upload 13 files

config.json

@@ -0,0 +1,39 @@
+{
+  "_name_or_path": "checkpoint-5110000",
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPT2LMHeadModel"
+  ],
+  "attn_pdrop": 0.1,
+  "bos_token_id": null,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 64000,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_ctx": 1024,
+  "n_embd": 1600,
+  "n_head": 25,
+  "n_inner": null,
+  "n_layer": 48,
+  "n_positions": 1600,
+  "reorder_and_upcast_attn": false,
+  "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "task_specific_params": {
+    "text-generation": {
+      "do_sample": true,
+      "max_length": 50
+    }
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.43.1",
+  "use_cache": true,
+  "vocab_size": 64013
+}

generation_config.json

@@ -0,0 +1,5 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": 64000,
+  "transformers_version": "4.43.1"
+}

model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4585ac58530bf64995988387b987d4ab87a3d331aca514d41265aa5c1d8e24ca
+size 4969621408

model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9d1b8362ea98220c61a86fa3920213e23788eb0dd41b0ad1f35ee84f192ba7d
+size 1352608960

model.safetensors.index.json

@@ -0,0 +1,587 @@
+{
+  "metadata": {
+    "total_size": 6322169600
+  },
+  "weight_map": {
+    "transformer.h.0.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.0.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.0.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.0.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.0.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.0.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.0.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.0.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.0.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.0.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.0.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.0.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.1.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.1.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.1.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.1.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.1.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.1.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.1.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.1.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.1.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.1.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.1.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.1.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.10.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.10.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.10.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.10.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.10.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.10.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.10.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.10.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.10.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.10.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.10.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.10.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.11.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.11.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.11.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.11.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.11.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.11.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.11.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.11.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.11.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.11.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.11.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.11.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.12.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.12.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.12.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.12.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.12.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.12.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.12.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.12.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.12.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.12.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.12.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.12.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.13.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.13.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.13.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.13.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.13.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.13.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.13.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.13.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.13.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.13.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.13.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.13.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.14.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.14.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.14.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.14.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.14.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.14.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.14.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.14.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.14.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.14.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.14.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.14.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.15.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.15.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.15.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.15.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.15.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.15.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.15.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.15.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.15.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.15.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.15.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.15.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.16.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.16.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.16.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.16.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.16.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.16.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.16.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.16.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.16.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.16.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.16.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.16.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.17.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.17.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.17.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.17.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.17.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.17.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.17.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.17.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.17.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.17.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.17.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.17.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.18.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.18.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.18.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.18.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.18.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.18.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.18.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.18.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.18.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.18.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.18.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.18.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.19.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.19.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.19.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.19.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.19.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.19.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.19.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.19.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.19.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.19.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.19.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.19.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.2.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.2.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.2.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.2.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.2.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.2.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.2.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.2.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.2.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.2.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.2.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.2.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.20.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.20.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.20.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.20.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.20.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.20.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.20.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.20.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.20.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.20.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.20.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.20.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.21.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.21.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.21.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.21.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.21.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.21.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.21.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.21.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.21.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.21.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.21.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.21.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.22.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.22.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.22.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.22.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.22.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.22.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.22.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.22.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.22.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.22.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.22.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.22.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.23.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.23.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.23.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.23.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.23.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.23.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.23.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.23.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.23.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.23.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.23.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.23.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.24.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.24.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.24.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.24.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.24.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.24.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.24.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.24.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.24.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.24.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.24.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.24.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.25.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.25.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.25.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.25.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.25.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.25.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.25.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.25.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.25.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.25.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.25.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.25.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.26.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.26.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.26.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.26.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.26.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.26.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.26.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.26.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.26.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.26.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.26.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.26.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.27.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.27.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.27.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.27.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.27.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.27.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.27.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.27.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.27.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.27.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.27.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.27.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.28.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.28.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.28.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.28.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.28.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.28.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.28.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.28.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.29.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.29.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.29.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.29.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.29.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.29.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.29.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.29.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.29.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.29.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.29.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.29.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.3.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.3.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.3.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.3.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.3.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.3.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.3.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.3.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.3.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.3.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.3.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.3.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.30.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.30.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.30.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.30.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.30.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.30.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.30.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.30.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.30.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.30.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.30.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.30.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.31.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.31.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.31.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.31.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.31.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.31.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.31.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.31.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.31.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.31.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.31.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.31.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.32.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.32.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.32.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.32.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.32.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.32.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.32.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.32.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.32.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.32.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.32.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.32.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.33.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.33.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.33.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.33.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.33.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.33.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.33.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.33.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.33.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.33.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.33.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.33.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.34.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.34.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.34.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.34.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.34.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.34.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.34.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.34.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.34.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.34.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.34.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.34.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.35.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.35.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.35.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.35.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.35.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.35.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.35.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.35.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.35.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.35.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.35.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.35.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.36.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.36.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.36.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.36.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.36.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.36.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.36.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.36.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.36.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.36.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.36.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.36.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.37.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.37.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.37.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.37.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.37.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.37.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.37.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.37.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.37.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.37.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.37.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.37.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.38.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.38.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.38.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.38.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.38.ln_1.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.38.ln_1.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.38.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.38.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.38.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.38.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.38.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.38.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.39.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.39.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.39.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.39.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.39.ln_1.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.39.ln_1.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.39.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.39.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.39.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.39.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.39.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.39.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.4.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.4.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.4.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.4.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.4.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.4.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.4.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.4.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.4.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.4.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.4.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.4.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.40.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.40.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.40.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.40.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.40.ln_1.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.40.ln_1.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.40.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.40.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.40.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.40.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.40.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.40.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.41.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.41.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.41.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.41.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.41.ln_1.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.41.ln_1.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.41.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.41.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.41.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.41.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.41.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.41.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.42.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.42.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.42.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.42.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.42.ln_1.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.42.ln_1.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.42.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.42.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.42.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.42.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.42.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.42.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.43.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.43.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.43.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.43.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.43.ln_1.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.43.ln_1.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.43.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.43.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.43.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.43.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.43.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.43.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.44.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.44.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.44.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.44.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.44.ln_1.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.44.ln_1.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.44.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.44.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.44.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.44.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.44.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.44.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.45.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.45.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.45.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.45.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.45.ln_1.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.45.ln_1.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.45.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.45.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.45.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.45.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.45.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.45.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.46.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.46.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.46.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.46.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.46.ln_1.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.46.ln_1.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.46.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.46.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.46.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.46.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.46.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.46.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.47.attn.c_attn.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.47.attn.c_attn.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.47.attn.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.47.attn.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.47.ln_1.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.47.ln_1.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.47.ln_2.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.47.ln_2.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.47.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.47.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.47.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
+    "transformer.h.47.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
+    "transformer.h.5.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.5.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.5.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.5.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.5.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.5.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.5.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.5.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.5.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.5.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.5.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.5.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.6.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.6.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.6.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.6.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.6.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.6.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.6.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.6.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.6.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.6.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.6.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.6.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.7.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.7.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.7.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.7.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.7.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.7.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.7.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.7.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.7.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.7.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.7.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.7.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.8.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.8.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.8.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.8.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.8.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.8.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.8.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.8.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.8.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.8.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.8.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.8.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.9.attn.c_attn.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.9.attn.c_attn.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.9.attn.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.9.attn.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.9.ln_1.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.9.ln_1.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.9.ln_2.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.9.ln_2.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.9.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.9.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.9.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
+    "transformer.h.9.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.ln_f.bias": "model-00002-of-00002.safetensors",
+    "transformer.ln_f.weight": "model-00002-of-00002.safetensors",
+    "transformer.wpe.weight": "model-00001-of-00002.safetensors",
+    "transformer.wte.weight": "model-00001-of-00002.safetensors"
+  }
+}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d8504a57030643a1313ec4b63f62f3977825cf248f1b378a2e13133f13749386
+size 12644692394

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbd215ac41a87b31b2cf521c3c9e31fb45aaed45d3b2deb1745d29ce8febeece
+size 14244

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d4c3980cd093b9f2878899bbbd52dede2273fc715dba17c7e4e055333d3bd6b
+size 1064

special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,120 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64000": {
+      "content": "<EOS>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64001": {
+      "content": "<PAD>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64002": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64003": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64004": {
+      "content": "<user>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64005": {
+      "content": "<model>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64006": {
+      "content": "<مستخدم>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64007": {
+      "content": "<نموذج>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64008": {
+      "content": "</مستخدم>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64009": {
+      "content": "</نموذج>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64010": {
+      "content": "<doc_sep>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64011": {
+      "content": "<فاصل>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64012": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "[PAD]",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}

trainer_state.json

@@ -0,0 +1,3778 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.9999943955298747,
+  "eval_steps": 500,
+  "global_step": 267642,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0037363134169146645,
+      "grad_norm": 72942.6796875,
+      "learning_rate": 9.999655549964018e-05,
+      "loss": 4.1478,
+      "step": 500
+    },
+    {
+      "epoch": 0.007472626833829329,
+      "grad_norm": 67110.5,
+      "learning_rate": 9.998622247314402e-05,
+      "loss": 3.6349,
+      "step": 1000
+    },
+    {
+      "epoch": 0.011208940250743994,
+      "grad_norm": 73347.7890625,
+      "learning_rate": 9.996900234419608e-05,
+      "loss": 3.4653,
+      "step": 1500
+    },
+    {
+      "epoch": 0.014945253667658658,
+      "grad_norm": 70530.5546875,
+      "learning_rate": 9.994489748538595e-05,
+      "loss": 3.3608,
+      "step": 2000
+    },
+    {
+      "epoch": 0.018681567084573324,
+      "grad_norm": 69644.8515625,
+      "learning_rate": 9.991391121788143e-05,
+      "loss": 3.2989,
+      "step": 2500
+    },
+    {
+      "epoch": 0.022417880501487988,
+      "grad_norm": 79574.8203125,
+      "learning_rate": 9.987604781097091e-05,
+      "loss": 3.1788,
+      "step": 3000
+    },
+    {
+      "epoch": 0.02615419391840265,
+      "grad_norm": 77448.375,
+      "learning_rate": 9.983131248147513e-05,
+      "loss": 3.1802,
+      "step": 3500
+    },
+    {
+      "epoch": 0.029890507335317316,
+      "grad_norm": 82995.328125,
+      "learning_rate": 9.977971139302844e-05,
+      "loss": 3.1495,
+      "step": 4000
+    },
+    {
+      "epoch": 0.03362682075223198,
+      "grad_norm": 77637.0625,
+      "learning_rate": 9.972125165522955e-05,
+      "loss": 3.0534,
+      "step": 4500
+    },
+    {
+      "epoch": 0.03736313416914665,
+      "grad_norm": 71801.3515625,
+      "learning_rate": 9.965594132266197e-05,
+      "loss": 3.0471,
+      "step": 5000
+    },
+    {
+      "epoch": 0.04109944758606131,
+      "grad_norm": 73166.8125,
+      "learning_rate": 9.958378939378425e-05,
+      "loss": 3.0568,
+      "step": 5500
+    },
+    {
+      "epoch": 0.044835761002975975,
+      "grad_norm": 82521.0390625,
+      "learning_rate": 9.950480580969022e-05,
+      "loss": 2.9987,
+      "step": 6000
+    },
+    {
+      "epoch": 0.048572074419890636,
+      "grad_norm": 73557.609375,
+      "learning_rate": 9.941900145273922e-05,
+      "loss": 2.9518,
+      "step": 6500
+    },
+    {
+      "epoch": 0.0523083878368053,
+      "grad_norm": 76367.0078125,
+      "learning_rate": 9.932638814505676e-05,
+      "loss": 2.9528,
+      "step": 7000
+    },
+    {
+      "epoch": 0.056044701253719964,
+      "grad_norm": 80728.796875,
+      "learning_rate": 9.922697864690573e-05,
+      "loss": 2.9501,
+      "step": 7500
+    },
+    {
+      "epoch": 0.05978101467063463,
+      "grad_norm": 71664.421875,
+      "learning_rate": 9.912078665492822e-05,
+      "loss": 2.9449,
+      "step": 8000
+    },
+    {
+      "epoch": 0.0635173280875493,
+      "grad_norm": 75999.953125,
+      "learning_rate": 9.90078268002584e-05,
+      "loss": 2.8971,
+      "step": 8500
+    },
+    {
+      "epoch": 0.06725364150446396,
+      "grad_norm": 67755.296875,
+      "learning_rate": 9.888811464650668e-05,
+      "loss": 2.8525,
+      "step": 9000
+    },
+    {
+      "epoch": 0.07098995492137862,
+      "grad_norm": 76703.8203125,
+      "learning_rate": 9.876166668761532e-05,
+      "loss": 2.8725,
+      "step": 9500
+    },
+    {
+      "epoch": 0.0747262683382933,
+      "grad_norm": 74384.6640625,
+      "learning_rate": 9.862850034558593e-05,
+      "loss": 2.8761,
+      "step": 10000
+    },
+    {
+      "epoch": 0.07846258175520796,
+      "grad_norm": 59161.4453125,
+      "learning_rate": 9.848863396807901e-05,
+      "loss": 2.8503,
+      "step": 10500
+    },
+    {
+      "epoch": 0.08219889517212262,
+      "grad_norm": 66133.3359375,
+      "learning_rate": 9.834208682588608e-05,
+      "loss": 2.8198,
+      "step": 11000
+    },
+    {
+      "epoch": 0.08593520858903728,
+      "grad_norm": 65402.43359375,
+      "learning_rate": 9.81888791102745e-05,
+      "loss": 2.8278,
+      "step": 11500
+    },
+    {
+      "epoch": 0.08967152200595195,
+      "grad_norm": 73789.2578125,
+      "learning_rate": 9.802903193020554e-05,
+      "loss": 2.808,
+      "step": 12000
+    },
+    {
+      "epoch": 0.09340783542286661,
+      "grad_norm": 71568.953125,
+      "learning_rate": 9.786256730942595e-05,
+      "loss": 2.7821,
+      "step": 12500
+    },
+    {
+      "epoch": 0.09714414883978127,
+      "grad_norm": 68736.6953125,
+      "learning_rate": 9.768950818343359e-05,
+      "loss": 2.7827,
+      "step": 13000
+    },
+    {
+      "epoch": 0.10088046225669595,
+      "grad_norm": 73235.984375,
+      "learning_rate": 9.750987839631733e-05,
+      "loss": 2.797,
+      "step": 13500
+    },
+    {
+      "epoch": 0.1046167756736106,
+      "grad_norm": 72581.59375,
+      "learning_rate": 9.732370269747181e-05,
+      "loss": 2.7515,
+      "step": 14000
+    },
+    {
+      "epoch": 0.10835308909052527,
+      "grad_norm": 77676.46875,
+      "learning_rate": 9.713100673818754e-05,
+      "loss": 2.7451,
+      "step": 14500
+    },
+    {
+      "epoch": 0.11208940250743993,
+      "grad_norm": 78359.078125,
+      "learning_rate": 9.693181706811652e-05,
+      "loss": 2.7544,
+      "step": 15000
+    },
+    {
+      "epoch": 0.1158257159243546,
+      "grad_norm": 67356.484375,
+      "learning_rate": 9.672616113161434e-05,
+      "loss": 2.7054,
+      "step": 15500
+    },
+    {
+      "epoch": 0.11956202934126926,
+      "grad_norm": 76034.65625,
+      "learning_rate": 9.651406726395895e-05,
+      "loss": 2.7419,
+      "step": 16000
+    },
+    {
+      "epoch": 0.12329834275818392,
+      "grad_norm": 77247.296875,
+      "learning_rate": 9.629556468744647e-05,
+      "loss": 2.7274,
+      "step": 16500
+    },
+    {
+      "epoch": 0.1270346561750986,
+      "grad_norm": 90000.84375,
+      "learning_rate": 9.607068350736501e-05,
+      "loss": 2.7118,
+      "step": 17000
+    },
+    {
+      "epoch": 0.13077096959201326,
+      "grad_norm": 82970.1953125,
+      "learning_rate": 9.583945470784683e-05,
+      "loss": 2.7036,
+      "step": 17500
+    },
+    {
+      "epoch": 0.13450728300892792,
+      "grad_norm": 70716.7578125,
+      "learning_rate": 9.560191014759922e-05,
+      "loss": 2.6936,
+      "step": 18000
+    },
+    {
+      "epoch": 0.13824359642584258,
+      "grad_norm": 69805.4140625,
+      "learning_rate": 9.535808255551515e-05,
+      "loss": 2.6633,
+      "step": 18500
+    },
+    {
+      "epoch": 0.14197990984275724,
+      "grad_norm": 69154.890625,
+      "learning_rate": 9.510800552616374e-05,
+      "loss": 2.6765,
+      "step": 19000
+    },
+    {
+      "epoch": 0.1457162232596719,
+      "grad_norm": 70633.8515625,
+      "learning_rate": 9.485171351516171e-05,
+      "loss": 2.6623,
+      "step": 19500
+    },
+    {
+      "epoch": 0.1494525366765866,
+      "grad_norm": 85893.5390625,
+      "learning_rate": 9.458924183442602e-05,
+      "loss": 2.6699,
+      "step": 20000
+    },
+    {
+      "epoch": 0.15318885009350125,
+      "grad_norm": 72002.125,
+      "learning_rate": 9.43206266473086e-05,
+      "loss": 2.6461,
+      "step": 20500
+    },
+    {
+      "epoch": 0.1569251635104159,
+      "grad_norm": 78129.734375,
+      "learning_rate": 9.404590496361382e-05,
+      "loss": 2.6687,
+      "step": 21000
+    },
+    {
+      "epoch": 0.16066147692733057,
+      "grad_norm": 64102.87890625,
+      "learning_rate": 9.376511463449922e-05,
+      "loss": 2.659,
+      "step": 21500
+    },
+    {
+      "epoch": 0.16439779034424523,
+      "grad_norm": 61640.66796875,
+      "learning_rate": 9.347829434726035e-05,
+      "loss": 2.6069,
+      "step": 22000
+    },
+    {
+      "epoch": 0.1681341037611599,
+      "grad_norm": 77315.6484375,
+      "learning_rate": 9.318548362000054e-05,
+      "loss": 2.6231,
+      "step": 22500
+    },
+    {
+      "epoch": 0.17187041717807455,
+      "grad_norm": 62667.19921875,
+      "learning_rate": 9.288672279618603e-05,
+      "loss": 2.6435,
+      "step": 23000
+    },
+    {
+      "epoch": 0.17560673059498924,
+      "grad_norm": 72954.6015625,
+      "learning_rate": 9.258205303908738e-05,
+      "loss": 2.6336,
+      "step": 23500
+    },
+    {
+      "epoch": 0.1793430440119039,
+      "grad_norm": 71380.78125,
+      "learning_rate": 9.227151632610811e-05,
+      "loss": 2.6389,
+      "step": 24000
+    },
+    {
+      "epoch": 0.18307935742881856,
+      "grad_norm": 72969.21875,
+      "learning_rate": 9.195515544300105e-05,
+      "loss": 2.6162,
+      "step": 24500
+    },
+    {
+      "epoch": 0.18681567084573322,
+      "grad_norm": 70472.6171875,
+      "learning_rate": 9.163301397797319e-05,
+      "loss": 2.6384,
+      "step": 25000
+    },
+    {
+      "epoch": 0.19055198426264788,
+      "grad_norm": 74207.796875,
+      "learning_rate": 9.130513631568021e-05,
+      "loss": 2.5962,
+      "step": 25500
+    },
+    {
+      "epoch": 0.19428829767956254,
+      "grad_norm": 66946.59375,
+      "learning_rate": 9.097156763111118e-05,
+      "loss": 2.607,
+      "step": 26000
+    },
+    {
+      "epoch": 0.1980246110964772,
+      "grad_norm": 75940.734375,
+      "learning_rate": 9.063235388336423e-05,
+      "loss": 2.5973,
+      "step": 26500
+    },
+    {
+      "epoch": 0.2017609245133919,
+      "grad_norm": 63575.37890625,
+      "learning_rate": 9.028754180931441e-05,
+      "loss": 2.6065,
+      "step": 27000
+    },
+    {
+      "epoch": 0.20549723793030655,
+      "grad_norm": 79057.234375,
+      "learning_rate": 8.993717891717425e-05,
+      "loss": 2.5717,
+      "step": 27500
+    },
+    {
+      "epoch": 0.2092335513472212,
+      "grad_norm": 72744.7578125,
+      "learning_rate": 8.958131347994808e-05,
+      "loss": 2.6101,
+      "step": 28000
+    },
+    {
+      "epoch": 0.21296986476413587,
+      "grad_norm": 77414.359375,
+      "learning_rate": 8.921999452878093e-05,
+      "loss": 2.5807,
+      "step": 28500
+    },
+    {
+      "epoch": 0.21670617818105053,
+      "grad_norm": 75271.28125,
+      "learning_rate": 8.885327184620314e-05,
+      "loss": 2.5952,
+      "step": 29000
+    },
+    {
+      "epoch": 0.2204424915979652,
+      "grad_norm": 69727.734375,
+      "learning_rate": 8.848119595927116e-05,
+      "loss": 2.587,
+      "step": 29500
+    },
+    {
+      "epoch": 0.22417880501487986,
+      "grad_norm": 83907.5859375,
+      "learning_rate": 8.810381813260606e-05,
+      "loss": 2.5377,
+      "step": 30000
+    },
+    {
+      "epoch": 0.22791511843179454,
+      "grad_norm": 93582.2265625,
+      "learning_rate": 8.772119036133022e-05,
+      "loss": 2.5808,
+      "step": 30500
+    },
+    {
+      "epoch": 0.2316514318487092,
+      "grad_norm": 69250.0703125,
+      "learning_rate": 8.733336536390349e-05,
+      "loss": 2.554,
+      "step": 31000
+    },
+    {
+      "epoch": 0.23538774526562387,
+      "grad_norm": 72463.203125,
+      "learning_rate": 8.69403965748596e-05,
+      "loss": 2.5899,
+      "step": 31500
+    },
+    {
+      "epoch": 0.23912405868253853,
+      "grad_norm": 71343.4765625,
+      "learning_rate": 8.654233813744392e-05,
+      "loss": 2.5627,
+      "step": 32000
+    },
+    {
+      "epoch": 0.2428603720994532,
+      "grad_norm": 78745.25,
+      "learning_rate": 8.613924489615373e-05,
+      "loss": 2.5353,
+      "step": 32500
+    },
+    {
+      "epoch": 0.24659668551636785,
+      "grad_norm": 81588.1875,
+      "learning_rate": 8.573117238918161e-05,
+      "loss": 2.5696,
+      "step": 33000
+    },
+    {
+      "epoch": 0.25033299893328254,
+      "grad_norm": 66313.8984375,
+      "learning_rate": 8.53181768407634e-05,
+      "loss": 2.5745,
+      "step": 33500
+    },
+    {
+      "epoch": 0.2540693123501972,
+      "grad_norm": 70478.8046875,
+      "learning_rate": 8.490031515343177e-05,
+      "loss": 2.5485,
+      "step": 34000
+    },
+    {
+      "epoch": 0.25780562576711186,
+      "grad_norm": 81928.53125,
+      "learning_rate": 8.447764490017597e-05,
+      "loss": 2.4977,
+      "step": 34500
+    },
+    {
+      "epoch": 0.2615419391840265,
+      "grad_norm": 65777.4921875,
+      "learning_rate": 8.405022431650959e-05,
+      "loss": 2.5664,
+      "step": 35000
+    },
+    {
+      "epoch": 0.2652782526009412,
+      "grad_norm": 81822.8828125,
+      "learning_rate": 8.361811229244682e-05,
+      "loss": 2.5249,
+      "step": 35500
+    },
+    {
+      "epoch": 0.26901456601785584,
+      "grad_norm": 66773.1328125,
+      "learning_rate": 8.318136836438853e-05,
+      "loss": 2.5385,
+      "step": 36000
+    },
+    {
+      "epoch": 0.2727508794347705,
+      "grad_norm": 78679.25,
+      "learning_rate": 8.274005270691942e-05,
+      "loss": 2.499,
+      "step": 36500
+    },
+    {
+      "epoch": 0.27648719285168516,
+      "grad_norm": 68307.140625,
+      "learning_rate": 8.229422612451711e-05,
+      "loss": 2.4764,
+      "step": 37000
+    },
+    {
+      "epoch": 0.2802235062685998,
+      "grad_norm": 67303.9453125,
+      "learning_rate": 8.184395004317458e-05,
+      "loss": 2.5165,
+      "step": 37500
+    },
+    {
+      "epoch": 0.2839598196855145,
+      "grad_norm": 69105.96875,
+      "learning_rate": 8.138928650193676e-05,
+      "loss": 2.5159,
+      "step": 38000
+    },
+    {
+      "epoch": 0.28769613310242914,
+      "grad_norm": 75505.984375,
+      "learning_rate": 8.093029814435293e-05,
+      "loss": 2.5268,
+      "step": 38500
+    },
+    {
+      "epoch": 0.2914324465193438,
+      "grad_norm": 80771.171875,
+      "learning_rate": 8.046704820984558e-05,
+      "loss": 2.5296,
+      "step": 39000
+    },
+    {
+      "epoch": 0.2951687599362585,
+      "grad_norm": 71207.0546875,
+      "learning_rate": 7.999960052499737e-05,
+      "loss": 2.518,
+      "step": 39500
+    },
+    {
+      "epoch": 0.2989050733531732,
+      "grad_norm": 69106.34375,
+      "learning_rate": 7.952801949475706e-05,
+      "loss": 2.4906,
+      "step": 40000
+    },
+    {
+      "epoch": 0.30264138677008784,
+      "grad_norm": 82679.9296875,
+      "learning_rate": 7.905237009356578e-05,
+      "loss": 2.4743,
+      "step": 40500
+    },
+    {
+      "epoch": 0.3063777001870025,
+      "grad_norm": 84038.28125,
+      "learning_rate": 7.857271785640486e-05,
+      "loss": 2.4726,
+      "step": 41000
+    },
+    {
+      "epoch": 0.31011401360391716,
+      "grad_norm": 75876.3203125,
+      "learning_rate": 7.808912886976641e-05,
+      "loss": 2.5081,
+      "step": 41500
+    },
+    {
+      "epoch": 0.3138503270208318,
+      "grad_norm": 78494.3046875,
+      "learning_rate": 7.760166976254804e-05,
+      "loss": 2.4795,
+      "step": 42000
+    },
+    {
+      "epoch": 0.3175866404377465,
+      "grad_norm": 68793.765625,
+      "learning_rate": 7.711040769687248e-05,
+      "loss": 2.4883,
+      "step": 42500
+    },
+    {
+      "epoch": 0.32132295385466114,
+      "grad_norm": 78543.03125,
+      "learning_rate": 7.661541035883427e-05,
+      "loss": 2.4752,
+      "step": 43000
+    },
+    {
+      "epoch": 0.3250592672715758,
+      "grad_norm": 78304.4140625,
+      "learning_rate": 7.611674594917376e-05,
+      "loss": 2.4747,
+      "step": 43500
+    },
+    {
+      "epoch": 0.32879558068849046,
+      "grad_norm": 77926.90625,
+      "learning_rate": 7.561448317388045e-05,
+      "loss": 2.4666,
+      "step": 44000
+    },
+    {
+      "epoch": 0.3325318941054051,
+      "grad_norm": 76948.8671875,
+      "learning_rate": 7.510869123472679e-05,
+      "loss": 2.4724,
+      "step": 44500
+    },
+    {
+      "epoch": 0.3362682075223198,
+      "grad_norm": 75918.4921875,
+      "learning_rate": 7.459943981973343e-05,
+      "loss": 2.4476,
+      "step": 45000
+    },
+    {
+      "epoch": 0.34000452093923444,
+      "grad_norm": 66944.8125,
+      "learning_rate": 7.408679909356764e-05,
+      "loss": 2.4982,
+      "step": 45500
+    },
+    {
+      "epoch": 0.3437408343561491,
+      "grad_norm": 68671.390625,
+      "learning_rate": 7.357083968787606e-05,
+      "loss": 2.4996,
+      "step": 46000
+    },
+    {
+      "epoch": 0.3474771477730638,
+      "grad_norm": 74974.25,
+      "learning_rate": 7.305163269155305e-05,
+      "loss": 2.44,
+      "step": 46500
+    },
+    {
+      "epoch": 0.3512134611899785,
+      "grad_norm": 83594.984375,
+      "learning_rate": 7.2529249640946e-05,
+      "loss": 2.466,
+      "step": 47000
+    },
+    {
+      "epoch": 0.35494977460689314,
+      "grad_norm": 67277.0546875,
+      "learning_rate": 7.200376250999917e-05,
+      "loss": 2.4398,
+      "step": 47500
+    },
+    {
+      "epoch": 0.3586860880238078,
+      "grad_norm": 64984.64453125,
+      "learning_rate": 7.1475243700337e-05,
+      "loss": 2.4536,
+      "step": 48000
+    },
+    {
+      "epoch": 0.36242240144072246,
+      "grad_norm": 76212.953125,
+      "learning_rate": 7.094376603128872e-05,
+      "loss": 2.4643,
+      "step": 48500
+    },
+    {
+      "epoch": 0.3661587148576371,
+      "grad_norm": 78486.7109375,
+      "learning_rate": 7.040940272985521e-05,
+      "loss": 2.4468,
+      "step": 49000
+    },
+    {
+      "epoch": 0.3698950282745518,
+      "grad_norm": 90899.328125,
+      "learning_rate": 6.987222742061982e-05,
+      "loss": 2.458,
+      "step": 49500
+    },
+    {
+      "epoch": 0.37363134169146645,
+      "grad_norm": 69692.6328125,
+      "learning_rate": 6.933231411560442e-05,
+      "loss": 2.4331,
+      "step": 50000
+    },
+    {
+      "epoch": 0.3773676551083811,
+      "grad_norm": 74715.71875,
+      "learning_rate": 6.878973720407191e-05,
+      "loss": 2.4379,
+      "step": 50500
+    },
+    {
+      "epoch": 0.38110396852529577,
+      "grad_norm": 72629.9140625,
+      "learning_rate": 6.824457144227698e-05,
+      "loss": 2.4424,
+      "step": 51000
+    },
+    {
+      "epoch": 0.3848402819422104,
+      "grad_norm": 74704.6328125,
+      "learning_rate": 6.769689194316615e-05,
+      "loss": 2.4905,
+      "step": 51500
+    },
+    {
+      "epoch": 0.3885765953591251,
+      "grad_norm": 75688.8125,
+      "learning_rate": 6.714677416602868e-05,
+      "loss": 2.4413,
+      "step": 52000
+    },
+    {
+      "epoch": 0.39231290877603975,
+      "grad_norm": 75325.578125,
+      "learning_rate": 6.659429390609984e-05,
+      "loss": 2.4505,
+      "step": 52500
+    },
+    {
+      "epoch": 0.3960492221929544,
+      "grad_norm": 70411.6953125,
+      "learning_rate": 6.603952728411776e-05,
+      "loss": 2.4358,
+      "step": 53000
+    },
+    {
+      "epoch": 0.3997855356098691,
+      "grad_norm": 71785.53125,
+      "learning_rate": 6.54825507358356e-05,
+      "loss": 2.4271,
+      "step": 53500
+    },
+    {
+      "epoch": 0.4035218490267838,
+      "grad_norm": 76337.5703125,
+      "learning_rate": 6.492344100149026e-05,
+      "loss": 2.4536,
+      "step": 54000
+    },
+    {
+      "epoch": 0.40725816244369845,
+      "grad_norm": 65393.2890625,
+      "learning_rate": 6.436227511522892e-05,
+      "loss": 2.4404,
+      "step": 54500
+    },
+    {
+      "epoch": 0.4109944758606131,
+      "grad_norm": 88779.578125,
+      "learning_rate": 6.379913039449547e-05,
+      "loss": 2.4277,
+      "step": 55000
+    },
+    {
+      "epoch": 0.41473078927752777,
+      "grad_norm": 70318.0546875,
+      "learning_rate": 6.32340844293777e-05,
+      "loss": 2.4209,
+      "step": 55500
+    },
+    {
+      "epoch": 0.4184671026944424,
+      "grad_norm": 77500.03125,
+      "learning_rate": 6.266721507191674e-05,
+      "loss": 2.4218,
+      "step": 56000
+    },
+    {
+      "epoch": 0.4222034161113571,
+      "grad_norm": 89454.5703125,
+      "learning_rate": 6.209860042538087e-05,
+      "loss": 2.409,
+      "step": 56500
+    },
+    {
+      "epoch": 0.42593972952827175,
+      "grad_norm": 72870.5,
+      "learning_rate": 6.152831883350426e-05,
+      "loss": 2.3841,
+      "step": 57000
+    },
+    {
+      "epoch": 0.4296760429451864,
+      "grad_norm": 75137.0703125,
+      "learning_rate": 6.095644886969283e-05,
+      "loss": 2.3922,
+      "step": 57500
+    },
+    {
+      "epoch": 0.43341235636210107,
+      "grad_norm": 64499.1328125,
+      "learning_rate": 6.038306932619846e-05,
+      "loss": 2.4042,
+      "step": 58000
+    },
+    {
+      "epoch": 0.43714866977901573,
+      "grad_norm": 74131.78125,
+      "learning_rate": 5.980825920326287e-05,
+      "loss": 2.3981,
+      "step": 58500
+    },
+    {
+      "epoch": 0.4408849831959304,
+      "grad_norm": 75569.671875,
+      "learning_rate": 5.923209769823308e-05,
+      "loss": 2.4412,
+      "step": 59000
+    },
+    {
+      "epoch": 0.44462129661284505,
+      "grad_norm": 94198.984375,
+      "learning_rate": 5.865466419464957e-05,
+      "loss": 2.4383,
+      "step": 59500
+    },
+    {
+      "epoch": 0.4483576100297597,
+      "grad_norm": 72894.984375,
+      "learning_rate": 5.807603825130873e-05,
+      "loss": 2.3944,
+      "step": 60000
+    },
+    {
+      "epoch": 0.45209392344667443,
+      "grad_norm": 75193.0234375,
+      "learning_rate": 5.7496299591301406e-05,
+      "loss": 2.3847,
+      "step": 60500
+    },
+    {
+      "epoch": 0.4558302368635891,
+      "grad_norm": 115278.234375,
+      "learning_rate": 5.6915528091028494e-05,
+      "loss": 2.3914,
+      "step": 61000
+    },
+    {
+      "epoch": 0.45956655028050375,
+      "grad_norm": 80114.7578125,
+      "learning_rate": 5.633380376919567e-05,
+      "loss": 2.394,
+      "step": 61500
+    },
+    {
+      "epoch": 0.4633028636974184,
+      "grad_norm": 77552.9921875,
+      "learning_rate": 5.5751206775788354e-05,
+      "loss": 2.3919,
+      "step": 62000
+    },
+    {
+      "epoch": 0.46703917711433307,
+      "grad_norm": 81552.4375,
+      "learning_rate": 5.516781738102871e-05,
+      "loss": 2.396,
+      "step": 62500
+    },
+    {
+      "epoch": 0.47077549053124773,
+      "grad_norm": 89432.1015625,
+      "learning_rate": 5.458371596431592e-05,
+      "loss": 2.4214,
+      "step": 63000
+    },
+    {
+      "epoch": 0.4745118039481624,
+      "grad_norm": 90061.796875,
+      "learning_rate": 5.3998983003151595e-05,
+      "loss": 2.3796,
+      "step": 63500
+    },
+    {
+      "epoch": 0.47824811736507705,
+      "grad_norm": 84587.1953125,
+      "learning_rate": 5.341369906205154e-05,
+      "loss": 2.39,
+      "step": 64000
+    },
+    {
+      "epoch": 0.4819844307819917,
+      "grad_norm": 63710.203125,
+      "learning_rate": 5.2827944781445583e-05,
+      "loss": 2.4094,
+      "step": 64500
+    },
+    {
+      "epoch": 0.4857207441989064,
+      "grad_norm": 69039.109375,
+      "learning_rate": 5.224180086656693e-05,
+      "loss": 2.3963,
+      "step": 65000
+    },
+    {
+      "epoch": 0.48945705761582103,
+      "grad_norm": 90888.671875,
+      "learning_rate": 5.165534807633261e-05,
+      "loss": 2.3701,
+      "step": 65500
+    },
+    {
+      "epoch": 0.4931933710327357,
+      "grad_norm": 75439.6875,
+      "learning_rate": 5.1068667212216527e-05,
+      "loss": 2.3588,
+      "step": 66000
+    },
+    {
+      "epoch": 0.49692968444965036,
+      "grad_norm": 77958.296875,
+      "learning_rate": 5.048183910711655e-05,
+      "loss": 2.3613,
+      "step": 66500
+    },
+    {
+      "epoch": 0.5006659978665651,
+      "grad_norm": 60339.5859375,
+      "learning_rate": 4.9894944614217454e-05,
+      "loss": 2.3501,
+      "step": 67000
+    },
+    {
+      "epoch": 0.5044023112834797,
+      "grad_norm": 88602.6015625,
+      "learning_rate": 4.930806459585094e-05,
+      "loss": 2.408,
+      "step": 67500
+    },
+    {
+      "epoch": 0.5081386247003944,
+      "grad_norm": 82704.09375,
+      "learning_rate": 4.872127991235435e-05,
+      "loss": 2.3974,
+      "step": 68000
+    },
+    {
+      "epoch": 0.511874938117309,
+      "grad_norm": 69181.0625,
+      "learning_rate": 4.8134671410929855e-05,
+      "loss": 2.337,
+      "step": 68500
+    },
+    {
+      "epoch": 0.5156112515342237,
+      "grad_norm": 83461.8984375,
+      "learning_rate": 4.75483199145052e-05,
+      "loss": 2.3468,
+      "step": 69000
+    },
+    {
+      "epoch": 0.5193475649511383,
+      "grad_norm": 70060.1875,
+      "learning_rate": 4.6962306210598024e-05,
+      "loss": 2.3686,
+      "step": 69500
+    },
+    {
+      "epoch": 0.523083878368053,
+      "grad_norm": 80763.5703125,
+      "learning_rate": 4.637671104018486e-05,
+      "loss": 2.3697,
+      "step": 70000
+    },
+    {
+      "epoch": 0.5268201917849676,
+      "grad_norm": 75881.0,
+      "learning_rate": 4.5791615086576726e-05,
+      "loss": 2.3789,
+      "step": 70500
+    },
+    {
+      "epoch": 0.5305565052018824,
+      "grad_norm": 68510.5390625,
+      "learning_rate": 4.5207098964302556e-05,
+      "loss": 2.3812,
+      "step": 71000
+    },
+    {
+      "epoch": 0.5342928186187971,
+      "grad_norm": 90292.1953125,
+      "learning_rate": 4.462324320800205e-05,
+      "loss": 2.3558,
+      "step": 71500
+    },
+    {
+      "epoch": 0.5380291320357117,
+      "grad_norm": 69267.453125,
+      "learning_rate": 4.4040128261329724e-05,
+      "loss": 2.3751,
+      "step": 72000
+    },
+    {
+      "epoch": 0.5417654454526264,
+      "grad_norm": 85016.8203125,
+      "learning_rate": 4.345783446587135e-05,
+      "loss": 2.3635,
+      "step": 72500
+    },
+    {
+      "epoch": 0.545501758869541,
+      "grad_norm": 103424.7109375,
+      "learning_rate": 4.2876442050074436e-05,
+      "loss": 2.3367,
+      "step": 73000
+    },
+    {
+      "epoch": 0.5492380722864557,
+      "grad_norm": 83033.2265625,
+      "learning_rate": 4.229603111819437e-05,
+      "loss": 2.371,
+      "step": 73500
+    },
+    {
+      "epoch": 0.5529743857033703,
+      "grad_norm": 89505.2890625,
+      "learning_rate": 4.1716681639257725e-05,
+      "loss": 2.3537,
+      "step": 74000
+    },
+    {
+      "epoch": 0.556710699120285,
+      "grad_norm": 86427.5390625,
+      "learning_rate": 4.113847343604404e-05,
+      "loss": 2.3518,
+      "step": 74500
+    },
+    {
+      "epoch": 0.5604470125371996,
+      "grad_norm": 95231.3203125,
+      "learning_rate": 4.05614861740879e-05,
+      "loss": 2.3475,
+      "step": 75000
+    },
+    {
+      "epoch": 0.5641833259541144,
+      "grad_norm": 73089.09375,
+      "learning_rate": 3.998579935070253e-05,
+      "loss": 2.3467,
+      "step": 75500
+    },
+    {
+      "epoch": 0.567919639371029,
+      "grad_norm": 82384.2109375,
+      "learning_rate": 3.941149228402674e-05,
+      "loss": 2.3278,
+      "step": 76000
+    },
+    {
+      "epoch": 0.5716559527879437,
+      "grad_norm": 84582.859375,
+      "learning_rate": 3.883864410209648e-05,
+      "loss": 2.3535,
+      "step": 76500
+    },
+    {
+      "epoch": 0.5753922662048583,
+      "grad_norm": 69513.640625,
+      "learning_rate": 3.826733373194247e-05,
+      "loss": 2.3257,
+      "step": 77000
+    },
+    {
+      "epoch": 0.579128579621773,
+      "grad_norm": 73027.9453125,
+      "learning_rate": 3.769763988871572e-05,
+      "loss": 2.3447,
+      "step": 77500
+    },
+    {
+      "epoch": 0.5828648930386876,
+      "grad_norm": 71039.1171875,
+      "learning_rate": 3.712964106484218e-05,
+      "loss": 2.3147,
+      "step": 78000
+    },
+    {
+      "epoch": 0.5866012064556023,
+      "grad_norm": 80346.1796875,
+      "learning_rate": 3.6563415519207985e-05,
+      "loss": 2.3144,
+      "step": 78500
+    },
+    {
+      "epoch": 0.590337519872517,
+      "grad_norm": 84110.34375,
+      "learning_rate": 3.599904126637692e-05,
+      "loss": 2.3333,
+      "step": 79000
+    },
+    {
+      "epoch": 0.5940738332894316,
+      "grad_norm": 82063.09375,
+      "learning_rate": 3.54365960658417e-05,
+      "loss": 2.3283,
+      "step": 79500
+    },
+    {
+      "epoch": 0.5978101467063464,
+      "grad_norm": 84510.8984375,
+      "learning_rate": 3.487615741131013e-05,
+      "loss": 2.3397,
+      "step": 80000
+    },
+    {
+      "epoch": 0.601546460123261,
+      "grad_norm": 74272.46875,
+      "learning_rate": 3.4317802520028117e-05,
+      "loss": 2.326,
+      "step": 80500
+    },
+    {
+      "epoch": 0.6052827735401757,
+      "grad_norm": 73607.890625,
+      "learning_rate": 3.37616083221406e-05,
+      "loss": 2.3238,
+      "step": 81000
+    },
+    {
+      "epoch": 0.6090190869570903,
+      "grad_norm": 84208.0078125,
+      "learning_rate": 3.320765145009216e-05,
+      "loss": 2.3112,
+      "step": 81500
+    },
+    {
+      "epoch": 0.612755400374005,
+      "grad_norm": 80192.90625,
+      "learning_rate": 3.265600822806866e-05,
+      "loss": 2.3361,
+      "step": 82000
+    },
+    {
+      "epoch": 0.6164917137909196,
+      "grad_norm": 83479.984375,
+      "learning_rate": 3.2106754661481086e-05,
+      "loss": 2.3383,
+      "step": 82500
+    },
+    {
+      "epoch": 0.6202280272078343,
+      "grad_norm": 88273.1015625,
+      "learning_rate": 3.1559966426493804e-05,
+      "loss": 2.3059,
+      "step": 83000
+    },
+    {
+      "epoch": 0.6239643406247489,
+      "grad_norm": 98511.15625,
+      "learning_rate": 3.101571885959766e-05,
+      "loss": 2.3113,
+      "step": 83500
+    },
+    {
+      "epoch": 0.6277006540416636,
+      "grad_norm": 76336.8046875,
+      "learning_rate": 3.047408694723032e-05,
+      "loss": 2.3254,
+      "step": 84000
+    },
+    {
+      "epoch": 0.6314369674585782,
+      "grad_norm": 80944.7734375,
+      "learning_rate": 2.99351453154444e-05,
+      "loss": 2.3428,
+      "step": 84500
+    },
+    {
+      "epoch": 0.635173280875493,
+      "grad_norm": 72115.8125,
+      "learning_rate": 2.9398968219625706e-05,
+      "loss": 2.3156,
+      "step": 85000
+    },
+    {
+      "epoch": 0.6389095942924077,
+      "grad_norm": 87770.3125,
+      "learning_rate": 2.8865629534262216e-05,
+      "loss": 2.3041,
+      "step": 85500
+    },
+    {
+      "epoch": 0.6426459077093223,
+      "grad_norm": 80694.6171875,
+      "learning_rate": 2.83352027427657e-05,
+      "loss": 2.3478,
+      "step": 86000
+    },
+    {
+      "epoch": 0.646382221126237,
+      "grad_norm": 77460.1171875,
+      "learning_rate": 2.7807760927347102e-05,
+      "loss": 2.302,
+      "step": 86500
+    },
+    {
+      "epoch": 0.6501185345431516,
+      "grad_norm": 82131.8671875,
+      "learning_rate": 2.7283376758947333e-05,
+      "loss": 2.3156,
+      "step": 87000
+    },
+    {
+      "epoch": 0.6538548479600663,
+      "grad_norm": 90870.65625,
+      "learning_rate": 2.6762122487224684e-05,
+      "loss": 2.331,
+      "step": 87500
+    },
+    {
+      "epoch": 0.6575911613769809,
+      "grad_norm": 98921.2109375,
+      "learning_rate": 2.624406993060019e-05,
+      "loss": 2.3073,
+      "step": 88000
+    },
+    {
+      "epoch": 0.6613274747938956,
+      "grad_norm": 73837.2109375,
+      "learning_rate": 2.5729290466362565e-05,
+      "loss": 2.301,
+      "step": 88500
+    },
+    {
+      "epoch": 0.6650637882108102,
+      "grad_norm": 82211.34375,
+      "learning_rate": 2.521785502083384e-05,
+      "loss": 2.3103,
+      "step": 89000
+    },
+    {
+      "epoch": 0.668800101627725,
+      "grad_norm": 86962.1484375,
+      "learning_rate": 2.4709834059597026e-05,
+      "loss": 2.3154,
+      "step": 89500
+    },
+    {
+      "epoch": 0.6725364150446396,
+      "grad_norm": 79155.3515625,
+      "learning_rate": 2.420529757778746e-05,
+      "loss": 2.348,
+      "step": 90000
+    },
+    {
+      "epoch": 0.6762727284615543,
+      "grad_norm": 97560.8125,
+      "learning_rate": 2.3704315090448886e-05,
+      "loss": 2.3417,
+      "step": 90500
+    },
+    {
+      "epoch": 0.6800090418784689,
+      "grad_norm": 124897.3203125,
+      "learning_rate": 2.3206955622955623e-05,
+      "loss": 2.3122,
+      "step": 91000
+    },
+    {
+      "epoch": 0.6837453552953836,
+      "grad_norm": 88038.5625,
+      "learning_rate": 2.271328770150227e-05,
+      "loss": 2.311,
+      "step": 91500
+    },
+    {
+      "epoch": 0.6874816687122982,
+      "grad_norm": 89480.2890625,
+      "learning_rate": 2.222337934366214e-05,
+      "loss": 2.3002,
+      "step": 92000
+    },
+    {
+      "epoch": 0.6912179821292129,
+      "grad_norm": 81038.4921875,
+      "learning_rate": 2.173729804901581e-05,
+      "loss": 2.3231,
+      "step": 92500
+    },
+    {
+      "epoch": 0.6949542955461276,
+      "grad_norm": 91312.8984375,
+      "learning_rate": 2.1255110789851084e-05,
+      "loss": 2.3125,
+      "step": 93000
+    },
+    {
+      "epoch": 0.6986906089630422,
+      "grad_norm": 84064.609375,
+      "learning_rate": 2.0776884001935437e-05,
+      "loss": 2.3379,
+      "step": 93500
+    },
+    {
+      "epoch": 0.702426922379957,
+      "grad_norm": 87531.078125,
+      "learning_rate": 2.0302683575362608e-05,
+      "loss": 2.2962,
+      "step": 94000
+    },
+    {
+      "epoch": 0.7061632357968716,
+      "grad_norm": 84962.34375,
+      "learning_rate": 1.983257484547421e-05,
+      "loss": 2.3032,
+      "step": 94500
+    },
+    {
+      "epoch": 0.7098995492137863,
+      "grad_norm": 88890.75,
+      "learning_rate": 1.9366622583857803e-05,
+      "loss": 2.3138,
+      "step": 95000
+    },
+    {
+      "epoch": 0.7136358626307009,
+      "grad_norm": 93013.0,
+      "learning_rate": 1.890489098942267e-05,
+      "loss": 2.3128,
+      "step": 95500
+    },
+    {
+      "epoch": 0.7173721760476156,
+      "grad_norm": 79075.3515625,
+      "learning_rate": 1.8447443679554567e-05,
+      "loss": 2.3013,
+      "step": 96000
+    },
+    {
+      "epoch": 0.7211084894645302,
+      "grad_norm": 94954.1640625,
+      "learning_rate": 1.7994343681350422e-05,
+      "loss": 2.2775,
+      "step": 96500
+    },
+    {
+      "epoch": 0.7248448028814449,
+      "grad_norm": 79968.1171875,
+      "learning_rate": 1.7545653422934537e-05,
+      "loss": 2.2979,
+      "step": 97000
+    },
+    {
+      "epoch": 0.7285811162983595,
+      "grad_norm": 90937.109375,
+      "learning_rate": 1.7101434724857147e-05,
+      "loss": 2.3466,
+      "step": 97500
+    },
+    {
+      "epoch": 0.7323174297152742,
+      "grad_norm": 85592.625,
+      "learning_rate": 1.6661748791576855e-05,
+      "loss": 2.3029,
+      "step": 98000
+    },
+    {
+      "epoch": 0.7360537431321889,
+      "grad_norm": 95187.1484375,
+      "learning_rate": 1.622665620302789e-05,
+      "loss": 2.3074,
+      "step": 98500
+    },
+    {
+      "epoch": 0.7397900565491036,
+      "grad_norm": 87762.859375,
+      "learning_rate": 1.5796216906273388e-05,
+      "loss": 2.2883,
+      "step": 99000
+    },
+    {
+      "epoch": 0.7435263699660182,
+      "grad_norm": 102495.0234375,
+      "learning_rate": 1.537049020724581e-05,
+      "loss": 2.2852,
+      "step": 99500
+    },
+    {
+      "epoch": 0.7472626833829329,
+      "grad_norm": 92540.0078125,
+      "learning_rate": 1.4949534762575912e-05,
+      "loss": 2.2914,
+      "step": 100000
+    },
+    {
+      "epoch": 0.7509989967998476,
+      "grad_norm": 81275.109375,
+      "learning_rate": 1.4533408571510904e-05,
+      "loss": 2.3119,
+      "step": 100500
+    },
+    {
+      "epoch": 0.7547353102167622,
+      "grad_norm": 82922.2578125,
+      "learning_rate": 1.4122168967923366e-05,
+      "loss": 2.3118,
+      "step": 101000
+    },
+    {
+      "epoch": 0.7584716236336769,
+      "grad_norm": 107727.1171875,
+      "learning_rate": 1.3715872612411806e-05,
+      "loss": 2.286,
+      "step": 101500
+    },
+    {
+      "epoch": 0.7622079370505915,
+      "grad_norm": 84810.40625,
+      "learning_rate": 1.3314575484493946e-05,
+      "loss": 2.3174,
+      "step": 102000
+    },
+    {
+      "epoch": 0.7659442504675062,
+      "grad_norm": 78007.40625,
+      "learning_rate": 1.2918332874893869e-05,
+      "loss": 2.3229,
+      "step": 102500
+    },
+    {
+      "epoch": 0.7696805638844209,
+      "grad_norm": 95875.2421875,
+      "learning_rate": 1.252719937792397e-05,
+      "loss": 2.2979,
+      "step": 103000
+    },
+    {
+      "epoch": 0.7734168773013356,
+      "grad_norm": 83561.375,
+      "learning_rate": 1.2141228883963118e-05,
+      "loss": 2.2667,
+      "step": 103500
+    },
+    {
+      "epoch": 0.7771531907182502,
+      "grad_norm": 73712.7578125,
+      "learning_rate": 1.1760474572031538e-05,
+      "loss": 2.2876,
+      "step": 104000
+    },
+    {
+      "epoch": 0.7808895041351649,
+      "grad_norm": 90907.0234375,
+      "learning_rate": 1.1384988902463806e-05,
+      "loss": 2.2799,
+      "step": 104500
+    },
+    {
+      "epoch": 0.7846258175520795,
+      "grad_norm": 84601.046875,
+      "learning_rate": 1.101482360968088e-05,
+      "loss": 2.3109,
+      "step": 105000
+    },
+    {
+      "epoch": 0.7883621309689942,
+      "grad_norm": 89593.640625,
+      "learning_rate": 1.0650029695062103e-05,
+      "loss": 2.3071,
+      "step": 105500
+    },
+    {
+      "epoch": 0.7920984443859088,
+      "grad_norm": 85944.5390625,
+      "learning_rate": 1.0290657419918308e-05,
+      "loss": 2.3275,
+      "step": 106000
+    },
+    {
+      "epoch": 0.7958347578028235,
+      "grad_norm": 71977.34375,
+      "learning_rate": 9.936756298566724e-06,
+      "loss": 2.3108,
+      "step": 106500
+    },
+    {
+      "epoch": 0.7995710712197382,
+      "grad_norm": 91883.7109375,
+      "learning_rate": 9.588375091508945e-06,
+      "loss": 2.2756,
+      "step": 107000
+    },
+    {
+      "epoch": 0.8033073846366529,
+      "grad_norm": 81686.0078125,
+      "learning_rate": 9.245561798712704e-06,
+      "loss": 2.3064,
+      "step": 107500
+    },
+    {
+      "epoch": 0.8070436980535676,
+      "grad_norm": 86408.875,
+      "learning_rate": 8.908363652998414e-06,
+      "loss": 2.3031,
+      "step": 108000
+    },
+    {
+      "epoch": 0.8107800114704822,
+      "grad_norm": 83010.9765625,
+      "learning_rate": 8.576827113531427e-06,
+      "loss": 2.2797,
+      "step": 108500
+    },
+    {
+      "epoch": 0.8145163248873969,
+      "grad_norm": 74965.84375,
+      "learning_rate": 8.25099785942094e-06,
+      "loss": 2.3011,
+      "step": 109000
+    },
+    {
+      "epoch": 0.8182526383043115,
+      "grad_norm": 91110.0859375,
+      "learning_rate": 7.930920783426265e-06,
+      "loss": 2.2572,
+      "step": 109500
+    },
+    {
+      "epoch": 0.8219889517212262,
+      "grad_norm": 96117.6640625,
+      "learning_rate": 7.616639985771556e-06,
+      "loss": 2.276,
+      "step": 110000
+    },
+    {
+      "epoch": 0.8257252651381408,
+      "grad_norm": 66883.046875,
+      "learning_rate": 7.30819876806963e-06,
+      "loss": 2.2586,
+      "step": 110500
+    },
+    {
+      "epoch": 0.8294615785550555,
+      "grad_norm": 86336.21875,
+      "learning_rate": 7.005639627355881e-06,
+      "loss": 2.2796,
+      "step": 111000
+    },
+    {
+      "epoch": 0.8331978919719701,
+      "grad_norm": 100607.5703125,
+      "learning_rate": 6.709004250233097e-06,
+      "loss": 2.2635,
+      "step": 111500
+    },
+    {
+      "epoch": 0.8369342053888849,
+      "grad_norm": 90908.2890625,
+      "learning_rate": 6.418333507127788e-06,
+      "loss": 2.2598,
+      "step": 112000
+    },
+    {
+      "epoch": 0.8406705188057995,
+      "grad_norm": 101526.234375,
+      "learning_rate": 6.133667446659131e-06,
+      "loss": 2.2648,
+      "step": 112500
+    },
+    {
+      "epoch": 0.8444068322227142,
+      "grad_norm": 87100.0234375,
+      "learning_rate": 5.855045290121047e-06,
+      "loss": 2.3005,
+      "step": 113000
+    },
+    {
+      "epoch": 0.8481431456396288,
+      "grad_norm": 95362.3984375,
+      "learning_rate": 5.58250542607826e-06,
+      "loss": 2.2747,
+      "step": 113500
+    },
+    {
+      "epoch": 0.8518794590565435,
+      "grad_norm": 105692.5625,
+      "learning_rate": 5.316085405077159e-06,
+      "loss": 2.2808,
+      "step": 114000
+    },
+    {
+      "epoch": 0.8556157724734582,
+      "grad_norm": 99888.703125,
+      "learning_rate": 5.055821934472071e-06,
+      "loss": 2.2707,
+      "step": 114500
+    },
+    {
+      "epoch": 0.8593520858903728,
+      "grad_norm": 89097.7421875,
+      "learning_rate": 4.801750873367728e-06,
+      "loss": 2.2996,
+      "step": 115000
+    },
+    {
+      "epoch": 0.8630883993072875,
+      "grad_norm": 94061.9453125,
+      "learning_rate": 4.553907227678594e-06,
+      "loss": 2.2585,
+      "step": 115500
+    },
+    {
+      "epoch": 0.8668247127242021,
+      "grad_norm": 89231.015625,
+      "learning_rate": 4.312325145305729e-06,
+      "loss": 2.2712,
+      "step": 116000
+    },
+    {
+      "epoch": 0.8705610261411169,
+      "grad_norm": 92370.8125,
+      "learning_rate": 4.077037911431913e-06,
+      "loss": 2.277,
+      "step": 116500
+    },
+    {
+      "epoch": 0.8742973395580315,
+      "grad_norm": 95057.8828125,
+      "learning_rate": 3.84807794393563e-06,
+      "loss": 2.2705,
+      "step": 117000
+    },
+    {
+      "epoch": 0.8780336529749462,
+      "grad_norm": 76279.6015625,
+      "learning_rate": 3.6254767889244813e-06,
+      "loss": 2.2692,
+      "step": 117500
+    },
+    {
+      "epoch": 0.8817699663918608,
+      "grad_norm": 99320.4375,
+      "learning_rate": 3.4092651163888154e-06,
+      "loss": 2.2841,
+      "step": 118000
+    },
+    {
+      "epoch": 0.8855062798087755,
+      "grad_norm": 97605.59375,
+      "learning_rate": 3.1994727159759986e-06,
+      "loss": 2.2942,
+      "step": 118500
+    },
+    {
+      "epoch": 0.8892425932256901,
+      "grad_norm": 87163.4765625,
+      "learning_rate": 2.996128492885969e-06,
+      "loss": 2.2918,
+      "step": 119000
+    },
+    {
+      "epoch": 0.8929789066426048,
+      "grad_norm": 78705.2421875,
+      "learning_rate": 2.7992604638886975e-06,
+      "loss": 2.259,
+      "step": 119500
+    },
+    {
+      "epoch": 0.8967152200595194,
+      "grad_norm": 94322.765625,
+      "learning_rate": 2.608895753464069e-06,
+      "loss": 2.2589,
+      "step": 120000
+    },
+    {
+      "epoch": 0.9004515334764341,
+      "grad_norm": 73016.4140625,
+      "learning_rate": 2.4250605900646204e-06,
+      "loss": 2.2938,
+      "step": 120500
+    },
+    {
+      "epoch": 0.9041878468933489,
+      "grad_norm": 90040.5546875,
+      "learning_rate": 2.2477803025018228e-06,
+      "loss": 2.2982,
+      "step": 121000
+    },
+    {
+      "epoch": 0.9079241603102635,
+      "grad_norm": 102110.484375,
+      "learning_rate": 2.0770793164562373e-06,
+      "loss": 2.2578,
+      "step": 121500
+    },
+    {
+      "epoch": 0.9116604737271782,
+      "grad_norm": 78370.1796875,
+      "learning_rate": 1.9129811511121697e-06,
+      "loss": 2.2854,
+      "step": 122000
+    },
+    {
+      "epoch": 0.9153967871440928,
+      "grad_norm": 87597.2578125,
+      "learning_rate": 1.7555084159172176e-06,
+      "loss": 2.2657,
+      "step": 122500
+    },
+    {
+      "epoch": 0.9191331005610075,
+      "grad_norm": 89568.2109375,
+      "learning_rate": 1.6046828074670894e-06,
+      "loss": 2.2674,
+      "step": 123000
+    },
+    {
+      "epoch": 0.9228694139779221,
+      "grad_norm": 80686.5234375,
+      "learning_rate": 1.4605251065162973e-06,
+      "loss": 2.2634,
+      "step": 123500
+    },
+    {
+      "epoch": 0.9266057273948368,
+      "grad_norm": 76351.1796875,
+      "learning_rate": 1.3230551751149568e-06,
+      "loss": 2.2889,
+      "step": 124000
+    },
+    {
+      "epoch": 0.9303420408117514,
+      "grad_norm": 82895.4921875,
+      "learning_rate": 1.1922919538721876e-06,
+      "loss": 2.2188,
+      "step": 124500
+    },
+    {
+      "epoch": 0.9340783542286661,
+      "grad_norm": 81276.09375,
+      "learning_rate": 1.0682534593464967e-06,
+      "loss": 2.268,
+      "step": 125000
+    },
+    {
+      "epoch": 0.9378146676455807,
+      "grad_norm": 83888.390625,
+      "learning_rate": 9.509567815634446e-07,
+      "loss": 2.2762,
+      "step": 125500
+    },
+    {
+      "epoch": 0.9415509810624955,
+      "grad_norm": 87934.8125,
+      "learning_rate": 8.404180816609819e-07,
+      "loss": 2.2759,
+      "step": 126000
+    },
+    {
+      "epoch": 0.9452872944794101,
+      "grad_norm": 89569.15625,
+      "learning_rate": 7.366525896627841e-07,
+      "loss": 2.2696,
+      "step": 126500
+    },
+    {
+      "epoch": 0.9490236078963248,
+      "grad_norm": 79137.96875,
+      "learning_rate": 6.396746023798139e-07,
+      "loss": 2.2951,
+      "step": 127000
+    },
+    {
+      "epoch": 0.9527599213132394,
+      "grad_norm": 74010.8984375,
+      "learning_rate": 5.49497481440564e-07,
+      "loss": 2.2875,
+      "step": 127500
+    },
+    {
+      "epoch": 0.9564962347301541,
+      "grad_norm": 70001.59375,
+      "learning_rate": 4.661336514500625e-07,
+      "loss": 2.2514,
+      "step": 128000
+    },
+    {
+      "epoch": 0.9602325481470688,
+      "grad_norm": 86343.359375,
+      "learning_rate": 3.895945982779925e-07,
+      "loss": 2.3027,
+      "step": 128500
+    },
+    {
+      "epoch": 0.9639688615639834,
+      "grad_norm": 92609.6171875,
+      "learning_rate": 3.19890867476208e-07,
+      "loss": 2.2857,
+      "step": 129000
+    },
+    {
+      "epoch": 0.9677051749808981,
+      "grad_norm": 91932.953125,
+      "learning_rate": 2.570320628257461e-07,
+      "loss": 2.2814,
+      "step": 129500
+    },
+    {
+      "epoch": 0.9714414883978127,
+      "grad_norm": 81636.1015625,
+      "learning_rate": 2.010268450136077e-07,
+      "loss": 2.2758,
+      "step": 130000
+    },
+    {
+      "epoch": 0.9751778018147275,
+      "grad_norm": 79906.6015625,
+      "learning_rate": 1.518829304395175e-07,
+      "loss": 2.2944,
+      "step": 130500
+    },
+    {
+      "epoch": 0.9789141152316421,
+      "grad_norm": 76946.6484375,
+      "learning_rate": 1.0960709015272464e-07,
+      "loss": 2.2305,
+      "step": 131000
+    },
+    {
+      "epoch": 0.9826504286485568,
+      "grad_norm": 81980.828125,
+      "learning_rate": 7.420514891911535e-08,
+      "loss": 2.2803,
+      "step": 131500
+    },
+    {
+      "epoch": 0.9863867420654714,
+      "grad_norm": 90421.125,
+      "learning_rate": 4.5681984418666224e-08,
+      "loss": 2.2767,
+      "step": 132000
+    },
+    {
+      "epoch": 0.9901230554823861,
+      "grad_norm": 81403.0390625,
+      "learning_rate": 2.404152657338732e-08,
+      "loss": 2.2399,
+      "step": 132500
+    },
+    {
+      "epoch": 0.9938593688993007,
+      "grad_norm": 98386.671875,
+      "learning_rate": 9.286757005871893e-09,
+      "loss": 2.2422,
+      "step": 133000
+    },
+    {
+      "epoch": 0.9975956823162154,
+      "grad_norm": 95571.890625,
+      "learning_rate": 1.4197086284917227e-09,
+      "loss": 2.2956,
+      "step": 133500
+    },
+    {
+      "epoch": 1.0013376002032555,
+      "grad_norm": 99771.578125,
+      "learning_rate": 4.9894944614217454e-05,
+      "loss": 2.211,
+      "step": 134000
+    },
+    {
+      "epoch": 1.00507391362017,
+      "grad_norm": 105597.21875,
+      "learning_rate": 4.960149774176924e-05,
+      "loss": 2.2131,
+      "step": 134500
+    },
+    {
+      "epoch": 1.0088102270370847,
+      "grad_norm": 87634.5,
+      "learning_rate": 4.930806459585094e-05,
+      "loss": 2.2453,
+      "step": 135000
+    },
+    {
+      "epoch": 1.0125465404539995,
+      "grad_norm": 86854.4375,
+      "learning_rate": 4.901465528385536e-05,
+      "loss": 2.1825,
+      "step": 135500
+    },
+    {
+      "epoch": 1.016282853870914,
+      "grad_norm": 75373.640625,
+      "learning_rate": 4.872127991235435e-05,
+      "loss": 2.1953,
+      "step": 136000
+    },
+    {
+      "epoch": 1.0200191672878287,
+      "grad_norm": 89905.53125,
+      "learning_rate": 4.842794858675067e-05,
+      "loss": 2.223,
+      "step": 136500
+    },
+    {
+      "epoch": 1.0237554807047435,
+      "grad_norm": 74529.2734375,
+      "learning_rate": 4.8134671410929855e-05,
+      "loss": 2.2232,
+      "step": 137000
+    },
+    {
+      "epoch": 1.0274917941216581,
+      "grad_norm": 86540.96875,
+      "learning_rate": 4.7841458486912313e-05,
+      "loss": 2.229,
+      "step": 137500
+    },
+    {
+      "epoch": 1.0312281075385727,
+      "grad_norm": 79359.9140625,
+      "learning_rate": 4.75483199145052e-05,
+      "loss": 2.2484,
+      "step": 138000
+    },
+    {
+      "epoch": 1.0349644209554874,
+      "grad_norm": 88735.703125,
+      "learning_rate": 4.725526579095468e-05,
+      "loss": 2.2141,
+      "step": 138500
+    },
+    {
+      "epoch": 1.0387007343724022,
+      "grad_norm": 84116.59375,
+      "learning_rate": 4.6962306210598024e-05,
+      "loss": 2.2362,
+      "step": 139000
+    },
+    {
+      "epoch": 1.0424370477893168,
+      "grad_norm": 85959.921875,
+      "learning_rate": 4.66694512645159e-05,
+      "loss": 2.2703,
+      "step": 139500
+    },
+    {
+      "epoch": 1.0461733612062314,
+      "grad_norm": 76734.796875,
+      "learning_rate": 4.637671104018486e-05,
+      "loss": 2.2324,
+      "step": 140000
+    },
+    {
+      "epoch": 1.049909674623146,
+      "grad_norm": 82724.0078125,
+      "learning_rate": 4.608409562112983e-05,
+      "loss": 2.2406,
+      "step": 140500
+    },
+    {
+      "epoch": 1.0536459880400608,
+      "grad_norm": 85404.375,
+      "learning_rate": 4.5791615086576726e-05,
+      "loss": 2.2049,
+      "step": 141000
+    },
+    {
+      "epoch": 1.0573823014569754,
+      "grad_norm": 73961.2890625,
+      "learning_rate": 4.549927951110542e-05,
+      "loss": 2.2176,
+      "step": 141500
+    },
+    {
+      "epoch": 1.06111861487389,
+      "grad_norm": 83664.9921875,
+      "learning_rate": 4.5207098964302556e-05,
+      "loss": 2.2473,
+      "step": 142000
+    },
+    {
+      "epoch": 1.0648549282908046,
+      "grad_norm": 89258.0390625,
+      "learning_rate": 4.491508351041477e-05,
+      "loss": 2.227,
+      "step": 142500
+    },
+    {
+      "epoch": 1.0685912417077195,
+      "grad_norm": 99317.4296875,
+      "learning_rate": 4.462324320800205e-05,
+      "loss": 2.2237,
+      "step": 143000
+    },
+    {
+      "epoch": 1.072327555124634,
+      "grad_norm": 90441.765625,
+      "learning_rate": 4.433158810959122e-05,
+      "loss": 2.2321,
+      "step": 143500
+    },
+    {
+      "epoch": 1.0760638685415487,
+      "grad_norm": 69907.328125,
+      "learning_rate": 4.4040128261329724e-05,
+      "loss": 2.2319,
+      "step": 144000
+    },
+    {
+      "epoch": 1.0798001819584635,
+      "grad_norm": 87409.109375,
+      "learning_rate": 4.3748873702639536e-05,
+      "loss": 2.2581,
+      "step": 144500
+    },
+    {
+      "epoch": 1.083536495375378,
+      "grad_norm": 91680.8828125,
+      "learning_rate": 4.345783446587135e-05,
+      "loss": 2.2381,
+      "step": 145000
+    },
+    {
+      "epoch": 1.0872728087922927,
+      "grad_norm": 84339.4140625,
+      "learning_rate": 4.316702057595907e-05,
+      "loss": 2.2398,
+      "step": 145500
+    },
+    {
+      "epoch": 1.0910091222092073,
+      "grad_norm": 99060.578125,
+      "learning_rate": 4.2876442050074436e-05,
+      "loss": 2.2251,
+      "step": 146000
+    },
+    {
+      "epoch": 1.0947454356261221,
+      "grad_norm": 103543.2421875,
+      "learning_rate": 4.258610889728198e-05,
+      "loss": 2.2536,
+      "step": 146500
+    },
+    {
+      "epoch": 1.0984817490430367,
+      "grad_norm": 89170.1953125,
+      "learning_rate": 4.229603111819437e-05,
+      "loss": 2.1822,
+      "step": 147000
+    },
+    {
+      "epoch": 1.1022180624599514,
+      "grad_norm": 95436.765625,
+      "learning_rate": 4.2006218704627756e-05,
+      "loss": 2.2338,
+      "step": 147500
+    },
+    {
+      "epoch": 1.105954375876866,
+      "grad_norm": 86420.7265625,
+      "learning_rate": 4.1716681639257725e-05,
+      "loss": 2.2153,
+      "step": 148000
+    },
+    {
+      "epoch": 1.1096906892937808,
+      "grad_norm": 76673.3359375,
+      "learning_rate": 4.1427429895275446e-05,
+      "loss": 2.2289,
+      "step": 148500
+    },
+    {
+      "epoch": 1.1134270027106954,
+      "grad_norm": 101906.8515625,
+      "learning_rate": 4.113847343604404e-05,
+      "loss": 2.2297,
+      "step": 149000
+    },
+    {
+      "epoch": 1.11716331612761,
+      "grad_norm": 88695.703125,
+      "learning_rate": 4.084982221475554e-05,
+      "loss": 2.2219,
+      "step": 149500
+    },
+    {
+      "epoch": 1.1208996295445246,
+      "grad_norm": 100434.46875,
+      "learning_rate": 4.05614861740879e-05,
+      "loss": 2.2344,
+      "step": 150000
+    },
+    {
+      "epoch": 1.1246359429614394,
+      "grad_norm": 90131.671875,
+      "learning_rate": 4.0273475245862594e-05,
+      "loss": 2.2137,
+      "step": 150500
+    },
+    {
+      "epoch": 1.128372256378354,
+      "grad_norm": 84749.8828125,
+      "learning_rate": 3.998579935070253e-05,
+      "loss": 2.2358,
+      "step": 151000
+    },
+    {
+      "epoch": 1.1321085697952686,
+      "grad_norm": 80951.0546875,
+      "learning_rate": 3.969846839769027e-05,
+      "loss": 2.2327,
+      "step": 151500
+    },
+    {
+      "epoch": 1.1358448832121835,
+      "grad_norm": 82394.359375,
+      "learning_rate": 3.941149228402674e-05,
+      "loss": 2.2448,
+      "step": 152000
+    },
+    {
+      "epoch": 1.139581196629098,
+      "grad_norm": 108267.5546875,
+      "learning_rate": 3.912488089469036e-05,
+      "loss": 2.211,
+      "step": 152500
+    },
+    {
+      "epoch": 1.1433175100460127,
+      "grad_norm": 73743.703125,
+      "learning_rate": 3.883864410209648e-05,
+      "loss": 2.2159,
+      "step": 153000
+    },
+    {
+      "epoch": 1.1470538234629273,
+      "grad_norm": 105991.84375,
+      "learning_rate": 3.855279176575732e-05,
+      "loss": 2.2191,
+      "step": 153500
+    },
+    {
+      "epoch": 1.150790136879842,
+      "grad_norm": 88528.921875,
+      "learning_rate": 3.826733373194247e-05,
+      "loss": 2.2397,
+      "step": 154000
+    },
+    {
+      "epoch": 1.1545264502967567,
+      "grad_norm": 99099.5390625,
+      "learning_rate": 3.798227983333955e-05,
+      "loss": 2.2083,
+      "step": 154500
+    },
+    {
+      "epoch": 1.1582627637136713,
+      "grad_norm": 96959.421875,
+      "learning_rate": 3.769763988871572e-05,
+      "loss": 2.2146,
+      "step": 155000
+    },
+    {
+      "epoch": 1.1619990771305861,
+      "grad_norm": 96111.9375,
+      "learning_rate": 3.741342370257932e-05,
+      "loss": 2.2421,
+      "step": 155500
+    },
+    {
+      "epoch": 1.1657353905475007,
+      "grad_norm": 80256.734375,
+      "learning_rate": 3.712964106484218e-05,
+      "loss": 2.2199,
+      "step": 156000
+    },
+    {
+      "epoch": 1.1694717039644154,
+      "grad_norm": 84992.546875,
+      "learning_rate": 3.684630175048249e-05,
+      "loss": 2.184,
+      "step": 156500
+    },
+    {
+      "epoch": 1.17320801738133,
+      "grad_norm": 94191.3671875,
+      "learning_rate": 3.6563415519207985e-05,
+      "loss": 2.2505,
+      "step": 157000
+    },
+    {
+      "epoch": 1.1769443307982446,
+      "grad_norm": 76174.2421875,
+      "learning_rate": 3.62809921151198e-05,
+      "loss": 2.2143,
+      "step": 157500
+    },
+    {
+      "epoch": 1.1806806442151594,
+      "grad_norm": 89977.34375,
+      "learning_rate": 3.599904126637692e-05,
+      "loss": 2.2338,
+      "step": 158000
+    },
+    {
+      "epoch": 1.184416957632074,
+      "grad_norm": 107886.5390625,
+      "learning_rate": 3.571757268486096e-05,
+      "loss": 2.2516,
+      "step": 158500
+    },
+    {
+      "epoch": 1.1881532710489886,
+      "grad_norm": 60902.64453125,
+      "learning_rate": 3.54365960658417e-05,
+      "loss": 2.197,
+      "step": 159000
+    },
+    {
+      "epoch": 1.1918895844659034,
+      "grad_norm": 91395.515625,
+      "learning_rate": 3.5156121087643154e-05,
+      "loss": 2.2238,
+      "step": 159500
+    },
+    {
+      "epoch": 1.195625897882818,
+      "grad_norm": 90301.53125,
+      "learning_rate": 3.487615741131013e-05,
+      "loss": 2.2317,
+      "step": 160000
+    },
+    {
+      "epoch": 1.1993622112997326,
+      "grad_norm": 91005.0625,
+      "learning_rate": 3.459671468027553e-05,
+      "loss": 2.2019,
+      "step": 160500
+    },
+    {
+      "epoch": 1.2030985247166472,
+      "grad_norm": 81654.03125,
+      "learning_rate": 3.4317802520028117e-05,
+      "loss": 2.2199,
+      "step": 161000
+    },
+    {
+      "epoch": 1.206834838133562,
+      "grad_norm": 100322.15625,
+      "learning_rate": 3.403943053778097e-05,
+      "loss": 2.2668,
+      "step": 161500
+    },
+    {
+      "epoch": 1.2105711515504767,
+      "grad_norm": 84712.84375,
+      "learning_rate": 3.37616083221406e-05,
+      "loss": 2.215,
+      "step": 162000
+    },
+    {
+      "epoch": 1.2143074649673913,
+      "grad_norm": 93723.7734375,
+      "learning_rate": 3.348434544277663e-05,
+      "loss": 2.2052,
+      "step": 162500
+    },
+    {
+      "epoch": 1.218043778384306,
+      "grad_norm": 98371.0859375,
+      "learning_rate": 3.320765145009216e-05,
+      "loss": 2.1767,
+      "step": 163000
+    },
+    {
+      "epoch": 1.2217800918012207,
+      "grad_norm": 78824.3359375,
+      "learning_rate": 3.293153587489488e-05,
+      "loss": 2.2341,
+      "step": 163500
+    },
+    {
+      "epoch": 1.2255164052181353,
+      "grad_norm": 93536.6640625,
+      "learning_rate": 3.265600822806866e-05,
+      "loss": 2.1954,
+      "step": 164000
+    },
+    {
+      "epoch": 1.22925271863505,
+      "grad_norm": 89381.3203125,
+      "learning_rate": 3.2381078000245965e-05,
+      "loss": 2.2026,
+      "step": 164500
+    },
+    {
+      "epoch": 1.2329890320519645,
+      "grad_norm": 110292.1484375,
+      "learning_rate": 3.2106754661481086e-05,
+      "loss": 2.2022,
+      "step": 165000
+    },
+    {
+      "epoch": 1.2367253454688794,
+      "grad_norm": 89096.8359375,
+      "learning_rate": 3.183304766092374e-05,
+      "loss": 2.1618,
+      "step": 165500
+    },
+    {
+      "epoch": 1.240461658885794,
+      "grad_norm": 97123.8671875,
+      "learning_rate": 3.1559966426493804e-05,
+      "loss": 2.192,
+      "step": 166000
+    },
+    {
+      "epoch": 1.2441979723027086,
+      "grad_norm": 86452.890625,
+      "learning_rate": 3.128752036455631e-05,
+      "loss": 2.203,
+      "step": 166500
+    },
+    {
+      "epoch": 1.2479342857196234,
+      "grad_norm": 99330.265625,
+      "learning_rate": 3.101571885959766e-05,
+      "loss": 2.1653,
+      "step": 167000
+    },
+    {
+      "epoch": 1.251670599136538,
+      "grad_norm": 82720.1015625,
+      "learning_rate": 3.0744571273902345e-05,
+      "loss": 2.1893,
+      "step": 167500
+    },
+    {
+      "epoch": 1.2554069125534526,
+      "grad_norm": 98328.6328125,
+      "learning_rate": 3.047408694723032e-05,
+      "loss": 2.2091,
+      "step": 168000
+    },
+    {
+      "epoch": 1.2591432259703672,
+      "grad_norm": 89367.40625,
+      "learning_rate": 3.0204275196495414e-05,
+      "loss": 2.1961,
+      "step": 168500
+    },
+    {
+      "epoch": 1.262879539387282,
+      "grad_norm": 95675.3203125,
+      "learning_rate": 2.99351453154444e-05,
+      "loss": 2.1824,
+      "step": 169000
+    },
+    {
+      "epoch": 1.2666158528041966,
+      "grad_norm": 83745.3359375,
+      "learning_rate": 2.9666706574336834e-05,
+      "loss": 2.1793,
+      "step": 169500
+    },
+    {
+      "epoch": 1.2703521662211112,
+      "grad_norm": 92235.09375,
+      "learning_rate": 2.9398968219625706e-05,
+      "loss": 2.2034,
+      "step": 170000
+    },
+    {
+      "epoch": 1.274088479638026,
+      "grad_norm": 79515.6328125,
+      "learning_rate": 2.913193947363908e-05,
+      "loss": 2.2091,
+      "step": 170500
+    },
+    {
+      "epoch": 1.2778247930549407,
+      "grad_norm": 92163.1640625,
+      "learning_rate": 2.8865629534262216e-05,
+      "loss": 2.2082,
+      "step": 171000
+    },
+    {
+      "epoch": 1.2815611064718553,
+      "grad_norm": 90488.1484375,
+      "learning_rate": 2.8600047574620998e-05,
+      "loss": 2.1976,
+      "step": 171500
+    },
+    {
+      "epoch": 1.2852974198887699,
+      "grad_norm": 93587.21875,
+      "learning_rate": 2.83352027427657e-05,
+      "loss": 2.2092,
+      "step": 172000
+    },
+    {
+      "epoch": 1.2890337333056845,
+      "grad_norm": 90367.0546875,
+      "learning_rate": 2.8071104161356087e-05,
+      "loss": 2.2086,
+      "step": 172500
+    },
+    {
+      "epoch": 1.2927700467225993,
+      "grad_norm": 94576.171875,
+      "learning_rate": 2.7807760927347102e-05,
+      "loss": 2.2071,
+      "step": 173000
+    },
+    {
+      "epoch": 1.296506360139514,
+      "grad_norm": 84362.703125,
+      "learning_rate": 2.7545182111675466e-05,
+      "loss": 2.1646,
+      "step": 173500
+    },
+    {
+      "epoch": 1.3002426735564285,
+      "grad_norm": 99929.3984375,
+      "learning_rate": 2.7283376758947333e-05,
+      "loss": 2.1721,
+      "step": 174000
+    },
+    {
+      "epoch": 1.3039789869733434,
+      "grad_norm": 99786.953125,
+      "learning_rate": 2.7022353887126668e-05,
+      "loss": 2.2208,
+      "step": 174500
+    },
+    {
+      "epoch": 1.307715300390258,
+      "grad_norm": 118692.8046875,
+      "learning_rate": 2.6762122487224684e-05,
+      "loss": 2.2036,
+      "step": 175000
+    },
+    {
+      "epoch": 1.3114516138071726,
+      "grad_norm": 88971.234375,
+      "learning_rate": 2.6502691522990032e-05,
+      "loss": 2.2018,
+      "step": 175500
+    },
+    {
+      "epoch": 1.3151879272240872,
+      "grad_norm": 106580.484375,
+      "learning_rate": 2.624406993060019e-05,
+      "loss": 2.1668,
+      "step": 176000
+    },
+    {
+      "epoch": 1.318924240641002,
+      "grad_norm": 94986.0390625,
+      "learning_rate": 2.598626661835354e-05,
+      "loss": 2.1782,
+      "step": 176500
+    },
+    {
+      "epoch": 1.3226605540579166,
+      "grad_norm": 87692.828125,
+      "learning_rate": 2.5729290466362565e-05,
+      "loss": 2.251,
+      "step": 177000
+    },
+    {
+      "epoch": 1.3263968674748312,
+      "grad_norm": 93603.875,
+      "learning_rate": 2.5473150326248006e-05,
+      "loss": 2.2143,
+      "step": 177500
+    },
+    {
+      "epoch": 1.330133180891746,
+      "grad_norm": 98669.8828125,
+      "learning_rate": 2.521785502083384e-05,
+      "loss": 2.1962,
+      "step": 178000
+    },
+    {
+      "epoch": 1.3338694943086606,
+      "grad_norm": 77537.9609375,
+      "learning_rate": 2.4963413343843522e-05,
+      "loss": 2.1533,
+      "step": 178500
+    },
+    {
+      "epoch": 1.3376058077255752,
+      "grad_norm": 107931.5625,
+      "learning_rate": 2.4709834059597026e-05,
+      "loss": 2.1715,
+      "step": 179000
+    },
+    {
+      "epoch": 1.3413421211424899,
+      "grad_norm": 91190.8125,
+      "learning_rate": 2.4457125902708872e-05,
+      "loss": 2.1757,
+      "step": 179500
+    },
+    {
+      "epoch": 1.3450784345594045,
+      "grad_norm": 84562.8046875,
+      "learning_rate": 2.420529757778746e-05,
+      "loss": 2.1666,
+      "step": 180000
+    },
+    {
+      "epoch": 1.3488147479763193,
+      "grad_norm": 99736.1640625,
+      "learning_rate": 2.3954357759134995e-05,
+      "loss": 2.1948,
+      "step": 180500
+    },
+    {
+      "epoch": 1.3525510613932339,
+      "grad_norm": 93177.078125,
+      "learning_rate": 2.3704315090448886e-05,
+      "loss": 2.1868,
+      "step": 181000
+    },
+    {
+      "epoch": 1.3562873748101485,
+      "grad_norm": 77689.5625,
+      "learning_rate": 2.345517818452394e-05,
+      "loss": 2.2091,
+      "step": 181500
+    },
+    {
+      "epoch": 1.3600236882270633,
+      "grad_norm": 92219.8671875,
+      "learning_rate": 2.3206955622955623e-05,
+      "loss": 2.1795,
+      "step": 182000
+    },
+    {
+      "epoch": 1.363760001643978,
+      "grad_norm": 96159.34375,
+      "learning_rate": 2.2959655955844656e-05,
+      "loss": 2.2186,
+      "step": 182500
+    },
+    {
+      "epoch": 1.3674963150608925,
+      "grad_norm": 84869.453125,
+      "learning_rate": 2.271328770150227e-05,
+      "loss": 2.1516,
+      "step": 183000
+    },
+    {
+      "epoch": 1.3712326284778071,
+      "grad_norm": 79080.1484375,
+      "learning_rate": 2.2467859346156968e-05,
+      "loss": 2.2031,
+      "step": 183500
+    },
+    {
+      "epoch": 1.374968941894722,
+      "grad_norm": 107557.9765625,
+      "learning_rate": 2.222337934366214e-05,
+      "loss": 2.1703,
+      "step": 184000
+    },
+    {
+      "epoch": 1.3787052553116366,
+      "grad_norm": 104010.9296875,
+      "learning_rate": 2.1979856115204845e-05,
+      "loss": 2.1637,
+      "step": 184500
+    },
+    {
+      "epoch": 1.3824415687285512,
+      "grad_norm": 93606.03125,
+      "learning_rate": 2.173729804901581e-05,
+      "loss": 2.1746,
+      "step": 185000
+    },
+    {
+      "epoch": 1.386177882145466,
+      "grad_norm": 100350.8515625,
+      "learning_rate": 2.1495713500080443e-05,
+      "loss": 2.1847,
+      "step": 185500
+    },
+    {
+      "epoch": 1.3899141955623806,
+      "grad_norm": 91779.265625,
+      "learning_rate": 2.1255110789851084e-05,
+      "loss": 2.1576,
+      "step": 186000
+    },
+    {
+      "epoch": 1.3936505089792952,
+      "grad_norm": 112973.15625,
+      "learning_rate": 2.1015498205960283e-05,
+      "loss": 2.177,
+      "step": 186500
+    },
+    {
+      "epoch": 1.3973868223962098,
+      "grad_norm": 91850.71875,
+      "learning_rate": 2.0776884001935437e-05,
+      "loss": 2.1904,
+      "step": 187000
+    },
+    {
+      "epoch": 1.4011231358131244,
+      "grad_norm": 92866.7265625,
+      "learning_rate": 2.0539276396914448e-05,
+      "loss": 2.1891,
+      "step": 187500
+    },
+    {
+      "epoch": 1.4048594492300392,
+      "grad_norm": 94535.1640625,
+      "learning_rate": 2.0302683575362608e-05,
+      "loss": 2.1768,
+      "step": 188000
+    },
+    {
+      "epoch": 1.4085957626469539,
+      "grad_norm": 99327.2890625,
+      "learning_rate": 2.0067113686790694e-05,
+      "loss": 2.195,
+      "step": 188500
+    },
+    {
+      "epoch": 1.4123320760638687,
+      "grad_norm": 95519.703125,
+      "learning_rate": 1.983257484547421e-05,
+      "loss": 2.1735,
+      "step": 189000
+    },
+    {
+      "epoch": 1.4160683894807833,
+      "grad_norm": 93955.65625,
+      "learning_rate": 1.9599075130173977e-05,
+      "loss": 2.2019,
+      "step": 189500
+    },
+    {
+      "epoch": 1.419804702897698,
+      "grad_norm": 143901.203125,
+      "learning_rate": 1.9366622583857803e-05,
+      "loss": 2.1768,
+      "step": 190000
+    },
+    {
+      "epoch": 1.4235410163146125,
+      "grad_norm": 77758.40625,
+      "learning_rate": 1.913522521342338e-05,
+      "loss": 2.1481,
+      "step": 190500
+    },
+    {
+      "epoch": 1.427277329731527,
+      "grad_norm": 102854.71875,
+      "learning_rate": 1.890489098942267e-05,
+      "loss": 2.1364,
+      "step": 191000
+    },
+    {
+      "epoch": 1.431013643148442,
+      "grad_norm": 76422.46875,
+      "learning_rate": 1.8675627845787126e-05,
+      "loss": 2.1446,
+      "step": 191500
+    },
+    {
+      "epoch": 1.4347499565653565,
+      "grad_norm": 103994.8359375,
+      "learning_rate": 1.8447443679554567e-05,
+      "loss": 2.1618,
+      "step": 192000
+    },
+    {
+      "epoch": 1.4384862699822711,
+      "grad_norm": 93646.765625,
+      "learning_rate": 1.8220346350597123e-05,
+      "loss": 2.2005,
+      "step": 192500
+    },
+    {
+      "epoch": 1.442222583399186,
+      "grad_norm": 104911.5234375,
+      "learning_rate": 1.7994343681350422e-05,
+      "loss": 2.185,
+      "step": 193000
+    },
+    {
+      "epoch": 1.4459588968161006,
+      "grad_norm": 103320.125,
+      "learning_rate": 1.7769443456544322e-05,
+      "loss": 2.1904,
+      "step": 193500
+    },
+    {
+      "epoch": 1.4496952102330152,
+      "grad_norm": 88957.3203125,
+      "learning_rate": 1.7545653422934537e-05,
+      "loss": 2.1924,
+      "step": 194000
+    },
+    {
+      "epoch": 1.4534315236499298,
+      "grad_norm": 98682.171875,
+      "learning_rate": 1.732298128903597e-05,
+      "loss": 2.1546,
+      "step": 194500
+    },
+    {
+      "epoch": 1.4571678370668444,
+      "grad_norm": 93626.0234375,
+      "learning_rate": 1.7101434724857147e-05,
+      "loss": 2.1653,
+      "step": 195000
+    },
+    {
+      "epoch": 1.4609041504837592,
+      "grad_norm": 84740.2265625,
+      "learning_rate": 1.6881021361635947e-05,
+      "loss": 2.1835,
+      "step": 195500
+    },
+    {
+      "epoch": 1.4646404639006738,
+      "grad_norm": 95559.859375,
+      "learning_rate": 1.6661748791576855e-05,
+      "loss": 2.1716,
+      "step": 196000
+    },
+    {
+      "epoch": 1.4683767773175886,
+      "grad_norm": 105746.453125,
+      "learning_rate": 1.6443624567589377e-05,
+      "loss": 2.1467,
+      "step": 196500
+    },
+    {
+      "epoch": 1.4721130907345032,
+      "grad_norm": 93445.3515625,
+      "learning_rate": 1.622665620302789e-05,
+      "loss": 2.1673,
+      "step": 197000
+    },
+    {
+      "epoch": 1.4758494041514179,
+      "grad_norm": 88159.8125,
+      "learning_rate": 1.6010851171432874e-05,
+      "loss": 2.1727,
+      "step": 197500
+    },
+    {
+      "epoch": 1.4795857175683325,
+      "grad_norm": 93973.4375,
+      "learning_rate": 1.5796216906273388e-05,
+      "loss": 2.1824,
+      "step": 198000
+    },
+    {
+      "epoch": 1.483322030985247,
+      "grad_norm": 95876.78125,
+      "learning_rate": 1.5582760800691155e-05,
+      "loss": 2.1858,
+      "step": 198500
+    },
+    {
+      "epoch": 1.487058344402162,
+      "grad_norm": 90519.78125,
+      "learning_rate": 1.537049020724581e-05,
+      "loss": 2.1984,
+      "step": 199000
+    },
+    {
+      "epoch": 1.4907946578190765,
+      "grad_norm": 88962.28125,
+      "learning_rate": 1.515941243766169e-05,
+      "loss": 2.2229,
+      "step": 199500
+    },
+    {
+      "epoch": 1.494530971235991,
+      "grad_norm": 113852.828125,
+      "learning_rate": 1.4949534762575912e-05,
+      "loss": 2.157,
+      "step": 200000
+    },
+    {
+      "epoch": 1.498267284652906,
+      "grad_norm": 97744.3984375,
+      "learning_rate": 1.4740864411288013e-05,
+      "loss": 2.1711,
+      "step": 200500
+    },
+    {
+      "epoch": 1.5020035980698205,
+      "grad_norm": 88245.1640625,
+      "learning_rate": 1.4533408571510904e-05,
+      "loss": 2.1867,
+      "step": 201000
+    },
+    {
+      "epoch": 1.5057399114867351,
+      "grad_norm": 89220.171875,
+      "learning_rate": 1.4327174389123227e-05,
+      "loss": 2.1382,
+      "step": 201500
+    },
+    {
+      "epoch": 1.5094762249036497,
+      "grad_norm": 108794.328125,
+      "learning_rate": 1.4122168967923366e-05,
+      "loss": 2.1608,
+      "step": 202000
+    },
+    {
+      "epoch": 1.5132125383205643,
+      "grad_norm": 108413.6640625,
+      "learning_rate": 1.3918399369384556e-05,
+      "loss": 2.1657,
+      "step": 202500
+    },
+    {
+      "epoch": 1.5169488517374792,
+      "grad_norm": 92711.84375,
+      "learning_rate": 1.3715872612411806e-05,
+      "loss": 2.1385,
+      "step": 203000
+    },
+    {
+      "epoch": 1.5206851651543938,
+      "grad_norm": 97720.765625,
+      "learning_rate": 1.3514595673100084e-05,
+      "loss": 2.1467,
+      "step": 203500
+    },
+    {
+      "epoch": 1.5244214785713086,
+      "grad_norm": 86545.078125,
+      "learning_rate": 1.3314575484493946e-05,
+      "loss": 2.152,
+      "step": 204000
+    },
+    {
+      "epoch": 1.5281577919882232,
+      "grad_norm": 96816.7421875,
+      "learning_rate": 1.3115818936348889e-05,
+      "loss": 2.1636,
+      "step": 204500
+    },
+    {
+      "epoch": 1.5318941054051378,
+      "grad_norm": 100713.7734375,
+      "learning_rate": 1.2918332874893869e-05,
+      "loss": 2.1784,
+      "step": 205000
+    },
+    {
+      "epoch": 1.5356304188220524,
+      "grad_norm": 97874.359375,
+      "learning_rate": 1.2722124102595518e-05,
+      "loss": 2.1903,
+      "step": 205500
+    },
+    {
+      "epoch": 1.539366732238967,
+      "grad_norm": 91848.2734375,
+      "learning_rate": 1.252719937792397e-05,
+      "loss": 2.1524,
+      "step": 206000
+    },
+    {
+      "epoch": 1.5431030456558816,
+      "grad_norm": 95539.515625,
+      "learning_rate": 1.2333565415119835e-05,
+      "loss": 2.1813,
+      "step": 206500
+    },
+    {
+      "epoch": 1.5468393590727965,
+      "grad_norm": 92111.3984375,
+      "learning_rate": 1.2141228883963118e-05,
+      "loss": 2.1615,
+      "step": 207000
+    },
+    {
+      "epoch": 1.5505756724897113,
+      "grad_norm": 90152.640625,
+      "learning_rate": 1.195019640954339e-05,
+      "loss": 2.1615,
+      "step": 207500
+    },
+    {
+      "epoch": 1.554311985906626,
+      "grad_norm": 96634.609375,
+      "learning_rate": 1.1760474572031538e-05,
+      "loss": 2.1455,
+      "step": 208000
+    },
+    {
+      "epoch": 1.5580482993235405,
+      "grad_norm": 102015.8984375,
+      "learning_rate": 1.1572069906453265e-05,
+      "loss": 2.1508,
+      "step": 208500
+    },
+    {
+      "epoch": 1.561784612740455,
+      "grad_norm": 90662.0859375,
+      "learning_rate": 1.1384988902463806e-05,
+      "loss": 2.1443,
+      "step": 209000
+    },
+    {
+      "epoch": 1.5655209261573697,
+      "grad_norm": 101081.84375,
+      "learning_rate": 1.1199238004124519e-05,
+      "loss": 2.191,
+      "step": 209500
+    },
+    {
+      "epoch": 1.5692572395742843,
+      "grad_norm": 86159.140625,
+      "learning_rate": 1.101482360968088e-05,
+      "loss": 2.1739,
+      "step": 210000
+    },
+    {
+      "epoch": 1.5729935529911991,
+      "grad_norm": 96140.9296875,
+      "learning_rate": 1.0831752071342033e-05,
+      "loss": 2.1597,
+      "step": 210500
+    },
+    {
+      "epoch": 1.5767298664081137,
+      "grad_norm": 93931.6484375,
+      "learning_rate": 1.0650029695062103e-05,
+      "loss": 2.1704,
+      "step": 211000
+    },
+    {
+      "epoch": 1.5804661798250286,
+      "grad_norm": 96550.375,
+      "learning_rate": 1.0469662740322888e-05,
+      "loss": 2.1762,
+      "step": 211500
+    },
+    {
+      "epoch": 1.5842024932419432,
+      "grad_norm": 91985.375,
+      "learning_rate": 1.0290657419918308e-05,
+      "loss": 2.1537,
+      "step": 212000
+    },
+    {
+      "epoch": 1.5879388066588578,
+      "grad_norm": 99446.453125,
+      "learning_rate": 1.0113019899740345e-05,
+      "loss": 2.1567,
+      "step": 212500
+    },
+    {
+      "epoch": 1.5916751200757724,
+      "grad_norm": 90499.4453125,
+      "learning_rate": 9.936756298566724e-06,
+      "loss": 2.1542,
+      "step": 213000
+    },
+    {
+      "epoch": 1.595411433492687,
+      "grad_norm": 91874.203125,
+      "learning_rate": 9.761872687850093e-06,
+      "loss": 2.1701,
+      "step": 213500
+    },
+    {
+      "epoch": 1.5991477469096018,
+      "grad_norm": 101494.265625,
+      "learning_rate": 9.588375091508945e-06,
+      "loss": 2.1785,
+      "step": 214000
+    },
+    {
+      "epoch": 1.6028840603265164,
+      "grad_norm": 101492.890625,
+      "learning_rate": 9.41626948572007e-06,
+      "loss": 2.1681,
+      "step": 214500
+    },
+    {
+      "epoch": 1.6066203737434313,
+      "grad_norm": 97085.359375,
+      "learning_rate": 9.245561798712704e-06,
+      "loss": 2.163,
+      "step": 215000
+    },
+    {
+      "epoch": 1.6103566871603459,
+      "grad_norm": 101823.4609375,
+      "learning_rate": 9.076257910564402e-06,
+      "loss": 2.1538,
+      "step": 215500
+    },
+    {
+      "epoch": 1.6140930005772605,
+      "grad_norm": 91494.078125,
+      "learning_rate": 8.908363652998414e-06,
+      "loss": 2.1235,
+      "step": 216000
+    },
+    {
+      "epoch": 1.617829313994175,
+      "grad_norm": 85581.09375,
+      "learning_rate": 8.74188480918282e-06,
+      "loss": 2.1736,
+      "step": 216500
+    },
+    {
+      "epoch": 1.6215656274110897,
+      "grad_norm": 92024.046875,
+      "learning_rate": 8.576827113531427e-06,
+      "loss": 2.177,
+      "step": 217000
+    },
+    {
+      "epoch": 1.6253019408280043,
+      "grad_norm": 74753.4296875,
+      "learning_rate": 8.413196251506073e-06,
+      "loss": 2.143,
+      "step": 217500
+    },
+    {
+      "epoch": 1.629038254244919,
+      "grad_norm": 107585.3125,
+      "learning_rate": 8.25099785942094e-06,
+      "loss": 2.1522,
+      "step": 218000
+    },
+    {
+      "epoch": 1.6327745676618337,
+      "grad_norm": 94180.15625,
+      "learning_rate": 8.090237524248346e-06,
+      "loss": 2.1343,
+      "step": 218500
+    },
+    {
+      "epoch": 1.6365108810787485,
+      "grad_norm": 102109.1953125,
+      "learning_rate": 7.930920783426265e-06,
+      "loss": 2.1643,
+      "step": 219000
+    },
+    {
+      "epoch": 1.6402471944956631,
+      "grad_norm": 107961.2109375,
+      "learning_rate": 7.773053124667712e-06,
+      "loss": 2.1483,
+      "step": 219500
+    },
+    {
+      "epoch": 1.6439835079125777,
+      "grad_norm": 89340.9296875,
+      "learning_rate": 7.616639985771556e-06,
+      "loss": 2.1602,
+      "step": 220000
+    },
+    {
+      "epoch": 1.6477198213294924,
+      "grad_norm": 103068.6015625,
+      "learning_rate": 7.461686754435332e-06,
+      "loss": 2.1549,
+      "step": 220500
+    },
+    {
+      "epoch": 1.651456134746407,
+      "grad_norm": 101308.2421875,
+      "learning_rate": 7.30819876806963e-06,
+      "loss": 2.1645,
+      "step": 221000
+    },
+    {
+      "epoch": 1.6551924481633218,
+      "grad_norm": 116589.4140625,
+      "learning_rate": 7.156181313614185e-06,
+      "loss": 2.1538,
+      "step": 221500
+    },
+    {
+      "epoch": 1.6589287615802364,
+      "grad_norm": 107515.390625,
+      "learning_rate": 7.005639627355881e-06,
+      "loss": 2.1506,
+      "step": 222000
+    },
+    {
+      "epoch": 1.6626650749971512,
+      "grad_norm": 105536.6484375,
+      "learning_rate": 6.856578894748284e-06,
+      "loss": 2.1563,
+      "step": 222500
+    },
+    {
+      "epoch": 1.6664013884140658,
+      "grad_norm": 125030.3046875,
+      "learning_rate": 6.709004250233097e-06,
+      "loss": 2.1574,
+      "step": 223000
+    },
+    {
+      "epoch": 1.6701377018309804,
+      "grad_norm": 95777.0546875,
+      "learning_rate": 6.562920777063231e-06,
+      "loss": 2.1502,
+      "step": 223500
+    },
+    {
+      "epoch": 1.673874015247895,
+      "grad_norm": 125466.265625,
+      "learning_rate": 6.418333507127788e-06,
+      "loss": 2.1569,
+      "step": 224000
+    },
+    {
+      "epoch": 1.6776103286648096,
+      "grad_norm": 82233.953125,
+      "learning_rate": 6.275247420778685e-06,
+      "loss": 2.1315,
+      "step": 224500
+    },
+    {
+      "epoch": 1.6813466420817242,
+      "grad_norm": 86722.7890625,
+      "learning_rate": 6.133667446659131e-06,
+      "loss": 2.1662,
+      "step": 225000
+    },
+    {
+      "epoch": 1.685082955498639,
+      "grad_norm": 95386.34375,
+      "learning_rate": 5.993598461533845e-06,
+      "loss": 2.147,
+      "step": 225500
+    },
+    {
+      "epoch": 1.6888192689155537,
+      "grad_norm": 82564.6484375,
+      "learning_rate": 5.855045290121047e-06,
+      "loss": 2.1426,
+      "step": 226000
+    },
+    {
+      "epoch": 1.6925555823324685,
+      "grad_norm": 86084.6875,
+      "learning_rate": 5.71801270492634e-06,
+      "loss": 2.1517,
+      "step": 226500
+    },
+    {
+      "epoch": 1.696291895749383,
+      "grad_norm": 90130.1796875,
+      "learning_rate": 5.58250542607826e-06,
+      "loss": 2.143,
+      "step": 227000
+    },
+    {
+      "epoch": 1.7000282091662977,
+      "grad_norm": 95907.765625,
+      "learning_rate": 5.448528121165691e-06,
+      "loss": 2.1631,
+      "step": 227500
+    },
+    {
+      "epoch": 1.7037645225832123,
+      "grad_norm": 88561.578125,
+      "learning_rate": 5.316085405077159e-06,
+      "loss": 2.135,
+      "step": 228000
+    },
+    {
+      "epoch": 1.707500836000127,
+      "grad_norm": 94637.078125,
+      "learning_rate": 5.185181839841746e-06,
+      "loss": 2.1314,
+      "step": 228500
+    },
+    {
+      "epoch": 1.7112371494170417,
+      "grad_norm": 96587.3125,
+      "learning_rate": 5.055821934472071e-06,
+      "loss": 2.1399,
+      "step": 229000
+    },
+    {
+      "epoch": 1.7149734628339564,
+      "grad_norm": 91652.40625,
+      "learning_rate": 4.92801014480892e-06,
+      "loss": 2.1798,
+      "step": 229500
+    },
+    {
+      "epoch": 1.7187097762508712,
+      "grad_norm": 104303.1015625,
+      "learning_rate": 4.801750873367728e-06,
+      "loss": 2.1749,
+      "step": 230000
+    },
+    {
+      "epoch": 1.7224460896677858,
+      "grad_norm": 85994.984375,
+      "learning_rate": 4.677048469187045e-06,
+      "loss": 2.161,
+      "step": 230500
+    },
+    {
+      "epoch": 1.7261824030847004,
+      "grad_norm": 105747.375,
+      "learning_rate": 4.553907227678594e-06,
+      "loss": 2.1273,
+      "step": 231000
+    },
+    {
+      "epoch": 1.729918716501615,
+      "grad_norm": 108115.8125,
+      "learning_rate": 4.432331390479416e-06,
+      "loss": 2.1493,
+      "step": 231500
+    },
+    {
+      "epoch": 1.7336550299185296,
+      "grad_norm": 97434.2109375,
+      "learning_rate": 4.312325145305729e-06,
+      "loss": 2.1339,
+      "step": 232000
+    },
+    {
+      "epoch": 1.7373913433354442,
+      "grad_norm": 100293.0546875,
+      "learning_rate": 4.193892625808659e-06,
+      "loss": 2.1501,
+      "step": 232500
+    },
+    {
+      "epoch": 1.741127656752359,
+      "grad_norm": 102609.921875,
+      "learning_rate": 4.077037911431913e-06,
+      "loss": 2.1309,
+      "step": 233000
+    },
+    {
+      "epoch": 1.7448639701692736,
+      "grad_norm": 115093.4453125,
+      "learning_rate": 3.961765027271197e-06,
+      "loss": 2.144,
+      "step": 233500
+    },
+    {
+      "epoch": 1.7486002835861885,
+      "grad_norm": 85552.25,
+      "learning_rate": 3.84807794393563e-06,
+      "loss": 2.0989,
+      "step": 234000
+    },
+    {
+      "epoch": 1.752336597003103,
+      "grad_norm": 104233.3359375,
+      "learning_rate": 3.735980577410908e-06,
+      "loss": 2.1801,
+      "step": 234500
+    },
+    {
+      "epoch": 1.7560729104200177,
+      "grad_norm": 103465.9375,
+      "learning_rate": 3.6254767889244813e-06,
+      "loss": 2.131,
+      "step": 235000
+    },
+    {
+      "epoch": 1.7598092238369323,
+      "grad_norm": 100807.5546875,
+      "learning_rate": 3.516570384812523e-06,
+      "loss": 2.1221,
+      "step": 235500
+    },
+    {
+      "epoch": 1.7635455372538469,
+      "grad_norm": 104049.0078125,
+      "learning_rate": 3.4092651163888154e-06,
+      "loss": 2.1382,
+      "step": 236000
+    },
+    {
+      "epoch": 1.7672818506707617,
+      "grad_norm": 100944.1015625,
+      "learning_rate": 3.303564679815557e-06,
+      "loss": 2.1736,
+      "step": 236500
+    },
+    {
+      "epoch": 1.7710181640876763,
+      "grad_norm": 104766.390625,
+      "learning_rate": 3.1994727159759986e-06,
+      "loss": 2.1125,
+      "step": 237000
+    },
+    {
+      "epoch": 1.7747544775045911,
+      "grad_norm": 95535.8515625,
+      "learning_rate": 3.0969928103490976e-06,
+      "loss": 2.1478,
+      "step": 237500
+    },
+    {
+      "epoch": 1.7784907909215057,
+      "grad_norm": 89555.0859375,
+      "learning_rate": 2.996128492885969e-06,
+      "loss": 2.1244,
+      "step": 238000
+    },
+    {
+      "epoch": 1.7822271043384204,
+      "grad_norm": 101483.0546875,
+      "learning_rate": 2.8968832378882926e-06,
+      "loss": 2.1287,
+      "step": 238500
+    },
+    {
+      "epoch": 1.785963417755335,
+      "grad_norm": 84929.1640625,
+      "learning_rate": 2.7992604638886975e-06,
+      "loss": 2.1421,
+      "step": 239000
+    },
+    {
+      "epoch": 1.7896997311722496,
+      "grad_norm": 96690.8984375,
+      "learning_rate": 2.703263533532929e-06,
+      "loss": 2.1681,
+      "step": 239500
+    },
+    {
+      "epoch": 1.7934360445891642,
+      "grad_norm": 105319.703125,
+      "learning_rate": 2.608895753464069e-06,
+      "loss": 2.1331,
+      "step": 240000
+    },
+    {
+      "epoch": 1.797172358006079,
+      "grad_norm": 101327.21875,
+      "learning_rate": 2.516160374208648e-06,
+      "loss": 2.1551,
+      "step": 240500
+    },
+    {
+      "epoch": 1.8009086714229936,
+      "grad_norm": 94917.8125,
+      "learning_rate": 2.4250605900646204e-06,
+      "loss": 2.1703,
+      "step": 241000
+    },
+    {
+      "epoch": 1.8046449848399084,
+      "grad_norm": 97282.4765625,
+      "learning_rate": 2.3355995389914342e-06,
+      "loss": 2.1457,
+      "step": 241500
+    },
+    {
+      "epoch": 1.808381298256823,
+      "grad_norm": 113476.78125,
+      "learning_rate": 2.2477803025018228e-06,
+      "loss": 2.1237,
+      "step": 242000
+    },
+    {
+      "epoch": 1.8121176116737376,
+      "grad_norm": 99987.9921875,
+      "learning_rate": 2.1616059055557636e-06,
+      "loss": 2.1792,
+      "step": 242500
+    },
+    {
+      "epoch": 1.8158539250906522,
+      "grad_norm": 110651.3828125,
+      "learning_rate": 2.0770793164562373e-06,
+      "loss": 2.135,
+      "step": 243000
+    },
+    {
+      "epoch": 1.8195902385075668,
+      "grad_norm": 96608.578125,
+      "learning_rate": 1.994203446746962e-06,
+      "loss": 2.1681,
+      "step": 243500
+    },
+    {
+      "epoch": 1.8233265519244817,
+      "grad_norm": 108624.8046875,
+      "learning_rate": 1.9129811511121697e-06,
+      "loss": 2.1457,
+      "step": 244000
+    },
+    {
+      "epoch": 1.8270628653413963,
+      "grad_norm": 113863.421875,
+      "learning_rate": 1.8334152272782069e-06,
+      "loss": 2.1444,
+      "step": 244500
+    },
+    {
+      "epoch": 1.830799178758311,
+      "grad_norm": 100759.1875,
+      "learning_rate": 1.7555084159172176e-06,
+      "loss": 2.1633,
+      "step": 245000
+    },
+    {
+      "epoch": 1.8345354921752257,
+      "grad_norm": 104006.0859375,
+      "learning_rate": 1.6792634005526943e-06,
+      "loss": 2.1561,
+      "step": 245500
+    },
+    {
+      "epoch": 1.8382718055921403,
+      "grad_norm": 93557.3203125,
+      "learning_rate": 1.6046828074670894e-06,
+      "loss": 2.1397,
+      "step": 246000
+    },
+    {
+      "epoch": 1.842008119009055,
+      "grad_norm": 92767.28125,
+      "learning_rate": 1.531769205611322e-06,
+      "loss": 2.1375,
+      "step": 246500
+    },
+    {
+      "epoch": 1.8457444324259695,
+      "grad_norm": 115375.0546875,
+      "learning_rate": 1.4605251065162973e-06,
+      "loss": 2.1149,
+      "step": 247000
+    },
+    {
+      "epoch": 1.8494807458428841,
+      "grad_norm": 85288.765625,
+      "learning_rate": 1.3909529642064045e-06,
+      "loss": 2.1298,
+      "step": 247500
+    },
+    {
+      "epoch": 1.853217059259799,
+      "grad_norm": 92626.7890625,
+      "learning_rate": 1.3230551751149568e-06,
+      "loss": 2.1488,
+      "step": 248000
+    },
+    {
+      "epoch": 1.8569533726767136,
+      "grad_norm": 105719.265625,
+      "learning_rate": 1.2568340780016953e-06,
+      "loss": 2.1397,
+      "step": 248500
+    },
+    {
+      "epoch": 1.8606896860936284,
+      "grad_norm": 89167.5703125,
+      "learning_rate": 1.1922919538721876e-06,
+      "loss": 2.1225,
+      "step": 249000
+    },
+    {
+      "epoch": 1.864425999510543,
+      "grad_norm": 100311.046875,
+      "learning_rate": 1.1294310258992747e-06,
+      "loss": 2.1445,
+      "step": 249500
+    },
+    {
+      "epoch": 1.8681623129274576,
+      "grad_norm": 87286.234375,
+      "learning_rate": 1.0682534593464967e-06,
+      "loss": 2.1369,
+      "step": 250000
+    },
+    {
+      "epoch": 1.8718986263443722,
+      "grad_norm": 103840.65625,
+      "learning_rate": 1.0087613614934987e-06,
+      "loss": 2.1134,
+      "step": 250500
+    },
+    {
+      "epoch": 1.8756349397612868,
+      "grad_norm": 97787.40625,
+      "learning_rate": 9.509567815634446e-07,
+      "loss": 2.1302,
+      "step": 251000
+    },
+    {
+      "epoch": 1.8793712531782016,
+      "grad_norm": 119866.546875,
+      "learning_rate": 8.948417106524554e-07,
+      "loss": 2.1391,
+      "step": 251500
+    },
+    {
+      "epoch": 1.8831075665951162,
+      "grad_norm": 101651.5234375,
+      "learning_rate": 8.404180816609819e-07,
+      "loss": 2.1628,
+      "step": 252000
+    },
+    {
+      "epoch": 1.886843880012031,
+      "grad_norm": 103326.453125,
+      "learning_rate": 7.876877692272799e-07,
+      "loss": 2.1341,
+      "step": 252500
+    },
+    {
+      "epoch": 1.8905801934289457,
+      "grad_norm": 100837.203125,
+      "learning_rate": 7.366525896627841e-07,
+      "loss": 2.125,
+      "step": 253000
+    },
+    {
+      "epoch": 1.8943165068458603,
+      "grad_norm": 101825.6015625,
+      "learning_rate": 6.873143008895633e-07,
+      "loss": 2.116,
+      "step": 253500
+    },
+    {
+      "epoch": 1.8980528202627749,
+      "grad_norm": 84607.0234375,
+      "learning_rate": 6.396746023798139e-07,
+      "loss": 2.1599,
+      "step": 254000
+    },
+    {
+      "epoch": 1.9017891336796895,
+      "grad_norm": 85267.1328125,
+      "learning_rate": 5.937351350972287e-07,
+      "loss": 2.1605,
+      "step": 254500
+    },
+    {
+      "epoch": 1.905525447096604,
+      "grad_norm": 112646.515625,
+      "learning_rate": 5.49497481440564e-07,
+      "loss": 2.1479,
+      "step": 255000
+    },
+    {
+      "epoch": 1.909261760513519,
+      "grad_norm": 101483.7578125,
+      "learning_rate": 5.069631651890783e-07,
+      "loss": 2.1306,
+      "step": 255500
+    },
+    {
+      "epoch": 1.9129980739304335,
+      "grad_norm": 104185.2265625,
+      "learning_rate": 4.661336514500625e-07,
+      "loss": 2.1222,
+      "step": 256000
+    },
+    {
+      "epoch": 1.9167343873473484,
+      "grad_norm": 83210.1484375,
+      "learning_rate": 4.270103466083697e-07,
+      "loss": 2.1573,
+      "step": 256500
+    },
+    {
+      "epoch": 1.920470700764263,
+      "grad_norm": 109216.5703125,
+      "learning_rate": 3.895945982779925e-07,
+      "loss": 2.166,
+      "step": 257000
+    },
+    {
+      "epoch": 1.9242070141811776,
+      "grad_norm": 100591.0546875,
+      "learning_rate": 3.5388769525560585e-07,
+      "loss": 2.1291,
+      "step": 257500
+    },
+    {
+      "epoch": 1.9279433275980922,
+      "grad_norm": 98919.7421875,
+      "learning_rate": 3.19890867476208e-07,
+      "loss": 2.1449,
+      "step": 258000
+    },
+    {
+      "epoch": 1.9316796410150068,
+      "grad_norm": 90432.4765625,
+      "learning_rate": 2.876052859707379e-07,
+      "loss": 2.1195,
+      "step": 258500
+    },
+    {
+      "epoch": 1.9354159544319216,
+      "grad_norm": 94623.453125,
+      "learning_rate": 2.570320628257461e-07,
+      "loss": 2.1409,
+      "step": 259000
+    },
+    {
+      "epoch": 1.9391522678488362,
+      "grad_norm": 112643.9765625,
+      "learning_rate": 2.2817225114507567e-07,
+      "loss": 2.1521,
+      "step": 259500
+    },
+    {
+      "epoch": 1.942888581265751,
+      "grad_norm": 83827.5078125,
+      "learning_rate": 2.010268450136077e-07,
+      "loss": 2.1273,
+      "step": 260000
+    },
+    {
+      "epoch": 1.9466248946826656,
+      "grad_norm": 97442.9453125,
+      "learning_rate": 1.7559677946300533e-07,
+      "loss": 2.1454,
+      "step": 260500
+    },
+    {
+      "epoch": 1.9503612080995802,
+      "grad_norm": 92114.421875,
+      "learning_rate": 1.518829304395175e-07,
+      "loss": 2.127,
+      "step": 261000
+    },
+    {
+      "epoch": 1.9540975215164949,
+      "grad_norm": 95378.5078125,
+      "learning_rate": 1.2988611477378066e-07,
+      "loss": 2.1213,
+      "step": 261500
+    },
+    {
+      "epoch": 1.9578338349334095,
+      "grad_norm": 87441.7890625,
+      "learning_rate": 1.0960709015272464e-07,
+      "loss": 2.1427,
+      "step": 262000
+    },
+    {
+      "epoch": 1.961570148350324,
+      "grad_norm": 87605.1640625,
+      "learning_rate": 9.104655509343252e-08,
+      "loss": 2.1541,
+      "step": 262500
+    },
+    {
+      "epoch": 1.9653064617672389,
+      "grad_norm": 101794.53125,
+      "learning_rate": 7.420514891911535e-08,
+      "loss": 2.152,
+      "step": 263000
+    },
+    {
+      "epoch": 1.9690427751841535,
+      "grad_norm": 94773.6328125,
+      "learning_rate": 5.908345173705754e-08,
+      "loss": 2.1471,
+      "step": 263500
+    },
+    {
+      "epoch": 1.9727790886010683,
+      "grad_norm": 105178.625,
+      "learning_rate": 4.5681984418666224e-08,
+      "loss": 2.1364,
+      "step": 264000
+    },
+    {
+      "epoch": 1.976515402017983,
+      "grad_norm": 97559.484375,
+      "learning_rate": 3.400120858149669e-08,
+      "loss": 2.17,
+      "step": 264500
+    },
+    {
+      "epoch": 1.9802517154348975,
+      "grad_norm": 120999.3125,
+      "learning_rate": 2.404152657338732e-08,
+      "loss": 2.1488,
+      "step": 265000
+    },
+    {
+      "epoch": 1.9839880288518121,
+      "grad_norm": 117999.5859375,
+      "learning_rate": 1.5803281458570684e-08,
+      "loss": 2.1481,
+      "step": 265500
+    },
+    {
+      "epoch": 1.9877243422687267,
+      "grad_norm": 94607.5078125,
+      "learning_rate": 9.286757005871893e-09,
+      "loss": 2.1104,
+      "step": 266000
+    },
+    {
+      "epoch": 1.9914606556856416,
+      "grad_norm": 94297.2109375,
+      "learning_rate": 4.492177678938614e-09,
+      "loss": 2.1349,
+      "step": 266500
+    },
+    {
+      "epoch": 1.9951969691025562,
+      "grad_norm": 95649.546875,
+      "learning_rate": 1.4197086284917227e-09,
+      "loss": 2.1224,
+      "step": 267000
+    },
+    {
+      "epoch": 1.998933282519471,
+      "grad_norm": 116585.421875,
+      "learning_rate": 6.945568664651347e-11,
+      "loss": 2.1757,
+      "step": 267500
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 267642,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 6.913150854495206e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd3718b3980300eba59bf2286f11a71efbe8c04d5c6b906fb9f99f03389ae634
+size 5240