Training in progress, epoch 6, checkpoint

checkpoint-1312/adapter_config.json

@@ -1,6 +1,6 @@
 {
   "auto_mapping": null,
-  "base_model_name_or_path": "Qwen/Qwen1.5-4B",
+  "base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,

checkpoint-1312/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a72d4627f9620571ed6d14e20477efcc6af0a76fb2d5a8088d15fa42a2358de
+size 143269386

checkpoint-1312/added_tokens.json

@@ -1,5 +1,5 @@
 {
-  "<|endoftext|>": 151643,
-  "<|im_end|>": 151645,
-  "<|im_start|>": 151644
+  "</s>": 2,
+  "<s>": 1,
+  "<unk>": 0
 }

checkpoint-1312/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:52e8020537449192e855af65d3e1c51a79a7099a5cd5f185f0c33359d0ce640b
-size 224537202
+oid sha256:caee3765dbd0269011f023930a4743df520eb2bdd1d46f2d3e09d4930335befd
+size 286585234

checkpoint-1312/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:71e412c33cda999decbbebfb149332716cdb81493fb7bd60f458fc1e77f24192
+oid sha256:9d3961ba55f68af864a992966b3a009a1dc58a5afc3d24428b45bfe9bbdd8a35
 size 15024

checkpoint-1312/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9b4c21c42fa20f03b7fc89b2130cb3d5524697acb468b8d4a164c3ba66360219
+oid sha256:4779a5baea73478e3675645339eedb3306cd97d45ca58de0f51a6cad68d7f8b7
 size 15024

checkpoint-1312/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:55d25714993f62d172f1d75d7dbb06effd62fa2aee3342c794114d3a6fa93894
+oid sha256:9e7bd75fcb33d2b3302f4d7ef31b9e1fbe1d703eb9977b7e0f8f78ee79676b6b
 size 15024

checkpoint-1312/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:318fd2fee2ccb0756ff2c7be93c33d2fa46ee5517f2e2c6489c60da0c939634c
+oid sha256:ca4a1254ca681a38b08c66e73439b3c9c2ce7707f99a1b49d2d85a5a27875f36
 size 15024

checkpoint-1312/special_tokens_map.json

@@ -1,14 +1,6 @@
 {
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>"
-  ],
-  "eos_token": {
-    "content": "<|endoftext|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
-  "pad_token": "<|endoftext|>"
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "</s>",
+  "unk_token": "<unk>"
 }

checkpoint-1312/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

checkpoint-1312/tokenizer_config.json

@@ -1,24 +1,23 @@
 {
-  "add_prefix_space": false,
   "added_tokens_decoder": {
-    "151643": {
-      "content": "<|endoftext|>",
+    "0": {
+      "content": "<unk>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "151644": {
-      "content": "<|im_start|>",
+    "1": {
+      "content": "<s>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "151645": {
-      "content": "<|im_end|>",
+    "2": {
+      "content": "</s>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -26,18 +25,16 @@
       "special": true
     }
   },
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>"
-  ],
-  "bos_token": null,
-  "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
   "clean_up_tokenization_spaces": false,
-  "eos_token": "<|endoftext|>",
-  "errors": "replace",
-  "model_max_length": 32768,
-  "pad_token": "<|endoftext|>",
-  "split_special_tokens": false,
-  "tokenizer_class": "Qwen2Tokenizer",
-  "unk_token": null
+  "eos_token": "</s>",
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "left",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": true
 }

checkpoint-1312/trainer_state.json

@@ -9,221 +9,194 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.5333333333333333,
-      "grad_norm": 0.9283596873283386,
+      "epoch": 0.53,
       "learning_rate": 3e-05,
-      "loss": 1.963,
+      "loss": 2.0369,
       "step": 100
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_accuracy": 0.5695384615384615,
-      "eval_loss": 1.6438876390457153,
-      "eval_runtime": 5.0534,
-      "eval_samples_per_second": 98.943,
-      "eval_steps_per_second": 12.467,
+      "epoch": 1.0,
+      "eval_accuracy": 0.6128205128205129,
+      "eval_loss": 1.2953312397003174,
+      "eval_runtime": 4.1114,
+      "eval_samples_per_second": 121.614,
+      "eval_steps_per_second": 15.323,
       "step": 187
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_exact_match": 16.2,
-      "eval_f1": 25.888571428571442,
+      "epoch": 1.0,
+      "eval_exact_match": 27.2,
+      "eval_f1": 37.44218938149973,
       "step": 187
     },
     {
-      "epoch": 1.0666666666666667,
-      "grad_norm": 1.0407154560089111,
+      "epoch": 1.07,
       "learning_rate": 3e-05,
-      "loss": 1.6668,
+      "loss": 1.3264,
       "step": 200
     },
     {
       "epoch": 1.6,
-      "grad_norm": 1.1281720399856567,
       "learning_rate": 3e-05,
-      "loss": 1.6099,
+      "loss": 1.2821,
       "step": 300
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5733333333333334,
-      "eval_loss": 1.618270754814148,
-      "eval_runtime": 5.1983,
-      "eval_samples_per_second": 96.186,
-      "eval_steps_per_second": 12.119,
+      "eval_accuracy": 0.6146153846153846,
+      "eval_loss": 1.2741013765335083,
+      "eval_runtime": 4.8302,
+      "eval_samples_per_second": 103.516,
+      "eval_steps_per_second": 13.043,
       "step": 375
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 16.6,
-      "eval_f1": 26.287619047619057,
+      "eval_exact_match": 25.6,
+      "eval_f1": 38.026666666666664,
       "step": 375
     },
     {
-      "epoch": 2.1333333333333333,
-      "grad_norm": 1.2508225440979004,
+      "epoch": 2.13,
       "learning_rate": 3e-05,
-      "loss": 1.5866,
+      "loss": 1.2348,
       "step": 400
     },
     {
-      "epoch": 2.6666666666666665,
-      "grad_norm": 1.3197834491729736,
+      "epoch": 2.67,
       "learning_rate": 3e-05,
-      "loss": 1.524,
+      "loss": 1.1987,
       "step": 500
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_accuracy": 0.5743589743589743,
-      "eval_loss": 1.6164497137069702,
-      "eval_runtime": 5.0495,
-      "eval_samples_per_second": 99.019,
-      "eval_steps_per_second": 12.476,
+      "epoch": 3.0,
+      "eval_accuracy": 0.6161538461538462,
+      "eval_loss": 1.2714948654174805,
+      "eval_runtime": 4.868,
+      "eval_samples_per_second": 102.712,
+      "eval_steps_per_second": 12.942,
       "step": 562
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_exact_match": 17.8,
-      "eval_f1": 27.35428571428571,
+      "epoch": 3.0,
+      "eval_exact_match": 28.0,
+      "eval_f1": 40.6962091503268,
       "step": 562
     },
     {
       "epoch": 3.2,
-      "grad_norm": 1.6320394277572632,
       "learning_rate": 3e-05,
-      "loss": 1.4752,
+      "loss": 1.1537,
       "step": 600
     },
     {
-      "epoch": 3.7333333333333334,
-      "grad_norm": 1.7101725339889526,
+      "epoch": 3.73,
       "learning_rate": 3e-05,
-      "loss": 1.3938,
+      "loss": 1.066,
       "step": 700
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.5728717948717948,
-      "eval_loss": 1.6376241445541382,
-      "eval_runtime": 4.9662,
-      "eval_samples_per_second": 100.68,
-      "eval_steps_per_second": 12.686,
+      "eval_accuracy": 0.6151282051282051,
+      "eval_loss": 1.3011157512664795,
+      "eval_runtime": 4.407,
+      "eval_samples_per_second": 113.455,
+      "eval_steps_per_second": 14.295,
       "step": 750
     },
     {
       "epoch": 4.0,
-      "eval_exact_match": 18.0,
-      "eval_f1": 27.498571428571427,
+      "eval_exact_match": 26.2,
+      "eval_f1": 38.19190476190477,
       "step": 750
     },
     {
-      "epoch": 4.266666666666667,
-      "grad_norm": 2.1727664470672607,
+      "epoch": 4.27,
       "learning_rate": 3e-05,
-      "loss": 1.342,
+      "loss": 1.001,
       "step": 800
     },
     {
       "epoch": 4.8,
-      "grad_norm": 2.341384172439575,
       "learning_rate": 3e-05,
-      "loss": 1.2685,
+      "loss": 0.9381,
       "step": 900
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_accuracy": 0.5699487179487179,
-      "eval_loss": 1.684618353843689,
-      "eval_runtime": 5.5457,
-      "eval_samples_per_second": 90.16,
-      "eval_steps_per_second": 11.36,
+      "epoch": 5.0,
+      "eval_accuracy": 0.6125641025641025,
+      "eval_loss": 1.3728467226028442,
+      "eval_runtime": 5.153,
+      "eval_samples_per_second": 97.03,
+      "eval_steps_per_second": 12.226,
       "step": 937
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_exact_match": 19.2,
-      "eval_f1": 28.190952380952375,
+      "epoch": 5.0,
+      "eval_exact_match": 24.6,
+      "eval_f1": 37.08666666666668,
       "step": 937
     },
     {
-      "epoch": 5.333333333333333,
-      "grad_norm": 2.5624172687530518,
+      "epoch": 5.33,
       "learning_rate": 3e-05,
-      "loss": 1.1847,
+      "loss": 0.8667,
       "step": 1000
     },
     {
-      "epoch": 5.866666666666667,
-      "grad_norm": 2.8159019947052,
+      "epoch": 5.87,
       "learning_rate": 3e-05,
-      "loss": 1.1591,
+      "loss": 0.8238,
       "step": 1100
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.5672820512820513,
-      "eval_loss": 1.7546838521957397,
-      "eval_runtime": 5.0186,
-      "eval_samples_per_second": 99.629,
-      "eval_steps_per_second": 12.553,
+      "eval_accuracy": 0.6090769230769231,
+      "eval_loss": 1.4598820209503174,
+      "eval_runtime": 4.0994,
+      "eval_samples_per_second": 121.97,
+      "eval_steps_per_second": 15.368,
       "step": 1125
     },
     {
       "epoch": 6.0,
-      "eval_exact_match": 16.6,
-      "eval_f1": 25.054761904761914,
+      "eval_exact_match": 23.0,
+      "eval_f1": 35.324444444444445,
       "step": 1125
     },
     {
       "epoch": 6.4,
-      "grad_norm": 3.056166410446167,
       "learning_rate": 3e-05,
-      "loss": 1.0486,
+      "loss": 0.7461,
       "step": 1200
     },
     {
-      "epoch": 6.933333333333334,
-      "grad_norm": 2.873281240463257,
+      "epoch": 6.93,
       "learning_rate": 3e-05,
-      "loss": 1.0444,
+      "loss": 0.7289,
       "step": 1300
     },
     {
-      "epoch": 6.997333333333334,
-      "eval_accuracy": 0.5643076923076923,
-      "eval_loss": 1.8394943475723267,
-      "eval_runtime": 5.5931,
-      "eval_samples_per_second": 89.396,
-      "eval_steps_per_second": 11.264,
+      "epoch": 7.0,
+      "eval_accuracy": 0.6064102564102564,
+      "eval_loss": 1.5455493927001953,
+      "eval_runtime": 4.1683,
+      "eval_samples_per_second": 119.952,
+      "eval_steps_per_second": 15.114,
       "step": 1312
     },
     {
-      "epoch": 6.997333333333334,
-      "eval_exact_match": 15.8,
-      "eval_f1": 23.925714285714296,
+      "epoch": 7.0,
+      "eval_exact_match": 21.8,
+      "eval_f1": 34.57222222222222,
       "step": 1312
     }
   ],
   "logging_steps": 100,
   "max_steps": 9350,
-  "num_input_tokens_seen": 0,
   "num_train_epochs": 50,
   "save_steps": 500,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": false
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 2.403137625260032e+16,
-  "train_batch_size": 2,
+  "total_flos": 4.899502260276429e+16,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-1312/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1a175de730fa6917b3a611bb9e53766640b7a5e4a426f825f2a602cf98661e47
-size 5304
+oid sha256:52c53f755ad4d3669bc1e92a0dd04ac5e6ef02b39406da28fac147ebfc373982
+size 4728