "6:12:09", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2750, "total_steps": 3177, "loss": 0.0802, "learning_rate": 9.329759002251726e-07, "epoch": 2.595563945257197, "percentage": 86.56, "elapsed_time": "1 day, 15:01:54", "remaining_time": "6:03:38", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2760, "total_steps": 3177, "loss": 0.0799, "learning_rate": 8.904414258568306e-07, "epoch": 2.6050023596035867, "percentage": 86.87, "elapsed_time": "1 day, 15:10:24", "remaining_time": "5:55:06", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2770, "total_steps": 3177, "loss": 0.0827, "learning_rate": 8.488540821986035e-07, "epoch": 2.614440773949976, "percentage": 87.19, "elapsed_time": "1 day, 15:18:53", "remaining_time": "5:46:35", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2780, "total_steps": 3177, "loss": 0.0792, "learning_rate": 8.082181931318311e-07, "epoch": 2.623879188296366, "percentage": 87.5, "elapsed_time": "1 day, 15:27:24", "remaining_time": "5:38:04", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2790, "total_steps": 3177, "loss": 0.079, "learning_rate": 7.685379836140872e-07, "epoch": 2.633317602642756, "percentage": 87.82, "elapsed_time": "1 day, 15:35:54", "remaining_time": "5:29:33", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2800, "total_steps": 3177, "loss": 0.0818, "learning_rate": 7.298175792398976e-07, "epoch": 2.642756016989146, "percentage": 88.13, "elapsed_time": "1 day, 15:44:24", "remaining_time": "5:21:02", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2810, "total_steps": 3177, "loss": 0.0804, "learning_rate": 6.920610058118105e-07, "epoch": 2.652194431335536, "percentage": 88.45, "elapsed_time": "1 day, 15:52:54", "remaining_time": "5:12:31", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2820, "total_steps": 3177, "loss": 0.0816, "learning_rate": 6.552721889218194e-07, "epoch": 2.6616328456819254, "percentage": 88.76, "elapsed_time": "1 day, 16:01:24", "remaining_time": "5:04:00", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2830, "total_steps": 3177, "loss": 0.08, "learning_rate": 6.194549535432137e-07, "epoch": 2.6710712600283153, "percentage": 89.08, "elapsed_time": "1 day, 16:09:54", "remaining_time": "4:55:29", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2840, "total_steps": 3177, "loss": 0.0823, "learning_rate": 5.846130236329073e-07, "epoch": 2.680509674374705, "percentage": 89.39, "elapsed_time": "1 day, 16:18:23", "remaining_time": "4:46:58", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2850, "total_steps": 3177, "loss": 0.0809, "learning_rate": 5.507500217442341e-07, "epoch": 2.6899480887210947, "percentage": 89.71, "elapsed_time": "1 day, 16:26:53", "remaining_time": "4:38:27", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2860, "total_steps": 3177, "loss": 0.0784, "learning_rate": 5.178694686503205e-07, "epoch": 2.6993865030674846, "percentage": 90.02, "elapsed_time": "1 day, 16:35:23", "remaining_time": "4:29:56", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2870, "total_steps": 3177, "loss": 0.081, "learning_rate": 4.85974782978027e-07, "epoch": 2.7088249174138745, "percentage": 90.34, "elapsed_time": "1 day, 16:43:53", "remaining_time": "4:21:25", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2880, "total_steps": 3177, "loss": 0.0778, "learning_rate": 4.5506928085250033e-07, "epoch": 2.718263331760264, "percentage": 90.65, "elapsed_time": "1 day, 16:52:22", "remaining_time": "4:12:54", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2890, "total_steps": 3177, "loss": 0.0832, "learning_rate": 4.251561755524036e-07, "epoch": 2.727701746106654, "percentage": 90.97, "elapsed_time": "1 day, 17:00:52", "remaining_time": "4:04:23", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2900, "total_steps": 3177, "loss": 0.0808, "learning_rate": 3.9623857717581813e-07, "epoch": 2.737140160453044, "percentage": 91.28, "elapsed_time": "1 day, 17:09:22", "remaining_time": "3:55:52", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2910, "total_steps": 3177, "loss": 0.081, "learning_rate": 3.6831949231689203e-07, "epoch": 2.746578574799434, "percentage": 91.6, "elapsed_time": "1 day, 17:17:51", "remaining_time": "3:47:21", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2920, "total_steps": 3177, "loss": 0.0821, "learning_rate": 3.414018237532335e-07, "epoch": 2.7560169891458237, "percentage": 91.91, "elapsed_time": "1 day, 17:26:21", "remaining_time": "3:38:50", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2930, "total_steps": 3177, "loss": 0.08, "learning_rate": 3.154883701441136e-07, "epoch": 2.765455403492213, "percentage": 92.23, "elapsed_time": "1 day, 17:34:51", "remaining_time": "3:30:19", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2940, "total_steps": 3177, "loss": 0.0798, "learning_rate": 2.905818257394799e-07, "epoch": 2.774893817838603, "percentage": 92.54, "elapsed_time": "1 day, 17:43:21", "remaining_time": "3:21:48", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2950, "total_steps": 3177, "loss": 0.0819, "learning_rate": 2.666847800998362e-07, "epoch": 2.784332232184993, "percentage": 92.85, "elapsed_time": "1 day, 17:51:51", "remaining_time": "3:13:17", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2960, "total_steps": 3177, "loss": 0.0807, "learning_rate": 2.437997178270035e-07, "epoch": 2.7937706465313825, "percentage": 93.17, "elapsed_time": "1 day, 18:00:21", "remaining_time": "3:04:46", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2970, "total_steps": 3177, "loss": 0.0806, "learning_rate": 2.219290183057865e-07, "epoch": 2.8032090608777724, "percentage": 93.48, "elapsed_time": "1 day, 18:08:52", "remaining_time": "2:56:15", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2980, "total_steps": 3177, "loss": 0.0778, "learning_rate": 2.0107495545659829e-07, "epoch": 2.8126474752241624, "percentage": 93.8, "elapsed_time": "1 day, 18:17:22", "remaining_time": "2:47:44", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 2990, "total_steps": 3177, "loss": 0.0777, "learning_rate": 1.8123969749902714e-07, "epoch": 2.8220858895705523, "percentage": 94.11, "elapsed_time": "1 day, 18:25:52", "remaining_time": "2:39:13", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3000, "total_steps": 3177, "loss": 0.0813, "learning_rate": 1.6242530672641143e-07, "epoch": 2.831524303916942, "percentage": 94.43, "elapsed_time": "1 day, 18:34:22", "remaining_time": "2:30:42", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3010, "total_steps": 3177, "loss": 0.0788, "learning_rate": 1.4463373929141766e-07, "epoch": 2.8409627182633317, "percentage": 94.74, "elapsed_time": "1 day, 18:42:51", "remaining_time": "2:22:11", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3020, "total_steps": 3177, "loss": 0.0819, "learning_rate": 1.2786684500265546e-07, "epoch": 2.8504011326097216, "percentage": 95.06, "elapsed_time": "1 day, 18:51:21", "remaining_time": "2:13:40", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3030, "total_steps": 3177, "loss": 0.0794, "learning_rate": 1.1212636713235581e-07, "epoch": 2.8598395469561115, "percentage": 95.37, "elapsed_time": "1 day, 18:59:50", "remaining_time": "2:05:09", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3040, "total_steps": 3177, "loss": 0.0814, "learning_rate": 9.741394223512057e-08, "epoch": 2.869277961302501, "percentage": 95.69, "elapsed_time": "1 day, 19:08:20", "remaining_time": "1:56:38", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3050, "total_steps": 3177, "loss": 0.0804, "learning_rate": 8.373109997776185e-08, "epoch": 2.878716375648891, "percentage": 96.0, "elapsed_time": "1 day, 19:16:51", "remaining_time": "1:48:07", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3060, "total_steps": 3177, "loss": 0.08, "learning_rate": 7.10792629802659e-08, "epoch": 2.888154789995281, "percentage": 96.32, "elapsed_time": "1 day, 19:25:21", "remaining_time": "1:39:37", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3070, "total_steps": 3177, "loss": 0.0809, "learning_rate": 5.945974666788479e-08, "epoch": 2.8975932043416703, "percentage": 96.63, "elapsed_time": "1 day, 19:33:52", "remaining_time": "1:31:06", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3080, "total_steps": 3177, "loss": 0.0817, "learning_rate": 4.887375913436132e-08, "epoch": 2.9070316186880603, "percentage": 96.95, "elapsed_time": "1 day, 19:42:22", "remaining_time": "1:22:35", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3090, "total_steps": 3177, "loss": 0.0823, "learning_rate": 3.932240101633178e-08, "epoch": 2.91647003303445, "percentage": 97.26, "elapsed_time": "1 day, 19:50:51", "remaining_time": "1:14:04", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3100, "total_steps": 3177, "loss": 0.0764, "learning_rate": 3.0806665378884106e-08, "epoch": 2.92590844738084, "percentage": 97.58, "elapsed_time": "1 day, 19:59:21", "remaining_time": "1:05:33", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3110, "total_steps": 3177, "loss": 0.0784, "learning_rate": 2.33274376123116e-08, "epoch": 2.93534686172723, "percentage": 97.89, "elapsed_time": "1 day, 20:07:50", "remaining_time": "0:57:02", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3120, "total_steps": 3177, "loss": 0.0795, "learning_rate": 1.68854953400599e-08, "epoch": 2.9447852760736195, "percentage": 98.21, "elapsed_time": "1 day, 20:16:20", "remaining_time": "0:48:31", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3130, "total_steps": 3177, "loss": 0.0801, "learning_rate": 1.1481508337869429e-08, "epoch": 2.9542236904200094, "percentage": 98.52, "elapsed_time": "1 day, 20:24:50", "remaining_time": "0:40:00", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3140, "total_steps": 3177, "loss": 0.0828, "learning_rate": 7.1160384641455475e-09, "epoch": 2.9636621047663994, "percentage": 98.84, "elapsed_time": "1 day, 20:33:20", "remaining_time": "0:31:30", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3150, "total_steps": 3177, "loss": 0.0808, "learning_rate": 3.7895396015374955e-09, "epoch": 2.973100519112789, "percentage": 99.15, "elapsed_time": "1 day, 20:41:50", "remaining_time": "0:22:59", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3160, "total_steps": 3177, "loss": 0.0803, "learning_rate": 1.502357609749483e-09, "epoch": 2.9825389334591788, "percentage": 99.46, "elapsed_time": "1 day, 20:50:20", "remaining_time": "0:14:28", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3170, "total_steps": 3177, "loss": 0.0813, "learning_rate": 2.5473028957945234e-10, "epoch": 2.9919773478055687, "percentage": 99.78, "elapsed_time": "1 day, 20:58:50", "remaining_time": "0:05:57", "throughput": "0.00", "total_tokens": 0}
+{"current_steps": 3177, "total_steps": 3177, "epoch": 2.9985842378480414, "percentage": 100.0, "elapsed_time": "1 day, 21:08:40", "remaining_time": "0:00:00", "throughput": "0.00", "total_tokens": 0}

trainer_state.json

@@ -0,0 +1,2261 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.9985842378480414,
+  "eval_steps": 500,
+  "global_step": 3177,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.009438414346389807,
+      "grad_norm": 0.7047261682571029,
+      "learning_rate": 2.0833333333333334e-06,
+      "loss": 0.3368,
+      "step": 10
+    },
+    {
+      "epoch": 0.018876828692779613,
+      "grad_norm": 0.44565794909772116,
+      "learning_rate": 4.166666666666667e-06,
+      "loss": 0.2696,
+      "step": 20
+    },
+    {
+      "epoch": 0.028315243039169418,
+      "grad_norm": 0.2469001773100438,
+      "learning_rate": 6.25e-06,
+      "loss": 0.2345,
+      "step": 30
+    },
+    {
+      "epoch": 0.037753657385559226,
+      "grad_norm": 0.17504888016598083,
+      "learning_rate": 8.333333333333334e-06,
+      "loss": 0.213,
+      "step": 40
+    },
+    {
+      "epoch": 0.04719207173194903,
+      "grad_norm": 0.14319985578848382,
+      "learning_rate": 1.0416666666666668e-05,
+      "loss": 0.194,
+      "step": 50
+    },
+    {
+      "epoch": 0.056630486078338836,
+      "grad_norm": 0.1594795618687463,
+      "learning_rate": 1.25e-05,
+      "loss": 0.1871,
+      "step": 60
+    },
+    {
+      "epoch": 0.06606890042472864,
+      "grad_norm": 0.14625288121586646,
+      "learning_rate": 1.4583333333333333e-05,
+      "loss": 0.1791,
+      "step": 70
+    },
+    {
+      "epoch": 0.07550731477111845,
+      "grad_norm": 0.44536813060583275,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 0.1777,
+      "step": 80
+    },
+    {
+      "epoch": 0.08494572911750826,
+      "grad_norm": 0.17858201185676476,
+      "learning_rate": 1.8750000000000002e-05,
+      "loss": 0.1743,
+      "step": 90
+    },
+    {
+      "epoch": 0.09438414346389806,
+      "grad_norm": 0.22679600482590775,
+      "learning_rate": 1.9999916822524766e-05,
+      "loss": 0.173,
+      "step": 100
+    },
+    {
+      "epoch": 0.10382255781028787,
+      "grad_norm": 0.2245846959065181,
+      "learning_rate": 1.999898109181919e-05,
+      "loss": 0.1697,
+      "step": 110
+    },
+    {
+      "epoch": 0.11326097215667767,
+      "grad_norm": 0.21274961909318216,
+      "learning_rate": 1.9997005756177228e-05,
+      "loss": 0.1669,
+      "step": 120
+    },
+    {
+      "epoch": 0.12269938650306748,
+      "grad_norm": 0.15535944242468744,
+      "learning_rate": 1.999399102097668e-05,
+      "loss": 0.1664,
+      "step": 130
+    },
+    {
+      "epoch": 0.13213780084945728,
+      "grad_norm": 0.1677351780597522,
+      "learning_rate": 1.9989937199662845e-05,
+      "loss": 0.1652,
+      "step": 140
+    },
+    {
+      "epoch": 0.1415762151958471,
+      "grad_norm": 0.15664753634679404,
+      "learning_rate": 1.998484471371593e-05,
+      "loss": 0.1619,
+      "step": 150
+    },
+    {
+      "epoch": 0.1510146295422369,
+      "grad_norm": 0.1936710831336268,
+      "learning_rate": 1.9978714092607234e-05,
+      "loss": 0.1606,
+      "step": 160
+    },
+    {
+      "epoch": 0.16045304388862672,
+      "grad_norm": 0.21595388620143963,
+      "learning_rate": 1.9971545973744102e-05,
+      "loss": 0.16,
+      "step": 170
+    },
+    {
+      "epoch": 0.16989145823501653,
+      "grad_norm": 0.19084712938417736,
+      "learning_rate": 1.9963341102403652e-05,
+      "loss": 0.1582,
+      "step": 180
+    },
+    {
+      "epoch": 0.1793298725814063,
+      "grad_norm": 0.19755743286804991,
+      "learning_rate": 1.9954100331655265e-05,
+      "loss": 0.1551,
+      "step": 190
+    },
+    {
+      "epoch": 0.18876828692779613,
+      "grad_norm": 0.21785901423412252,
+      "learning_rate": 1.9943824622271934e-05,
+      "loss": 0.1559,
+      "step": 200
+    },
+    {
+      "epoch": 0.19820670127418594,
+      "grad_norm": 0.1472275149396309,
+      "learning_rate": 1.9932515042630335e-05,
+      "loss": 0.1534,
+      "step": 210
+    },
+    {
+      "epoch": 0.20764511562057575,
+      "grad_norm": 0.14184865805893884,
+      "learning_rate": 1.9920172768599763e-05,
+      "loss": 0.1545,
+      "step": 220
+    },
+    {
+      "epoch": 0.21708352996696556,
+      "grad_norm": 0.1833363522186809,
+      "learning_rate": 1.9906799083419865e-05,
+      "loss": 0.1543,
+      "step": 230
+    },
+    {
+      "epoch": 0.22652194431335534,
+      "grad_norm": 0.141212536462394,
+      "learning_rate": 1.989239537756723e-05,
+      "loss": 0.1538,
+      "step": 240
+    },
+    {
+      "epoch": 0.23596035865974516,
+      "grad_norm": 0.1556507383209179,
+      "learning_rate": 1.987696314861082e-05,
+      "loss": 0.1541,
+      "step": 250
+    },
+    {
+      "epoch": 0.24539877300613497,
+      "grad_norm": 0.17173283637998002,
+      "learning_rate": 1.986050400105626e-05,
+      "loss": 0.1518,
+      "step": 260
+    },
+    {
+      "epoch": 0.25483718735252475,
+      "grad_norm": 0.150629748875022,
+      "learning_rate": 1.9843019646179014e-05,
+      "loss": 0.1501,
+      "step": 270
+    },
+    {
+      "epoch": 0.26427560169891456,
+      "grad_norm": 0.19334006281958221,
+      "learning_rate": 1.9824511901846475e-05,
+      "loss": 0.1483,
+      "step": 280
+    },
+    {
+      "epoch": 0.2737140160453044,
+      "grad_norm": 0.12527664802237196,
+      "learning_rate": 1.9804982692328944e-05,
+      "loss": 0.1514,
+      "step": 290
+    },
+    {
+      "epoch": 0.2831524303916942,
+      "grad_norm": 0.19858199254198736,
+      "learning_rate": 1.9784434048099565e-05,
+      "loss": 0.151,
+      "step": 300
+    },
+    {
+      "epoch": 0.292590844738084,
+      "grad_norm": 0.1677799134496006,
+      "learning_rate": 1.976286810562323e-05,
+      "loss": 0.1498,
+      "step": 310
+    },
+    {
+      "epoch": 0.3020292590844738,
+      "grad_norm": 0.15743349400071904,
+      "learning_rate": 1.9740287107134417e-05,
+      "loss": 0.1513,
+      "step": 320
+    },
+    {
+      "epoch": 0.3114676734308636,
+      "grad_norm": 0.1616061413079364,
+      "learning_rate": 1.97166934004041e-05,
+      "loss": 0.1489,
+      "step": 330
+    },
+    {
+      "epoch": 0.32090608777725343,
+      "grad_norm": 0.1878283137943562,
+      "learning_rate": 1.9692089438495622e-05,
+      "loss": 0.1449,
+      "step": 340
+    },
+    {
+      "epoch": 0.33034450212364325,
+      "grad_norm": 0.16166637953640253,
+      "learning_rate": 1.9666477779509655e-05,
+      "loss": 0.1469,
+      "step": 350
+    },
+    {
+      "epoch": 0.33978291647003306,
+      "grad_norm": 0.12292532976422958,
+      "learning_rate": 1.963986108631823e-05,
+      "loss": 0.1468,
+      "step": 360
+    },
+    {
+      "epoch": 0.3492213308164228,
+      "grad_norm": 0.16469874208354635,
+      "learning_rate": 1.9612242126287876e-05,
+      "loss": 0.1483,
+      "step": 370
+    },
+    {
+      "epoch": 0.3586597451628126,
+      "grad_norm": 0.13369950841636608,
+      "learning_rate": 1.958362377099191e-05,
+      "loss": 0.1443,
+      "step": 380
+    },
+    {
+      "epoch": 0.36809815950920244,
+      "grad_norm": 0.12551292002845085,
+      "learning_rate": 1.9554008995911837e-05,
+      "loss": 0.1463,
+      "step": 390
+    },
+    {
+      "epoch": 0.37753657385559225,
+      "grad_norm": 0.14851921568961518,
+      "learning_rate": 1.9523400880128032e-05,
+      "loss": 0.1471,
+      "step": 400
+    },
+    {
+      "epoch": 0.38697498820198206,
+      "grad_norm": 0.13162403368451694,
+      "learning_rate": 1.949180260599957e-05,
+      "loss": 0.1452,
+      "step": 410
+    },
+    {
+      "epoch": 0.3964134025483719,
+      "grad_norm": 0.12724887401811377,
+      "learning_rate": 1.945921745883337e-05,
+      "loss": 0.1455,
+      "step": 420
+    },
+    {
+      "epoch": 0.4058518168947617,
+      "grad_norm": 0.11779906396951238,
+      "learning_rate": 1.9425648826542618e-05,
+      "loss": 0.1435,
+      "step": 430
+    },
+    {
+      "epoch": 0.4152902312411515,
+      "grad_norm": 0.1610933457769652,
+      "learning_rate": 1.939110019929451e-05,
+      "loss": 0.1436,
+      "step": 440
+    },
+    {
+      "epoch": 0.4247286455875413,
+      "grad_norm": 0.12250958203512534,
+      "learning_rate": 1.935557516914739e-05,
+      "loss": 0.1451,
+      "step": 450
+    },
+    {
+      "epoch": 0.4341670599339311,
+      "grad_norm": 0.1380572524992858,
+      "learning_rate": 1.931907742967727e-05,
+      "loss": 0.1444,
+      "step": 460
+    },
+    {
+      "epoch": 0.44360547428032093,
+      "grad_norm": 0.13646993698111895,
+      "learning_rate": 1.92816107755938e-05,
+      "loss": 0.142,
+      "step": 470
+    },
+    {
+      "epoch": 0.4530438886267107,
+      "grad_norm": 0.11765542306036501,
+      "learning_rate": 1.9243179102345753e-05,
+      "loss": 0.1406,
+      "step": 480
+    },
+    {
+      "epoch": 0.4624823029731005,
+      "grad_norm": 0.1266567901893174,
+      "learning_rate": 1.9203786405715984e-05,
+      "loss": 0.144,
+      "step": 490
+    },
+    {
+      "epoch": 0.4719207173194903,
+      "grad_norm": 0.1113634311573256,
+      "learning_rate": 1.9163436781405992e-05,
+      "loss": 0.1428,
+      "step": 500
+    },
+    {
+      "epoch": 0.4813591316658801,
+      "grad_norm": 0.13808836428511967,
+      "learning_rate": 1.912213442461009e-05,
+      "loss": 0.1399,
+      "step": 510
+    },
+    {
+      "epoch": 0.49079754601226994,
+      "grad_norm": 0.1226613837593307,
+      "learning_rate": 1.9079883629579224e-05,
+      "loss": 0.1396,
+      "step": 520
+    },
+    {
+      "epoch": 0.5002359603586597,
+      "grad_norm": 0.14272835200919645,
+      "learning_rate": 1.9036688789174496e-05,
+      "loss": 0.1403,
+      "step": 530
+    },
+    {
+      "epoch": 0.5096743747050495,
+      "grad_norm": 0.12981510040553715,
+      "learning_rate": 1.899255439441043e-05,
+      "loss": 0.1399,
+      "step": 540
+    },
+    {
+      "epoch": 0.5191127890514393,
+      "grad_norm": 0.1190871345092575,
+      "learning_rate": 1.8947485033988034e-05,
+      "loss": 0.1376,
+      "step": 550
+    },
+    {
+      "epoch": 0.5285512033978291,
+      "grad_norm": 0.1271477738963388,
+      "learning_rate": 1.8901485393817724e-05,
+      "loss": 0.1415,
+      "step": 560
+    },
+    {
+      "epoch": 0.5379896177442189,
+      "grad_norm": 0.12965211048846748,
+      "learning_rate": 1.8854560256532098e-05,
+      "loss": 0.1423,
+      "step": 570
+    },
+    {
+      "epoch": 0.5474280320906088,
+      "grad_norm": 0.13373262160455968,
+      "learning_rate": 1.880671450098871e-05,
+      "loss": 0.139,
+      "step": 580
+    },
+    {
+      "epoch": 0.5568664464369986,
+      "grad_norm": 0.1322939697550499,
+      "learning_rate": 1.8757953101762786e-05,
+      "loss": 0.1396,
+      "step": 590
+    },
+    {
+      "epoch": 0.5663048607833884,
+      "grad_norm": 0.11918437239832326,
+      "learning_rate": 1.8708281128630023e-05,
+      "loss": 0.138,
+      "step": 600
+    },
+    {
+      "epoch": 0.5757432751297782,
+      "grad_norm": 0.12338738357381479,
+      "learning_rate": 1.865770374603948e-05,
+      "loss": 0.1406,
+      "step": 610
+    },
+    {
+      "epoch": 0.585181689476168,
+      "grad_norm": 0.11573754594906395,
+      "learning_rate": 1.8606226212576612e-05,
+      "loss": 0.138,
+      "step": 620
+    },
+    {
+      "epoch": 0.5946201038225578,
+      "grad_norm": 0.1419588706141848,
+      "learning_rate": 1.8553853880416555e-05,
+      "loss": 0.1408,
+      "step": 630
+    },
+    {
+      "epoch": 0.6040585181689476,
+      "grad_norm": 0.13998266637185536,
+      "learning_rate": 1.8500592194767625e-05,
+      "loss": 0.1394,
+      "step": 640
+    },
+    {
+      "epoch": 0.6134969325153374,
+      "grad_norm": 0.11868995175822014,
+      "learning_rate": 1.8446446693305194e-05,
+      "loss": 0.1384,
+      "step": 650
+    },
+    {
+      "epoch": 0.6229353468617272,
+      "grad_norm": 0.1328472026088287,
+      "learning_rate": 1.8391423005595928e-05,
+      "loss": 0.1393,
+      "step": 660
+    },
+    {
+      "epoch": 0.6323737612081171,
+      "grad_norm": 0.11726921800593894,
+      "learning_rate": 1.833552685251246e-05,
+      "loss": 0.1398,
+      "step": 670
+    },
+    {
+      "epoch": 0.6418121755545069,
+      "grad_norm": 0.11466260187649016,
+      "learning_rate": 1.827876404563861e-05,
+      "loss": 0.1369,
+      "step": 680
+    },
+    {
+      "epoch": 0.6512505899008967,
+      "grad_norm": 0.11234281014514101,
+      "learning_rate": 1.8221140486665125e-05,
+      "loss": 0.1346,
+      "step": 690
+    },
+    {
+      "epoch": 0.6606890042472865,
+      "grad_norm": 0.11159741277810285,
+      "learning_rate": 1.8162662166776085e-05,
+      "loss": 0.1357,
+      "step": 700
+    },
+    {
+      "epoch": 0.6701274185936763,
+      "grad_norm": 0.12752868267859116,
+      "learning_rate": 1.8103335166026002e-05,
+      "loss": 0.1389,
+      "step": 710
+    },
+    {
+      "epoch": 0.6795658329400661,
+      "grad_norm": 0.12084535348559353,
+      "learning_rate": 1.804316565270765e-05,
+      "loss": 0.1375,
+      "step": 720
+    },
+    {
+      "epoch": 0.6890042472864559,
+      "grad_norm": 0.12102077085461252,
+      "learning_rate": 1.798215988271075e-05,
+      "loss": 0.1364,
+      "step": 730
+    },
+    {
+      "epoch": 0.6984426616328456,
+      "grad_norm": 0.11713742692774234,
+      "learning_rate": 1.7920324198871546e-05,
+      "loss": 0.138,
+      "step": 740
+    },
+    {
+      "epoch": 0.7078810759792354,
+      "grad_norm": 0.11656355822805255,
+      "learning_rate": 1.785766503031332e-05,
+      "loss": 0.1346,
+      "step": 750
+    },
+    {
+      "epoch": 0.7173194903256253,
+      "grad_norm": 0.11377328844943654,
+      "learning_rate": 1.7794188891777964e-05,
+      "loss": 0.1352,
+      "step": 760
+    },
+    {
+      "epoch": 0.7267579046720151,
+      "grad_norm": 0.12103799646507679,
+      "learning_rate": 1.7729902382948617e-05,
+      "loss": 0.1353,
+      "step": 770
+    },
+    {
+      "epoch": 0.7361963190184049,
+      "grad_norm": 0.1073585292390918,
+      "learning_rate": 1.76648121877635e-05,
+      "loss": 0.1352,
+      "step": 780
+    },
+    {
+      "epoch": 0.7456347333647947,
+      "grad_norm": 0.11214075940260533,
+      "learning_rate": 1.759892507372099e-05,
+      "loss": 0.1341,
+      "step": 790
+    },
+    {
+      "epoch": 0.7550731477111845,
+      "grad_norm": 0.11706899066793994,
+      "learning_rate": 1.7532247891175968e-05,
+      "loss": 0.1333,
+      "step": 800
+    },
+    {
+      "epoch": 0.7645115620575743,
+      "grad_norm": 0.11789888505768062,
+      "learning_rate": 1.746478757262761e-05,
+      "loss": 0.136,
+      "step": 810
+    },
+    {
+      "epoch": 0.7739499764039641,
+      "grad_norm": 0.11237848535926774,
+      "learning_rate": 1.739655113199858e-05,
+      "loss": 0.1336,
+      "step": 820
+    },
+    {
+      "epoch": 0.7833883907503539,
+      "grad_norm": 0.10753154987431834,
+      "learning_rate": 1.7327545663905813e-05,
+      "loss": 0.1331,
+      "step": 830
+    },
+    {
+      "epoch": 0.7928268050967437,
+      "grad_norm": 0.1441522552747225,
+      "learning_rate": 1.7257778342922853e-05,
+      "loss": 0.1328,
+      "step": 840
+    },
+    {
+      "epoch": 0.8022652194431336,
+      "grad_norm": 0.1269707942863234,
+      "learning_rate": 1.7187256422833928e-05,
+      "loss": 0.1319,
+      "step": 850
+    },
+    {
+      "epoch": 0.8117036337895234,
+      "grad_norm": 0.11091236494221275,
+      "learning_rate": 1.711598723587975e-05,
+      "loss": 0.1324,
+      "step": 860
+    },
+    {
+      "epoch": 0.8211420481359132,
+      "grad_norm": 0.10854265306012167,
+      "learning_rate": 1.7043978191995177e-05,
+      "loss": 0.1325,
+      "step": 870
+    },
+    {
+      "epoch": 0.830580462482303,
+      "grad_norm": 0.1110467712060928,
+      "learning_rate": 1.6971236778038806e-05,
+      "loss": 0.1315,
+      "step": 880
+    },
+    {
+      "epoch": 0.8400188768286928,
+      "grad_norm": 0.12129611756408008,
+      "learning_rate": 1.6897770557014535e-05,
+      "loss": 0.1328,
+      "step": 890
+    },
+    {
+      "epoch": 0.8494572911750826,
+      "grad_norm": 0.106781748696916,
+      "learning_rate": 1.682358716728525e-05,
+      "loss": 0.1351,
+      "step": 900
+    },
+    {
+      "epoch": 0.8588957055214724,
+      "grad_norm": 0.11020118439519076,
+      "learning_rate": 1.674869432177864e-05,
+      "loss": 0.1325,
+      "step": 910
+    },
+    {
+      "epoch": 0.8683341198678622,
+      "grad_norm": 0.11750342557908768,
+      "learning_rate": 1.667309980718529e-05,
+      "loss": 0.1312,
+      "step": 920
+    },
+    {
+      "epoch": 0.877772534214252,
+      "grad_norm": 0.12853148875033116,
+      "learning_rate": 1.6596811483149077e-05,
+      "loss": 0.1317,
+      "step": 930
+    },
+    {
+      "epoch": 0.8872109485606419,
+      "grad_norm": 0.11393786746070304,
+      "learning_rate": 1.651983728145e-05,
+      "loss": 0.1355,
+      "step": 940
+    },
+    {
+      "epoch": 0.8966493629070316,
+      "grad_norm": 0.11013100846033319,
+      "learning_rate": 1.6442185205179507e-05,
+      "loss": 0.1309,
+      "step": 950
+    },
+    {
+      "epoch": 0.9060877772534214,
+      "grad_norm": 0.10641286141618259,
+      "learning_rate": 1.6363863327908405e-05,
+      "loss": 0.1339,
+      "step": 960
+    },
+    {
+      "epoch": 0.9155261915998112,
+      "grad_norm": 0.11308702339858176,
+      "learning_rate": 1.6284879792847433e-05,
+      "loss": 0.1299,
+      "step": 970
+    },
+    {
+      "epoch": 0.924964605946201,
+      "grad_norm": 0.11315250527471539,
+      "learning_rate": 1.620524281200062e-05,
+      "loss": 0.1305,
+      "step": 980
+    },
+    {
+      "epoch": 0.9344030202925908,
+      "grad_norm": 0.0972875949252274,
+      "learning_rate": 1.6124960665311447e-05,
+      "loss": 0.1322,
+      "step": 990
+    },
+    {
+      "epoch": 0.9438414346389806,
+      "grad_norm": 0.10070713866086979,
+      "learning_rate": 1.6044041699802005e-05,
+      "loss": 0.129,
+      "step": 1000
+    },
+    {
+      "epoch": 0.9532798489853704,
+      "grad_norm": 0.11109074990403733,
+      "learning_rate": 1.5962494328705123e-05,
+      "loss": 0.1321,
+      "step": 1010
+    },
+    {
+      "epoch": 0.9627182633317602,
+      "grad_norm": 0.1199186598391774,
+      "learning_rate": 1.588032703058964e-05,
+      "loss": 0.1334,
+      "step": 1020
+    },
+    {
+      "epoch": 0.9721566776781501,
+      "grad_norm": 0.10777396066893469,
+      "learning_rate": 1.5797548348478893e-05,
+      "loss": 0.1325,
+      "step": 1030
+    },
+    {
+      "epoch": 0.9815950920245399,
+      "grad_norm": 0.11999882098060052,
+      "learning_rate": 1.571416688896246e-05,
+      "loss": 0.132,
+      "step": 1040
+    },
+    {
+      "epoch": 0.9910335063709297,
+      "grad_norm": 0.10911342083469809,
+      "learning_rate": 1.563019132130136e-05,
+      "loss": 0.1301,
+      "step": 1050
+    },
+    {
+      "epoch": 1.0004719207173194,
+      "grad_norm": 0.11482143235010223,
+      "learning_rate": 1.5545630376526665e-05,
+      "loss": 0.1282,
+      "step": 1060
+    },
+    {
+      "epoch": 1.0099103350637093,
+      "grad_norm": 0.11699392564682201,
+      "learning_rate": 1.5460492846531748e-05,
+      "loss": 0.1142,
+      "step": 1070
+    },
+    {
+      "epoch": 1.019348749410099,
+      "grad_norm": 0.09530170230868218,
+      "learning_rate": 1.5374787583158188e-05,
+      "loss": 0.1157,
+      "step": 1080
+    },
+    {
+      "epoch": 1.028787163756489,
+      "grad_norm": 0.09486822390799159,
+      "learning_rate": 1.5288523497275392e-05,
+      "loss": 0.1143,
+      "step": 1090
+    },
+    {
+      "epoch": 1.0382255781028786,
+      "grad_norm": 0.09531194088354993,
+      "learning_rate": 1.5201709557854178e-05,
+      "loss": 0.1128,
+      "step": 1100
+    },
+    {
+      "epoch": 1.0476639924492686,
+      "grad_norm": 0.11206649112299381,
+      "learning_rate": 1.5114354791034225e-05,
+      "loss": 0.1161,
+      "step": 1110
+    },
+    {
+      "epoch": 1.0571024067956583,
+      "grad_norm": 0.10196697892456508,
+      "learning_rate": 1.5026468279185615e-05,
+      "loss": 0.1159,
+      "step": 1120
+    },
+    {
+      "epoch": 1.0665408211420482,
+      "grad_norm": 0.10326390731228648,
+      "learning_rate": 1.4938059159964555e-05,
+      "loss": 0.1161,
+      "step": 1130
+    },
+    {
+      "epoch": 1.0759792354884379,
+      "grad_norm": 0.09969628777021497,
+      "learning_rate": 1.4849136625363297e-05,
+      "loss": 0.1141,
+      "step": 1140
+    },
+    {
+      "epoch": 1.0854176498348278,
+      "grad_norm": 0.0939091288214549,
+      "learning_rate": 1.4759709920754453e-05,
+      "loss": 0.1125,
+      "step": 1150
+    },
+    {
+      "epoch": 1.0948560641812175,
+      "grad_norm": 0.09506323252337912,
+      "learning_rate": 1.4669788343929736e-05,
+      "loss": 0.1141,
+      "step": 1160
+    },
+    {
+      "epoch": 1.1042944785276074,
+      "grad_norm": 0.10154441851850998,
+      "learning_rate": 1.4579381244133265e-05,
+      "loss": 0.1128,
+      "step": 1170
+    },
+    {
+      "epoch": 1.1137328928739971,
+      "grad_norm": 0.11822773479215737,
+      "learning_rate": 1.4488498021089514e-05,
+      "loss": 0.1137,
+      "step": 1180
+    },
+    {
+      "epoch": 1.123171307220387,
+      "grad_norm": 0.1009800661484683,
+      "learning_rate": 1.4397148124025997e-05,
+      "loss": 0.1143,
+      "step": 1190
+    },
+    {
+      "epoch": 1.1326097215667768,
+      "grad_norm": 0.10126420585679885,
+      "learning_rate": 1.4305341050690845e-05,
+      "loss": 0.117,
+      "step": 1200
+    },
+    {
+      "epoch": 1.1420481359131667,
+      "grad_norm": 0.09588874040008494,
+      "learning_rate": 1.421308634636529e-05,
+      "loss": 0.1137,
+      "step": 1210
+    },
+    {
+      "epoch": 1.1514865502595564,
+      "grad_norm": 0.10922875430592754,
+      "learning_rate": 1.412039360287126e-05,
+      "loss": 0.1145,
+      "step": 1220
+    },
+    {
+      "epoch": 1.1609249646059463,
+      "grad_norm": 0.11298890031757797,
+      "learning_rate": 1.4027272457574082e-05,
+      "loss": 0.1138,
+      "step": 1230
+    },
+    {
+      "epoch": 1.170363378952336,
+      "grad_norm": 0.10857815603466196,
+      "learning_rate": 1.3933732592380485e-05,
+      "loss": 0.1135,
+      "step": 1240
+    },
+    {
+      "epoch": 1.1798017932987257,
+      "grad_norm": 0.10213279825434321,
+      "learning_rate": 1.3839783732731966e-05,
+      "loss": 0.1134,
+      "step": 1250
+    },
+    {
+      "epoch": 1.1892402076451156,
+      "grad_norm": 0.10027833977041692,
+      "learning_rate": 1.3745435646593613e-05,
+      "loss": 0.1136,
+      "step": 1260
+    },
+    {
+      "epoch": 1.1986786219915055,
+      "grad_norm": 0.09590585357482817,
+      "learning_rate": 1.3650698143438534e-05,
+      "loss": 0.113,
+      "step": 1270
+    },
+    {
+      "epoch": 1.2081170363378952,
+      "grad_norm": 0.10321073236266613,
+      "learning_rate": 1.3555581073227942e-05,
+      "loss": 0.1167,
+      "step": 1280
+    },
+    {
+      "epoch": 1.217555450684285,
+      "grad_norm": 0.09327710523878686,
+      "learning_rate": 1.346009432538705e-05,
+      "loss": 0.1147,
+      "step": 1290
+    },
+    {
+      "epoch": 1.2269938650306749,
+      "grad_norm": 0.0933025296287067,
+      "learning_rate": 1.3364247827776854e-05,
+      "loss": 0.1145,
+      "step": 1300
+    },
+    {
+      "epoch": 1.2364322793770646,
+      "grad_norm": 0.09493771082819326,
+      "learning_rate": 1.3268051545661937e-05,
+      "loss": 0.1141,
+      "step": 1310
+    },
+    {
+      "epoch": 1.2458706937234545,
+      "grad_norm": 0.10053484854502866,
+      "learning_rate": 1.3171515480674342e-05,
+      "loss": 0.1122,
+      "step": 1320
+    },
+    {
+      "epoch": 1.2553091080698442,
+      "grad_norm": 0.1108335770631105,
+      "learning_rate": 1.3074649669773716e-05,
+      "loss": 0.1173,
+      "step": 1330
+    },
+    {
+      "epoch": 1.2647475224162341,
+      "grad_norm": 0.10521299166726314,
+      "learning_rate": 1.297746418420374e-05,
+      "loss": 0.1103,
+      "step": 1340
+    },
+    {
+      "epoch": 1.2741859367626238,
+      "grad_norm": 0.10478814209881943,
+      "learning_rate": 1.2879969128445025e-05,
+      "loss": 0.1122,
+      "step": 1350
+    },
+    {
+      "epoch": 1.2836243511090137,
+      "grad_norm": 0.0969588608522638,
+      "learning_rate": 1.2782174639164528e-05,
+      "loss": 0.1112,
+      "step": 1360
+    },
+    {
+      "epoch": 1.2930627654554034,
+      "grad_norm": 0.10783687256200783,
+      "learning_rate": 1.2684090884161636e-05,
+      "loss": 0.1125,
+      "step": 1370
+    },
+    {
+      "epoch": 1.3025011798017934,
+      "grad_norm": 0.10076692580892369,
+      "learning_rate": 1.2585728061311003e-05,
+      "loss": 0.1107,
+      "step": 1380
+    },
+    {
+      "epoch": 1.311939594148183,
+      "grad_norm": 0.09895358395270354,
+      "learning_rate": 1.248709639750228e-05,
+      "loss": 0.1122,
+      "step": 1390
+    },
+    {
+      "epoch": 1.321378008494573,
+      "grad_norm": 0.10215738990006902,
+      "learning_rate": 1.2388206147576796e-05,
+      "loss": 0.1124,
+      "step": 1400
+    },
+    {
+      "epoch": 1.3308164228409627,
+      "grad_norm": 0.09611638665301472,
+      "learning_rate": 1.2289067593261358e-05,
+      "loss": 0.1151,
+      "step": 1410
+    },
+    {
+      "epoch": 1.3402548371873526,
+      "grad_norm": 0.09899073401009041,
+      "learning_rate": 1.2189691042099265e-05,
+      "loss": 0.1124,
+      "step": 1420
+    },
+    {
+      "epoch": 1.3496932515337423,
+      "grad_norm": 0.1157109248340035,
+      "learning_rate": 1.209008682637859e-05,
+      "loss": 0.1154,
+      "step": 1430
+    },
+    {
+      "epoch": 1.359131665880132,
+      "grad_norm": 0.09358448441833775,
+      "learning_rate": 1.1990265302057948e-05,
+      "loss": 0.1127,
+      "step": 1440
+    },
+    {
+      "epoch": 1.368570080226522,
+      "grad_norm": 0.10318474117014907,
+      "learning_rate": 1.1890236847689762e-05,
+      "loss": 0.1134,
+      "step": 1450
+    },
+    {
+      "epoch": 1.3780084945729119,
+      "grad_norm": 0.10507403584009326,
+      "learning_rate": 1.1790011863341197e-05,
+      "loss": 0.1145,
+      "step": 1460
+    },
+    {
+      "epoch": 1.3874469089193016,
+      "grad_norm": 0.09210160678217687,
+      "learning_rate": 1.1689600769512855e-05,
+      "loss": 0.1128,
+      "step": 1470
+    },
+    {
+      "epoch": 1.3968853232656913,
+      "grad_norm": 0.09890271495599744,
+      "learning_rate": 1.1589014006055337e-05,
+      "loss": 0.1158,
+      "step": 1480
+    },
+    {
+      "epoch": 1.4063237376120812,
+      "grad_norm": 0.09768042621781026,
+      "learning_rate": 1.1488262031083816e-05,
+      "loss": 0.1107,
+      "step": 1490
+    },
+    {
+      "epoch": 1.415762151958471,
+      "grad_norm": 0.09501221720590393,
+      "learning_rate": 1.1387355319890685e-05,
+      "loss": 0.1138,
+      "step": 1500
+    },
+    {
+      "epoch": 1.4252005663048608,
+      "grad_norm": 0.09098951194154016,
+      "learning_rate": 1.1286304363856418e-05,
+      "loss": 0.112,
+      "step": 1510
+    },
+    {
+      "epoch": 1.4346389806512505,
+      "grad_norm": 0.08856225785605727,
+      "learning_rate": 1.1185119669358792e-05,
+      "loss": 0.1137,
+      "step": 1520
+    },
+    {
+      "epoch": 1.4440773949976404,
+      "grad_norm": 0.09111265321801558,
+      "learning_rate": 1.1083811756680523e-05,
+      "loss": 0.1093,
+      "step": 1530
+    },
+    {
+      "epoch": 1.4535158093440301,
+      "grad_norm": 0.09329783759350743,
+      "learning_rate": 1.0982391158915441e-05,
+      "loss": 0.1138,
+      "step": 1540
+    },
+    {
+      "epoch": 1.46295422369042,
+      "grad_norm": 0.09114905230583735,
+      "learning_rate": 1.0880868420873375e-05,
+      "loss": 0.1135,
+      "step": 1550
+    },
+    {
+      "epoch": 1.4723926380368098,
+      "grad_norm": 0.10049302399783284,
+      "learning_rate": 1.0779254097983788e-05,
+      "loss": 0.1104,
+      "step": 1560
+    },
+    {
+      "epoch": 1.4818310523831997,
+      "grad_norm": 0.08811790258148439,
+      "learning_rate": 1.0677558755198327e-05,
+      "loss": 0.114,
+      "step": 1570
+    },
+    {
+      "epoch": 1.4912694667295894,
+      "grad_norm": 0.09402068008649977,
+      "learning_rate": 1.0575792965892349e-05,
+      "loss": 0.1112,
+      "step": 1580
+    },
+    {
+      "epoch": 1.500707881075979,
+      "grad_norm": 0.09144293350756144,
+      "learning_rate": 1.0473967310765629e-05,
+      "loss": 0.1099,
+      "step": 1590
+    },
+    {
+      "epoch": 1.510146295422369,
+      "grad_norm": 0.08887643971229772,
+      "learning_rate": 1.0372092376742247e-05,
+      "loss": 0.1109,
+      "step": 1600
+    },
+    {
+      "epoch": 1.519584709768759,
+      "grad_norm": 0.09042876745354687,
+      "learning_rate": 1.0270178755869861e-05,
+      "loss": 0.1123,
+      "step": 1610
+    },
+    {
+      "epoch": 1.5290231241151486,
+      "grad_norm": 0.08799872003450031,
+      "learning_rate": 1.0168237044218452e-05,
+      "loss": 0.1088,
+      "step": 1620
+    },
+    {
+      "epoch": 1.5384615384615383,
+      "grad_norm": 0.08558681617619375,
+      "learning_rate": 1.0066277840778626e-05,
+      "loss": 0.1125,
+      "step": 1630
+    },
+    {
+      "epoch": 1.5478999528079282,
+      "grad_norm": 0.08917702916102198,
+      "learning_rate": 9.964311746359631e-06,
+      "loss": 0.1078,
+      "step": 1640
+    },
+    {
+      "epoch": 1.5573383671543182,
+      "grad_norm": 0.09365014825092945,
+      "learning_rate": 9.862349362487172e-06,
+      "loss": 0.108,
+      "step": 1650
+    },
+    {
+      "epoch": 1.5667767815007079,
+      "grad_norm": 0.08881624424784158,
+      "learning_rate": 9.760401290301164e-06,
+      "loss": 0.1073,
+      "step": 1660
+    },
+    {
+      "epoch": 1.5762151958470976,
+      "grad_norm": 0.09040629927788134,
+      "learning_rate": 9.658478129453532e-06,
+      "loss": 0.1095,
+      "step": 1670
+    },
+    {
+      "epoch": 1.5856536101934875,
+      "grad_norm": 0.09668654356646131,
+      "learning_rate": 9.556590477006123e-06,
+      "loss": 0.109,
+      "step": 1680
+    },
+    {
+      "epoch": 1.5950920245398774,
+      "grad_norm": 0.08945527030759594,
+      "learning_rate": 9.454748926328962e-06,
+      "loss": 0.1111,
+      "step": 1690
+    },
+    {
+      "epoch": 1.6045304388862671,
+      "grad_norm": 0.09096700102455489,
+      "learning_rate": 9.352964065998801e-06,
+      "loss": 0.1091,
+      "step": 1700
+    },
+    {
+      "epoch": 1.6139688532326568,
+      "grad_norm": 0.098227701202526,
+      "learning_rate": 9.251246478698242e-06,
+      "loss": 0.1124,
+      "step": 1710
+    },
+    {
+      "epoch": 1.6234072675790467,
+      "grad_norm": 0.08702267003826049,
+      "learning_rate": 9.149606740115444e-06,
+      "loss": 0.1091,
+      "step": 1720
+    },
+    {
+      "epoch": 1.6328456819254367,
+      "grad_norm": 0.0912741057496456,
+      "learning_rate": 9.04805541784454e-06,
+      "loss": 0.1084,
+      "step": 1730
+    },
+    {
+      "epoch": 1.6422840962718264,
+      "grad_norm": 0.09473199957469115,
+      "learning_rate": 8.946603070286926e-06,
+      "loss": 0.1071,
+      "step": 1740
+    },
+    {
+      "epoch": 1.651722510618216,
+      "grad_norm": 0.09786418318351309,
+      "learning_rate": 8.845260245553493e-06,
+      "loss": 0.1106,
+      "step": 1750
+    },
+    {
+      "epoch": 1.661160924964606,
+      "grad_norm": 0.09376970400308367,
+      "learning_rate": 8.744037480367922e-06,
+      "loss": 0.1095,
+      "step": 1760
+    },
+    {
+      "epoch": 1.670599339310996,
+      "grad_norm": 0.09927967174299174,
+      "learning_rate": 8.642945298971168e-06,
+      "loss": 0.1086,
+      "step": 1770
+    },
+    {
+      "epoch": 1.6800377536573856,
+      "grad_norm": 0.08941325714107755,
+      "learning_rate": 8.54199421202726e-06,
+      "loss": 0.1096,
+      "step": 1780
+    },
+    {
+      "epoch": 1.6894761680037753,
+      "grad_norm": 0.09076378690689199,
+      "learning_rate": 8.441194715530472e-06,
+      "loss": 0.111,
+      "step": 1790
+    },
+    {
+      "epoch": 1.6989145823501652,
+      "grad_norm": 0.08350889813597084,
+      "learning_rate": 8.340557289714055e-06,
+      "loss": 0.1089,
+      "step": 1800
+    },
+    {
+      "epoch": 1.708352996696555,
+      "grad_norm": 0.1102767062542333,
+      "learning_rate": 8.240092397960601e-06,
+      "loss": 0.1077,
+      "step": 1810
+    },
+    {
+      "epoch": 1.7177914110429446,
+      "grad_norm": 0.0928092514701947,
+      "learning_rate": 8.139810485714142e-06,
+      "loss": 0.109,
+      "step": 1820
+    },
+    {
+      "epoch": 1.7272298253893346,
+      "grad_norm": 0.08827741386544802,
+      "learning_rate": 8.03972197939414e-06,
+      "loss": 0.1103,
+      "step": 1830
+    },
+    {
+      "epoch": 1.7366682397357245,
+      "grad_norm": 0.09074532027658296,
+      "learning_rate": 7.939837285311425e-06,
+      "loss": 0.106,
+      "step": 1840
+    },
+    {
+      "epoch": 1.7461066540821142,
+      "grad_norm": 0.08863060213083432,
+      "learning_rate": 7.840166788586244e-06,
+      "loss": 0.1111,
+      "step": 1850
+    },
+    {
+      "epoch": 1.7555450684285039,
+      "grad_norm": 0.0873194235737418,
+      "learning_rate": 7.740720852068524e-06,
+      "loss": 0.1107,
+      "step": 1860
+    },
+    {
+      "epoch": 1.7649834827748938,
+      "grad_norm": 0.09463118593541094,
+      "learning_rate": 7.641509815260412e-06,
+      "loss": 0.1067,
+      "step": 1870
+    },
+    {
+      "epoch": 1.7744218971212837,
+      "grad_norm": 0.08277999614215281,
+      "learning_rate": 7.542543993241278e-06,
+      "loss": 0.1092,
+      "step": 1880
+    },
+    {
+      "epoch": 1.7838603114676734,
+      "grad_norm": 0.08350764579820083,
+      "learning_rate": 7.443833675595254e-06,
+      "loss": 0.1033,
+      "step": 1890
+    },
+    {
+      "epoch": 1.7932987258140631,
+      "grad_norm": 0.0892168369121696,
+      "learning_rate": 7.3453891253413935e-06,
+      "loss": 0.1088,
+      "step": 1900
+    },
+    {
+      "epoch": 1.802737140160453,
+      "grad_norm": 0.09068391166704463,
+      "learning_rate": 7.247220577866625e-06,
+      "loss": 0.1074,
+      "step": 1910
+    },
+    {
+      "epoch": 1.812175554506843,
+      "grad_norm": 0.09208367707491026,
+      "learning_rate": 7.149338239861579e-06,
+      "loss": 0.1069,
+      "step": 1920
+    },
+    {
+      "epoch": 1.8216139688532327,
+      "grad_norm": 0.09334448658561058,
+      "learning_rate": 7.051752288259366e-06,
+      "loss": 0.1051,
+      "step": 1930
+    },
+    {
+      "epoch": 1.8310523831996224,
+      "grad_norm": 0.0867013966015152,
+      "learning_rate": 6.954472869177479e-06,
+      "loss": 0.1071,
+      "step": 1940
+    },
+    {
+      "epoch": 1.8404907975460123,
+      "grad_norm": 0.08513824070105314,
+      "learning_rate": 6.857510096862901e-06,
+      "loss": 0.108,
+      "step": 1950
+    },
+    {
+      "epoch": 1.8499292118924022,
+      "grad_norm": 0.08880515925379688,
+      "learning_rate": 6.760874052640494e-06,
+      "loss": 0.1081,
+      "step": 1960
+    },
+    {
+      "epoch": 1.859367626238792,
+      "grad_norm": 0.09395118992476542,
+      "learning_rate": 6.664574783864862e-06,
+      "loss": 0.1079,
+      "step": 1970
+    },
+    {
+      "epoch": 1.8688060405851816,
+      "grad_norm": 0.09253843263366972,
+      "learning_rate": 6.568622302875682e-06,
+      "loss": 0.1068,
+      "step": 1980
+    },
+    {
+      "epoch": 1.8782444549315715,
+      "grad_norm": 0.09103342170778085,
+      "learning_rate": 6.473026585956736e-06,
+      "loss": 0.106,
+      "step": 1990
+    },
+    {
+      "epoch": 1.8876828692779613,
+      "grad_norm": 0.08266259287550586,
+      "learning_rate": 6.377797572298661e-06,
+      "loss": 0.1076,
+      "step": 2000
+    },
+    {
+      "epoch": 1.897121283624351,
+      "grad_norm": 0.08360823987901486,
+      "learning_rate": 6.282945162965548e-06,
+      "loss": 0.1079,
+      "step": 2010
+    },
+    {
+      "epoch": 1.9065596979707409,
+      "grad_norm": 0.09004863363551058,
+      "learning_rate": 6.188479219865529e-06,
+      "loss": 0.1064,
+      "step": 2020
+    },
+    {
+      "epoch": 1.9159981123171308,
+      "grad_norm": 0.0931263680297109,
+      "learning_rate": 6.094409564725435e-06,
+      "loss": 0.1054,
+      "step": 2030
+    },
+    {
+      "epoch": 1.9254365266635205,
+      "grad_norm": 0.09051504504706874,
+      "learning_rate": 6.0007459780695885e-06,
+      "loss": 0.1082,
+      "step": 2040
+    },
+    {
+      "epoch": 1.9348749410099102,
+      "grad_norm": 0.0904194328886728,
+      "learning_rate": 5.907498198202939e-06,
+      "loss": 0.1081,
+      "step": 2050
+    },
+    {
+      "epoch": 1.9443133553563001,
+      "grad_norm": 0.08711490496456058,
+      "learning_rate": 5.8146759201985525e-06,
+      "loss": 0.1069,
+      "step": 2060
+    },
+    {
+      "epoch": 1.95375176970269,
+      "grad_norm": 0.08686808156213838,
+      "learning_rate": 5.722288794889603e-06,
+      "loss": 0.1064,
+      "step": 2070
+    },
+    {
+      "epoch": 1.9631901840490797,
+      "grad_norm": 0.08817123970600604,
+      "learning_rate": 5.630346427865965e-06,
+      "loss": 0.1045,
+      "step": 2080
+    },
+    {
+      "epoch": 1.9726285983954694,
+      "grad_norm": 0.08339231796976605,
+      "learning_rate": 5.538858378475508e-06,
+      "loss": 0.1066,
+      "step": 2090
+    },
+    {
+      "epoch": 1.9820670127418594,
+      "grad_norm": 0.08824324960180577,
+      "learning_rate": 5.447834158830202e-06,
+      "loss": 0.1037,
+      "step": 2100
+    },
+    {
+      "epoch": 1.9915054270882493,
+      "grad_norm": 0.09311929590923752,
+      "learning_rate": 5.357283232817147e-06,
+      "loss": 0.1054,
+      "step": 2110
+    },
+    {
+      "epoch": 2.0009438414346388,
+      "grad_norm": 0.1334431719993605,
+      "learning_rate": 5.267215015114574e-06,
+      "loss": 0.1031,
+      "step": 2120
+    },
+    {
+      "epoch": 2.0103822557810287,
+      "grad_norm": 0.10361861757914052,
+      "learning_rate": 5.177638870213008e-06,
+      "loss": 0.0868,
+      "step": 2130
+    },
+    {
+      "epoch": 2.0198206701274186,
+      "grad_norm": 0.084172782289755,
+      "learning_rate": 5.088564111441645e-06,
+      "loss": 0.0834,
+      "step": 2140
+    },
+    {
+      "epoch": 2.0292590844738085,
+      "grad_norm": 0.08727410339549913,
+      "learning_rate": 5.000000000000003e-06,
+      "loss": 0.0852,
+      "step": 2150
+    },
+    {
+      "epoch": 2.038697498820198,
+      "grad_norm": 0.08994662846331379,
+      "learning_rate": 4.911955743995042e-06,
+      "loss": 0.0845,
+      "step": 2160
+    },
+    {
+      "epoch": 2.048135913166588,
+      "grad_norm": 0.08715835430152602,
+      "learning_rate": 4.824440497483802e-06,
+      "loss": 0.0847,
+      "step": 2170
+    },
+    {
+      "epoch": 2.057574327512978,
+      "grad_norm": 0.0943214139106939,
+      "learning_rate": 4.737463359521618e-06,
+      "loss": 0.0845,
+      "step": 2180
+    },
+    {
+      "epoch": 2.067012741859368,
+      "grad_norm": 0.09049982951899538,
+      "learning_rate": 4.6510333732160915e-06,
+      "loss": 0.085,
+      "step": 2190
+    },
+    {
+      "epoch": 2.0764511562057573,
+      "grad_norm": 0.08823142440331375,
+      "learning_rate": 4.565159524786888e-06,
+      "loss": 0.0867,
+      "step": 2200
+    },
+    {
+      "epoch": 2.085889570552147,
+      "grad_norm": 0.08727162151071796,
+      "learning_rate": 4.479850742631396e-06,
+      "loss": 0.0834,
+      "step": 2210
+    },
+    {
+      "epoch": 2.095327984898537,
+      "grad_norm": 0.08243182067038303,
+      "learning_rate": 4.395115896396457e-06,
+      "loss": 0.0849,
+      "step": 2220
+    },
+    {
+      "epoch": 2.104766399244927,
+      "grad_norm": 0.08720723034547441,
+      "learning_rate": 4.310963796056168e-06,
+      "loss": 0.084,
+      "step": 2230
+    },
+    {
+      "epoch": 2.1142048135913165,
+      "grad_norm": 0.08588933137845103,
+      "learning_rate": 4.227403190995901e-06,
+      "loss": 0.0875,
+      "step": 2240
+    },
+    {
+      "epoch": 2.1236432279377064,
+      "grad_norm": 0.09200761679022347,
+      "learning_rate": 4.14444276910263e-06,
+      "loss": 0.0853,
+      "step": 2250
+    },
+    {
+      "epoch": 2.1330816422840964,
+      "grad_norm": 0.08831298949051568,
+      "learning_rate": 4.06209115586162e-06,
+      "loss": 0.0867,
+      "step": 2260
+    },
+    {
+      "epoch": 2.1425200566304863,
+      "grad_norm": 0.0893828115241757,
+      "learning_rate": 3.980356913459642e-06,
+      "loss": 0.0865,
+      "step": 2270
+    },
+    {
+      "epoch": 2.1519584709768758,
+      "grad_norm": 0.09119946740323005,
+      "learning_rate": 3.899248539894756e-06,
+      "loss": 0.0848,
+      "step": 2280
+    },
+    {
+      "epoch": 2.1613968853232657,
+      "grad_norm": 0.08882008929472095,
+      "learning_rate": 3.818774468092754e-06,
+      "loss": 0.0843,
+      "step": 2290
+    },
+    {
+      "epoch": 2.1708352996696556,
+      "grad_norm": 0.08922739717614447,
+      "learning_rate": 3.738943065030376e-06,
+      "loss": 0.0811,
+      "step": 2300
+    },
+    {
+      "epoch": 2.180273714016045,
+      "grad_norm": 0.08622476744102522,
+      "learning_rate": 3.659762630865411e-06,
+      "loss": 0.083,
+      "step": 2310
+    },
+    {
+      "epoch": 2.189712128362435,
+      "grad_norm": 0.08399251697806781,
+      "learning_rate": 3.5812413980736916e-06,
+      "loss": 0.0827,
+      "step": 2320
+    },
+    {
+      "epoch": 2.199150542708825,
+      "grad_norm": 0.09006770579644241,
+      "learning_rate": 3.5033875305931662e-06,
+      "loss": 0.0849,
+      "step": 2330
+    },
+    {
+      "epoch": 2.208588957055215,
+      "grad_norm": 0.08747795506814363,
+      "learning_rate": 3.4262091229750973e-06,
+      "loss": 0.0822,
+      "step": 2340
+    },
+    {
+      "epoch": 2.2180273714016043,
+      "grad_norm": 0.08840099962821243,
+      "learning_rate": 3.3497141995424397e-06,
+      "loss": 0.0835,
+      "step": 2350
+    },
+    {
+      "epoch": 2.2274657857479943,
+      "grad_norm": 0.08986511740506226,
+      "learning_rate": 3.2739107135555603e-06,
+      "loss": 0.0841,
+      "step": 2360
+    },
+    {
+      "epoch": 2.236904200094384,
+      "grad_norm": 0.08924913381765429,
+      "learning_rate": 3.1988065463853204e-06,
+      "loss": 0.0849,
+      "step": 2370
+    },
+    {
+      "epoch": 2.246342614440774,
+      "grad_norm": 0.08622955784811655,
+      "learning_rate": 3.1244095066936396e-06,
+      "loss": 0.0848,
+      "step": 2380
+    },
+    {
+      "epoch": 2.2557810287871636,
+      "grad_norm": 0.08819044440789944,
+      "learning_rate": 3.050727329621637e-06,
+      "loss": 0.0835,
+      "step": 2390
+    },
+    {
+      "epoch": 2.2652194431335535,
+      "grad_norm": 0.08900963097417651,
+      "learning_rate": 2.977767675985377e-06,
+      "loss": 0.0805,
+      "step": 2400
+    },
+    {
+      "epoch": 2.2746578574799434,
+      "grad_norm": 0.0889820260869723,
+      "learning_rate": 2.905538131479376e-06,
+      "loss": 0.0844,
+      "step": 2410
+    },
+    {
+      "epoch": 2.2840962718263333,
+      "grad_norm": 0.08508956855605507,
+      "learning_rate": 2.8340462058879214e-06,
+      "loss": 0.082,
+      "step": 2420
+    },
+    {
+      "epoch": 2.293534686172723,
+      "grad_norm": 0.0859290434178059,
+      "learning_rate": 2.76329933230425e-06,
+      "loss": 0.0819,
+      "step": 2430
+    },
+    {
+      "epoch": 2.3029731005191127,
+      "grad_norm": 0.08310233373376713,
+      "learning_rate": 2.6933048663577297e-06,
+      "loss": 0.0811,
+      "step": 2440
+    },
+    {
+      "epoch": 2.3124115148655027,
+      "grad_norm": 0.08745596999299074,
+      "learning_rate": 2.6240700854490988e-06,
+      "loss": 0.0824,
+      "step": 2450
+    },
+    {
+      "epoch": 2.3218499292118926,
+      "grad_norm": 0.08451577327209396,
+      "learning_rate": 2.5556021879938074e-06,
+      "loss": 0.0828,
+      "step": 2460
+    },
+    {
+      "epoch": 2.331288343558282,
+      "grad_norm": 0.09129945645731877,
+      "learning_rate": 2.4879082926735974e-06,
+      "loss": 0.0837,
+      "step": 2470
+    },
+    {
+      "epoch": 2.340726757904672,
+      "grad_norm": 0.083910957600724,
+      "learning_rate": 2.4209954376963797e-06,
+      "loss": 0.0816,
+      "step": 2480
+    },
+    {
+      "epoch": 2.350165172251062,
+      "grad_norm": 0.08447673558716574,
+      "learning_rate": 2.354870580064439e-06,
+      "loss": 0.0808,
+      "step": 2490
+    },
+    {
+      "epoch": 2.3596035865974514,
+      "grad_norm": 0.08671014168934867,
+      "learning_rate": 2.289540594851122e-06,
+      "loss": 0.0814,
+      "step": 2500
+    },
+    {
+      "epoch": 2.3690420009438413,
+      "grad_norm": 0.08155772855747594,
+      "learning_rate": 2.225012274486028e-06,
+      "loss": 0.0791,
+      "step": 2510
+    },
+    {
+      "epoch": 2.3784804152902312,
+      "grad_norm": 0.08555794455636312,
+      "learning_rate": 2.1612923280487883e-06,
+      "loss": 0.0843,
+      "step": 2520
+    },
+    {
+      "epoch": 2.387918829636621,
+      "grad_norm": 0.08566290155470521,
+      "learning_rate": 2.0983873805715216e-06,
+      "loss": 0.0837,
+      "step": 2530
+    },
+    {
+      "epoch": 2.397357243983011,
+      "grad_norm": 0.08531077534490512,
+      "learning_rate": 2.0363039723500155e-06,
+      "loss": 0.0838,
+      "step": 2540
+    },
+    {
+      "epoch": 2.4067956583294006,
+      "grad_norm": 0.08375539552341336,
+      "learning_rate": 1.9750485582637245e-06,
+      "loss": 0.0822,
+      "step": 2550
+    },
+    {
+      "epoch": 2.4162340726757905,
+      "grad_norm": 0.08503293134727295,
+      "learning_rate": 1.9146275071046626e-06,
+      "loss": 0.0849,
+      "step": 2560
+    },
+    {
+      "epoch": 2.4256724870221804,
+      "grad_norm": 0.08165485153496227,
+      "learning_rate": 1.8550471009152138e-06,
+      "loss": 0.0803,
+      "step": 2570
+    },
+    {
+      "epoch": 2.43511090136857,
+      "grad_norm": 0.08309504205099255,
+      "learning_rate": 1.7963135343349914e-06,
+      "loss": 0.0789,
+      "step": 2580
+    },
+    {
+      "epoch": 2.44454931571496,
+      "grad_norm": 0.08294260063335909,
+      "learning_rate": 1.73843291395678e-06,
+      "loss": 0.0823,
+      "step": 2590
+    },
+    {
+      "epoch": 2.4539877300613497,
+      "grad_norm": 0.08537750016844735,
+      "learning_rate": 1.6814112576916142e-06,
+      "loss": 0.0825,
+      "step": 2600
+    },
+    {
+      "epoch": 2.4634261444077397,
+      "grad_norm": 0.08433197560779272,
+      "learning_rate": 1.6252544941430982e-06,
+      "loss": 0.0813,
+      "step": 2610
+    },
+    {
+      "epoch": 2.472864558754129,
+      "grad_norm": 0.08321923421106348,
+      "learning_rate": 1.5699684619909983e-06,
+      "loss": 0.0826,
+      "step": 2620
+    },
+    {
+      "epoch": 2.482302973100519,
+      "grad_norm": 0.08373732680479463,
+      "learning_rate": 1.5155589093841939e-06,
+      "loss": 0.0802,
+      "step": 2630
+    },
+    {
+      "epoch": 2.491741387446909,
+      "grad_norm": 0.09112397953039225,
+      "learning_rate": 1.4620314933430269e-06,
+      "loss": 0.081,
+      "step": 2640
+    },
+    {
+      "epoch": 2.501179801793299,
+      "grad_norm": 0.08246005822696881,
+      "learning_rate": 1.4093917791711497e-06,
+      "loss": 0.0808,
+      "step": 2650
+    },
+    {
+      "epoch": 2.5106182161396884,
+      "grad_norm": 0.08399111524479627,
+      "learning_rate": 1.357645239876879e-06,
+      "loss": 0.08,
+      "step": 2660
+    },
+    {
+      "epoch": 2.5200566304860783,
+      "grad_norm": 0.08520663335129833,
+      "learning_rate": 1.3067972556041753e-06,
+      "loss": 0.0818,
+      "step": 2670
+    },
+    {
+      "epoch": 2.5294950448324682,
+      "grad_norm": 0.08299784168003543,
+      "learning_rate": 1.2568531130732498e-06,
+      "loss": 0.0807,
+      "step": 2680
+    },
+    {
+      "epoch": 2.5389334591788577,
+      "grad_norm": 0.08594453806774255,
+      "learning_rate": 1.207818005030904e-06,
+      "loss": 0.0802,
+      "step": 2690
+    },
+    {
+      "epoch": 2.5483718735252476,
+      "grad_norm": 0.08039958287229476,
+      "learning_rate": 1.1596970297106458e-06,
+      "loss": 0.0818,
+      "step": 2700
+    },
+    {
+      "epoch": 2.5578102878716376,
+      "grad_norm": 0.08085895400758615,
+      "learning_rate": 1.1124951903025981e-06,
+      "loss": 0.0806,
+      "step": 2710
+    },
+    {
+      "epoch": 2.5672487022180275,
+      "grad_norm": 0.08458215660632902,
+      "learning_rate": 1.0662173944333288e-06,
+      "loss": 0.081,
+      "step": 2720
+    },
+    {
+      "epoch": 2.5766871165644174,
+      "grad_norm": 0.0826446041832577,
+      "learning_rate": 1.0208684536555968e-06,
+      "loss": 0.081,
+      "step": 2730
+    },
+    {
+      "epoch": 2.586125530910807,
+      "grad_norm": 0.08442284791430928,
+      "learning_rate": 9.764530829480822e-07,
+      "loss": 0.0832,
+      "step": 2740
+    },
+    {
+      "epoch": 2.595563945257197,
+      "grad_norm": 0.08336636617446731,
+      "learning_rate": 9.329759002251726e-07,
+      "loss": 0.0802,
+      "step": 2750
+    },
+    {
+      "epoch": 2.6050023596035867,
+      "grad_norm": 0.0846992268529094,
+      "learning_rate": 8.904414258568306e-07,
+      "loss": 0.0799,
+      "step": 2760
+    },
+    {
+      "epoch": 2.614440773949976,
+      "grad_norm": 0.08300701320734614,
+      "learning_rate": 8.488540821986035e-07,
+      "loss": 0.0827,
+      "step": 2770
+    },
+    {
+      "epoch": 2.623879188296366,
+      "grad_norm": 0.08443465909555346,
+      "learning_rate": 8.082181931318311e-07,
+      "loss": 0.0792,
+      "step": 2780
+    },
+    {
+      "epoch": 2.633317602642756,
+      "grad_norm": 0.08476186997139065,
+      "learning_rate": 7.685379836140872e-07,
+      "loss": 0.079,
+      "step": 2790
+    },
+    {
+      "epoch": 2.642756016989146,
+      "grad_norm": 0.08472149270611445,
+      "learning_rate": 7.298175792398976e-07,
+      "loss": 0.0818,
+      "step": 2800
+    },
+    {
+      "epoch": 2.652194431335536,
+      "grad_norm": 0.08032205169166948,
+      "learning_rate": 6.920610058118105e-07,
+      "loss": 0.0804,
+      "step": 2810
+    },
+    {
+      "epoch": 2.6616328456819254,
+      "grad_norm": 0.08367697987102687,
+      "learning_rate": 6.552721889218194e-07,
+      "loss": 0.0816,
+      "step": 2820
+    },
+    {
+      "epoch": 2.6710712600283153,
+      "grad_norm": 0.08668064800964889,
+      "learning_rate": 6.194549535432137e-07,
+      "loss": 0.08,
+      "step": 2830
+    },
+    {
+      "epoch": 2.680509674374705,
+      "grad_norm": 0.0828804628433115,
+      "learning_rate": 5.846130236329073e-07,
+      "loss": 0.0823,
+      "step": 2840
+    },
+    {
+      "epoch": 2.6899480887210947,
+      "grad_norm": 0.08378018124842795,
+      "learning_rate": 5.507500217442341e-07,
+      "loss": 0.0809,
+      "step": 2850
+    },
+    {
+      "epoch": 2.6993865030674846,
+      "grad_norm": 0.08220414182971268,
+      "learning_rate": 5.178694686503205e-07,
+      "loss": 0.0784,
+      "step": 2860
+    },
+    {
+      "epoch": 2.7088249174138745,
+      "grad_norm": 0.08076213129882256,
+      "learning_rate": 4.85974782978027e-07,
+      "loss": 0.081,
+      "step": 2870
+    },
+    {
+      "epoch": 2.718263331760264,
+      "grad_norm": 0.08189865863600006,
+      "learning_rate": 4.5506928085250033e-07,
+      "loss": 0.0778,
+      "step": 2880
+    },
+    {
+      "epoch": 2.727701746106654,
+      "grad_norm": 0.08777777733586507,
+      "learning_rate": 4.251561755524036e-07,
+      "loss": 0.0832,
+      "step": 2890
+    },
+    {
+      "epoch": 2.737140160453044,
+      "grad_norm": 0.08438414185840341,
+      "learning_rate": 3.9623857717581813e-07,
+      "loss": 0.0808,
+      "step": 2900
+    },
+    {
+      "epoch": 2.746578574799434,
+      "grad_norm": 0.08245920785308986,
+      "learning_rate": 3.6831949231689203e-07,
+      "loss": 0.081,
+      "step": 2910
+    },
+    {
+      "epoch": 2.7560169891458237,
+      "grad_norm": 0.08356304284398074,
+      "learning_rate": 3.414018237532335e-07,
+      "loss": 0.0821,
+      "step": 2920
+    },
+    {
+      "epoch": 2.765455403492213,
+      "grad_norm": 0.08633702580648989,
+      "learning_rate": 3.154883701441136e-07,
+      "loss": 0.08,
+      "step": 2930
+    },
+    {
+      "epoch": 2.774893817838603,
+      "grad_norm": 0.0821243443006684,
+      "learning_rate": 2.905818257394799e-07,
+      "loss": 0.0798,
+      "step": 2940
+    },
+    {
+      "epoch": 2.784332232184993,
+      "grad_norm": 0.08233971677127812,
+      "learning_rate": 2.666847800998362e-07,
+      "loss": 0.0819,
+      "step": 2950
+    },
+    {
+      "epoch": 2.7937706465313825,
+      "grad_norm": 0.08176593022884952,
+      "learning_rate": 2.437997178270035e-07,
+      "loss": 0.0807,
+      "step": 2960
+    },
+    {
+      "epoch": 2.8032090608777724,
+      "grad_norm": 0.08485236840414098,
+      "learning_rate": 2.219290183057865e-07,
+      "loss": 0.0806,
+      "step": 2970
+    },
+    {
+      "epoch": 2.8126474752241624,
+      "grad_norm": 0.08134044819035506,
+      "learning_rate": 2.0107495545659829e-07,
+      "loss": 0.0778,
+      "step": 2980
+    },
+    {
+      "epoch": 2.8220858895705523,
+      "grad_norm": 0.08200978384685277,
+      "learning_rate": 1.8123969749902714e-07,
+      "loss": 0.0777,
+      "step": 2990
+    },
+    {
+      "epoch": 2.831524303916942,
+      "grad_norm": 0.08409107175117113,
+      "learning_rate": 1.6242530672641143e-07,
+      "loss": 0.0813,
+      "step": 3000
+    },
+    {
+      "epoch": 2.8409627182633317,
+      "grad_norm": 0.08288246905003016,
+      "learning_rate": 1.4463373929141766e-07,
+      "loss": 0.0788,
+      "step": 3010
+    },
+    {
+      "epoch": 2.8504011326097216,
+      "grad_norm": 0.08212980978941326,
+      "learning_rate": 1.2786684500265546e-07,
+      "loss": 0.0819,
+      "step": 3020
+    },
+    {
+      "epoch": 2.8598395469561115,
+      "grad_norm": 0.0832626354470357,
+      "learning_rate": 1.1212636713235581e-07,
+      "loss": 0.0794,
+      "step": 3030
+    },
+    {
+      "epoch": 2.869277961302501,
+      "grad_norm": 0.0836378792472762,
+      "learning_rate": 9.741394223512057e-08,
+      "loss": 0.0814,
+      "step": 3040
+    },
+    {
+      "epoch": 2.878716375648891,
+      "grad_norm": 0.08017963643030881,
+      "learning_rate": 8.373109997776185e-08,
+      "loss": 0.0804,
+      "step": 3050
+    },
+    {
+      "epoch": 2.888154789995281,
+      "grad_norm": 0.08105963755024889,
+      "learning_rate": 7.10792629802659e-08,
+      "loss": 0.08,
+      "step": 3060
+    },
+    {
+      "epoch": 2.8975932043416703,
+      "grad_norm": 0.08196798527897163,
+      "learning_rate": 5.945974666788479e-08,
+      "loss": 0.0809,
+      "step": 3070
+    },
+    {
+      "epoch": 2.9070316186880603,
+      "grad_norm": 0.08445010118063752,
+      "learning_rate": 4.887375913436132e-08,
+      "loss": 0.0817,
+      "step": 3080
+    },
+    {
+      "epoch": 2.91647003303445,
+      "grad_norm": 0.0830340739185833,
+      "learning_rate": 3.932240101633178e-08,
+      "loss": 0.0823,
+      "step": 3090
+    },
+    {
+      "epoch": 2.92590844738084,
+      "grad_norm": 0.08107816362561618,
+      "learning_rate": 3.0806665378884106e-08,
+      "loss": 0.0764,
+      "step": 3100
+    },
+    {
+      "epoch": 2.93534686172723,
+      "grad_norm": 0.08498347611137054,
+      "learning_rate": 2.33274376123116e-08,
+      "loss": 0.0784,
+      "step": 3110
+    },
+    {
+      "epoch": 2.9447852760736195,
+      "grad_norm": 0.08214101571060746,
+      "learning_rate": 1.68854953400599e-08,
+      "loss": 0.0795,
+      "step": 3120
+    },
+    {
+      "epoch": 2.9542236904200094,
+      "grad_norm": 0.0849230883706386,
+      "learning_rate": 1.1481508337869429e-08,
+      "loss": 0.0801,
+      "step": 3130
+    },
+    {
+      "epoch": 2.9636621047663994,
+      "grad_norm": 0.09022687172778407,
+      "learning_rate": 7.1160384641455475e-09,
+      "loss": 0.0828,
+      "step": 3140
+    },
+    {
+      "epoch": 2.973100519112789,
+      "grad_norm": 0.08399540827662202,
+      "learning_rate": 3.7895396015374955e-09,
+      "loss": 0.0808,
+      "step": 3150
+    },
+    {
+      "epoch": 2.9825389334591788,
+      "grad_norm": 0.08027494442622467,
+      "learning_rate": 1.502357609749483e-09,
+      "loss": 0.0803,
+      "step": 3160
+    },
+    {
+      "epoch": 2.9919773478055687,
+      "grad_norm": 0.08480380742956374,
+      "learning_rate": 2.5473028957945234e-10,
+      "loss": 0.0813,
+      "step": 3170
+    },
+    {
+      "epoch": 2.9985842378480414,
+      "step": 3177,
+      "total_flos": 5.875342281120154e+16,
+      "train_loss": 0.11385704865424108,
+      "train_runtime": 162519.6807,
+      "train_samples_per_second": 3.441,
+      "train_steps_per_second": 0.02
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 3177,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 50000000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 5.875342281120154e+16,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:756f5b50a6aefa13e703583604e4feb8437b69a33cc7b4d6adcfda6a54a39860
+size 7224