Training in progress, epoch 6, checkpoint

checkpoint-1125/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-1125/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f9fa7d8fe8812cd0e05f7a551acaf4c4d81f6c18c2dcb5e5d2ac40f9cc7af46
+size 143269386

checkpoint-1125/added_tokens.json

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

checkpoint-1125/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:52f082b10a855e4a0fa794b860de384798765e04d69f221b7fdd0bc5e95a04e5
-size 224537202
+oid sha256:790d24e69a22fbae97139a88c9ad92456df2b0d80bccd5452698c15923228651
+size 286585234

checkpoint-1125/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9d32f81adf26383d34578744e70a5144b5391ca3792ef0248a9022bcfcfd651d
+oid sha256:90da429f441fd2180d50e25621a6b3d6407a1e9ba401831c0bada7293a7b7d99
 size 15024

checkpoint-1125/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dad9edca65451454582b0f88936ea9ef6d373ea1c6da9fe97cce96a4f9b9ba9f
+oid sha256:8390f3449ea994277a0532425defbccc9fa812200c051e5a08e6107932e1ed13
 size 15024

checkpoint-1125/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dd699ad8dfe32e9ff1cf95826e701a5c739978d9ca44e675315ca2a06cdad4bc
+oid sha256:ea9b24868f5919de2ae27a117a574d59f4e3ee8890ed31acf13530bb3560760c
 size 15024

checkpoint-1125/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9eb0cabc7c6f6bd71e75deec6c648d39fe758c9df7429b843fa0faaac1ad4d87
+oid sha256:f7e8747b84599866ef890d296267222d7844783f39d17f6e3bd5c2110decb1ae
 size 15024

checkpoint-1125/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-1125/tokenizer.model

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

checkpoint-1125/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-1125/trainer_state.json

@@ -9,216 +9,191 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.5333333333333333,
-      "grad_norm": 0.271308958530426,
+      "epoch": 0.53,
       "learning_rate": 5e-05,
-      "loss": 1.8262,
+      "loss": 1.3988,
       "step": 100
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_accuracy": 0.6047174887892377,
-      "eval_loss": 1.7027028799057007,
-      "eval_runtime": 6.6338,
-      "eval_samples_per_second": 75.372,
-      "eval_steps_per_second": 9.497,
+      "epoch": 1.0,
+      "eval_accuracy": 0.6660784313725491,
+      "eval_loss": 1.2107256650924683,
+      "eval_runtime": 7.728,
+      "eval_samples_per_second": 64.7,
+      "eval_steps_per_second": 8.152,
       "step": 187
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_exact_match": 14.6,
-      "eval_f1": 23.665714285714298,
-      "eval_qa_bleu": 10.24028390049623,
-      "eval_qa_exact_match": 0.114,
-      "eval_recite_bleu": 12.525534079250063,
+      "epoch": 1.0,
+      "eval_exact_match": 19.4,
+      "eval_f1": 27.911959489312448,
+      "eval_qa_bleu": 2.842200543476061,
+      "eval_qa_exact_match": 0.152,
+      "eval_recite_bleu": 12.149078895446083,
       "eval_recite_exact_match": 0.0,
       "step": 187
     },
     {
-      "epoch": 1.0666666666666667,
-      "grad_norm": 0.301651269197464,
+      "epoch": 1.07,
       "learning_rate": 5e-05,
-      "loss": 1.7424,
+      "loss": 1.238,
       "step": 200
     },
     {
       "epoch": 1.6,
-      "grad_norm": 0.35949045419692993,
       "learning_rate": 5e-05,
-      "loss": 1.6978,
+      "loss": 1.1977,
       "step": 300
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.6068699551569506,
-      "eval_loss": 1.6845285892486572,
-      "eval_runtime": 5.7524,
-      "eval_samples_per_second": 86.92,
-      "eval_steps_per_second": 10.952,
+      "eval_accuracy": 0.6676078431372549,
+      "eval_loss": 1.1987448930740356,
+      "eval_runtime": 7.4628,
+      "eval_samples_per_second": 66.999,
+      "eval_steps_per_second": 8.442,
       "step": 375
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 14.6,
-      "eval_f1": 23.74190476190477,
-      "eval_qa_bleu": 9.70432698897927,
-      "eval_qa_exact_match": 0.114,
-      "eval_recite_bleu": 13.014822566317982,
-      "eval_recite_exact_match": 0.006,
+      "eval_exact_match": 21.4,
+      "eval_f1": 27.979969946667268,
+      "eval_qa_bleu": 3.4273838915653356,
+      "eval_qa_exact_match": 0.158,
+      "eval_recite_bleu": 11.474877690710539,
+      "eval_recite_exact_match": 0.0,
       "step": 375
     },
     {
-      "epoch": 2.1333333333333333,
-      "grad_norm": 0.41301965713500977,
+      "epoch": 2.13,
       "learning_rate": 5e-05,
-      "loss": 1.694,
+      "loss": 1.1628,
       "step": 400
     },
     {
-      "epoch": 2.6666666666666665,
-      "grad_norm": 0.5057485699653625,
+      "epoch": 2.67,
       "learning_rate": 5e-05,
-      "loss": 1.6324,
+      "loss": 1.1458,
       "step": 500
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_accuracy": 0.6079551569506726,
-      "eval_loss": 1.6851012706756592,
-      "eval_runtime": 5.5314,
-      "eval_samples_per_second": 90.394,
-      "eval_steps_per_second": 11.39,
+      "epoch": 3.0,
+      "eval_accuracy": 0.6675764705882353,
+      "eval_loss": 1.1964411735534668,
+      "eval_runtime": 6.6677,
+      "eval_samples_per_second": 74.988,
+      "eval_steps_per_second": 9.448,
       "step": 562
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_exact_match": 14.2,
-      "eval_f1": 23.29190476190477,
-      "eval_qa_bleu": 10.724307586340888,
-      "eval_qa_exact_match": 0.104,
-      "eval_recite_bleu": 12.83531066630853,
-      "eval_recite_exact_match": 0.004,
+      "epoch": 3.0,
+      "eval_exact_match": 23.4,
+      "eval_f1": 31.525761083856903,
+      "eval_qa_bleu": 2.0606924929592334,
+      "eval_qa_exact_match": 0.17,
+      "eval_recite_bleu": 12.027903845545522,
+      "eval_recite_exact_match": 0.0,
       "step": 562
     },
     {
       "epoch": 3.2,
-      "grad_norm": 0.6363334655761719,
       "learning_rate": 5e-05,
-      "loss": 1.6063,
+      "loss": 1.1179,
       "step": 600
     },
     {
-      "epoch": 3.7333333333333334,
-      "grad_norm": 0.6466453671455383,
+      "epoch": 3.73,
       "learning_rate": 5e-05,
-      "loss": 1.5562,
+      "loss": 1.0707,
       "step": 700
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.6071121076233184,
-      "eval_loss": 1.6969172954559326,
-      "eval_runtime": 6.2612,
-      "eval_samples_per_second": 79.857,
-      "eval_steps_per_second": 10.062,
+      "eval_accuracy": 0.6665960784313726,
+      "eval_loss": 1.208784580230713,
+      "eval_runtime": 6.6546,
+      "eval_samples_per_second": 75.136,
+      "eval_steps_per_second": 9.467,
       "step": 750
     },
     {
       "epoch": 4.0,
-      "eval_exact_match": 14.0,
-      "eval_f1": 21.34333333333334,
-      "eval_qa_bleu": 10.250340106635626,
-      "eval_qa_exact_match": 0.084,
-      "eval_recite_bleu": 12.741830585824024,
-      "eval_recite_exact_match": 0.004,
+      "eval_exact_match": 23.4,
+      "eval_f1": 33.18893657873498,
+      "eval_qa_bleu": 3.8183981024520506,
+      "eval_qa_exact_match": 0.176,
+      "eval_recite_bleu": 14.071857337313705,
+      "eval_recite_exact_match": 0.0,
       "step": 750
     },
     {
-      "epoch": 4.266666666666667,
-      "grad_norm": 0.8162885904312134,
+      "epoch": 4.27,
       "learning_rate": 5e-05,
-      "loss": 1.5052,
+      "loss": 1.0179,
       "step": 800
     },
     {
       "epoch": 4.8,
-      "grad_norm": 0.8799439072608948,
       "learning_rate": 5e-05,
-      "loss": 1.469,
+      "loss": 1.0066,
       "step": 900
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_accuracy": 0.6059282511210762,
-      "eval_loss": 1.7216957807540894,
-      "eval_runtime": 6.3012,
-      "eval_samples_per_second": 79.35,
-      "eval_steps_per_second": 9.998,
+      "epoch": 5.0,
+      "eval_accuracy": 0.6651764705882353,
+      "eval_loss": 1.2342489957809448,
+      "eval_runtime": 6.699,
+      "eval_samples_per_second": 74.638,
+      "eval_steps_per_second": 9.404,
       "step": 937
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_exact_match": 12.6,
-      "eval_f1": 21.041428571428575,
-      "eval_qa_bleu": 8.239984727774104,
-      "eval_qa_exact_match": 0.094,
-      "eval_recite_bleu": 12.327373697720192,
-      "eval_recite_exact_match": 0.002,
+      "epoch": 5.0,
+      "eval_exact_match": 24.0,
+      "eval_f1": 32.88692081962977,
+      "eval_qa_bleu": 3.6747877275980905,
+      "eval_qa_exact_match": 0.174,
+      "eval_recite_bleu": 14.36158051041567,
+      "eval_recite_exact_match": 0.006,
       "step": 937
     },
     {
-      "epoch": 5.333333333333333,
-      "grad_norm": 0.9834449291229248,
+      "epoch": 5.33,
       "learning_rate": 5e-05,
-      "loss": 1.3845,
+      "loss": 0.9595,
       "step": 1000
     },
     {
-      "epoch": 5.866666666666667,
-      "grad_norm": 0.9377098679542542,
+      "epoch": 5.87,
       "learning_rate": 5e-05,
-      "loss": 1.3927,
+      "loss": 0.9353,
       "step": 1100
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.6047354260089686,
-      "eval_loss": 1.7564114332199097,
-      "eval_runtime": 6.3447,
-      "eval_samples_per_second": 78.806,
-      "eval_steps_per_second": 9.93,
+      "eval_accuracy": 0.6628392156862745,
+      "eval_loss": 1.2645816802978516,
+      "eval_runtime": 7.4539,
+      "eval_samples_per_second": 67.079,
+      "eval_steps_per_second": 8.452,
       "step": 1125
     },
     {
       "epoch": 6.0,
-      "eval_exact_match": 13.4,
-      "eval_f1": 21.70317460317461,
-      "eval_qa_bleu": 10.522554916579404,
-      "eval_qa_exact_match": 0.104,
-      "eval_recite_bleu": 12.383604385812577,
-      "eval_recite_exact_match": 0.006,
+      "eval_exact_match": 23.4,
+      "eval_f1": 33.50380952380953,
+      "eval_qa_bleu": 13.968612067397348,
+      "eval_qa_exact_match": 0.168,
+      "eval_recite_bleu": 15.026821124315997,
+      "eval_recite_exact_match": 0.012,
       "step": 1125
     }
   ],
   "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": 1.1331927290989773e+17,
-  "train_batch_size": 1,
+  "total_flos": 2.6070114090195354e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-1125/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:01f9f0dc44c57fb44afb03320f4d02bf48066009e3bbcd96d8e4377b0171c5a5
-size 5304
+oid sha256:ae9143de0ad8fd0f027f30f757d262a6e02edc4ceeaea2c21697f4f44b5b1eeb
+size 4728