Training in progress, epoch 10, checkpoint

checkpoint-2062/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-2062/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b67176073d0bd16d1db5193ece95f1ec338d9f4345a5d27077ff337fff7aefc
+size 143269386

checkpoint-2062/added_tokens.json

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

checkpoint-2062/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d5751e0249d32eb88849c5cca8367f7ea98ea9ffe8752fe7bcadfb9363cde616
-size 224537202
+oid sha256:9fb19f71cdcd51719dccb18b753e4e97e223b2dfe2db7889a6e6737d1016e1b3
+size 286585234

checkpoint-2062/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e572cafff48afd1f69bdc3e7a06a879dc8392b15e1c0e74b3ffff21d85373f5c
+oid sha256:64ddc84174387dc872a8422d7a9d32c03c69f3bcdbaea4dfb0025a142a5a72c3
 size 15024

checkpoint-2062/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fc7199b52bb0ddb30f94ae094b9ef5c6f19b5d84c9be1a685cef82ef56e98e73
+oid sha256:0c80fa4d254795e8dac89b310118199450a3b1bcca69c7e120d1ac80fc2a1e63
 size 15024

checkpoint-2062/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eee6ff2f735f407fa2d9212e380654ce098faee8080fd9a899b2973d7f408fa6
+oid sha256:4918e4fa10ee439c789cf62f9db648b6d32ea6fd09274a54e84694792ac25796
 size 15024

checkpoint-2062/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6ad3519f90b378f2ac2c02eb503d597dc80720162f7321ad032313d54f63e456
+oid sha256:691333c78ce5e1d307821ee320949776f25e23b5c63226e6b9560657cb4681a1
 size 15024

checkpoint-2062/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-2062/tokenizer.model

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

checkpoint-2062/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-2062/trainer_state.json

@@ -9,330 +9,296 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.5333333333333333,
-      "grad_norm": 0.5194531679153442,
+      "epoch": 0.53,
       "learning_rate": 0.0003,
-      "loss": 1.7259,
+      "loss": 1.4433,
       "step": 100
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_accuracy": 0.5753333333333334,
-      "eval_loss": 1.6105104684829712,
-      "eval_runtime": 5.418,
-      "eval_samples_per_second": 92.285,
-      "eval_steps_per_second": 11.628,
+      "epoch": 1.0,
+      "eval_accuracy": 0.6165641025641025,
+      "eval_loss": 1.2625622749328613,
+      "eval_runtime": 4.3923,
+      "eval_samples_per_second": 113.835,
+      "eval_steps_per_second": 14.343,
       "step": 187
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_exact_match": 17.2,
-      "eval_f1": 26.890173160173163,
+      "epoch": 1.0,
+      "eval_exact_match": 26.4,
+      "eval_f1": 39.52571428571429,
       "step": 187
     },
     {
-      "epoch": 1.0666666666666667,
-      "grad_norm": 0.6734262704849243,
+      "epoch": 1.07,
       "learning_rate": 0.0003,
-      "loss": 1.6156,
+      "loss": 1.2659,
       "step": 200
     },
     {
       "epoch": 1.6,
-      "grad_norm": 0.7422465682029724,
       "learning_rate": 0.0003,
-      "loss": 1.3358,
+      "loss": 1.0761,
       "step": 300
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5722564102564103,
-      "eval_loss": 1.6464648246765137,
-      "eval_runtime": 5.1856,
-      "eval_samples_per_second": 96.422,
-      "eval_steps_per_second": 12.149,
+      "eval_accuracy": 0.615897435897436,
+      "eval_loss": 1.295995831489563,
+      "eval_runtime": 4.4145,
+      "eval_samples_per_second": 113.264,
+      "eval_steps_per_second": 14.271,
       "step": 375
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 19.0,
-      "eval_f1": 27.762380952380955,
+      "eval_exact_match": 25.4,
+      "eval_f1": 38.36603174603176,
       "step": 375
     },
     {
-      "epoch": 2.1333333333333333,
-      "grad_norm": 0.828364908695221,
+      "epoch": 2.13,
       "learning_rate": 0.0003,
-      "loss": 1.2607,
+      "loss": 0.9947,
       "step": 400
     },
     {
-      "epoch": 2.6666666666666665,
-      "grad_norm": 0.9025559425354004,
+      "epoch": 2.67,
       "learning_rate": 0.0003,
-      "loss": 0.9854,
+      "loss": 0.7909,
       "step": 500
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_accuracy": 0.5707692307692308,
-      "eval_loss": 1.7620763778686523,
-      "eval_runtime": 5.2589,
-      "eval_samples_per_second": 95.077,
-      "eval_steps_per_second": 11.98,
+      "epoch": 3.0,
+      "eval_accuracy": 0.6093846153846154,
+      "eval_loss": 1.42075777053833,
+      "eval_runtime": 4.4832,
+      "eval_samples_per_second": 111.529,
+      "eval_steps_per_second": 14.053,
       "step": 562
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_exact_match": 18.2,
-      "eval_f1": 27.53142857142857,
+      "epoch": 3.0,
+      "eval_exact_match": 23.6,
+      "eval_f1": 36.76761904761906,
       "step": 562
     },
     {
       "epoch": 3.2,
-      "grad_norm": 0.8217537999153137,
       "learning_rate": 0.0003,
-      "loss": 0.9099,
+      "loss": 0.7343,
       "step": 600
     },
     {
-      "epoch": 3.7333333333333334,
-      "grad_norm": 0.8315110206604004,
+      "epoch": 3.73,
       "learning_rate": 0.0003,
-      "loss": 0.7508,
+      "loss": 0.6117,
       "step": 700
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.5663076923076923,
-      "eval_loss": 1.9380683898925781,
-      "eval_runtime": 5.3338,
-      "eval_samples_per_second": 93.742,
-      "eval_steps_per_second": 11.811,
+      "eval_accuracy": 0.6065128205128205,
+      "eval_loss": 1.509896159172058,
+      "eval_runtime": 4.8185,
+      "eval_samples_per_second": 103.766,
+      "eval_steps_per_second": 13.075,
       "step": 750
     },
     {
       "epoch": 4.0,
-      "eval_exact_match": 16.0,
-      "eval_f1": 24.611904761904768,
+      "eval_exact_match": 22.6,
+      "eval_f1": 35.788253968253976,
       "step": 750
     },
     {
-      "epoch": 4.266666666666667,
-      "grad_norm": 0.6763857007026672,
+      "epoch": 4.27,
       "learning_rate": 0.0003,
-      "loss": 0.6912,
+      "loss": 0.5688,
       "step": 800
     },
     {
       "epoch": 4.8,
-      "grad_norm": 0.8408166766166687,
       "learning_rate": 0.0003,
-      "loss": 0.6396,
+      "loss": 0.535,
       "step": 900
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_accuracy": 0.5634358974358974,
-      "eval_loss": 1.9925932884216309,
-      "eval_runtime": 5.3717,
-      "eval_samples_per_second": 93.08,
-      "eval_steps_per_second": 11.728,
+      "epoch": 5.0,
+      "eval_accuracy": 0.6073333333333333,
+      "eval_loss": 1.6383951902389526,
+      "eval_runtime": 4.1044,
+      "eval_samples_per_second": 121.82,
+      "eval_steps_per_second": 15.349,
       "step": 937
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_exact_match": 16.0,
-      "eval_f1": 24.313809523809528,
+      "epoch": 5.0,
+      "eval_exact_match": 23.8,
+      "eval_f1": 37.04809523809523,
       "step": 937
     },
     {
-      "epoch": 5.333333333333333,
-      "grad_norm": 0.6721271276473999,
+      "epoch": 5.33,
       "learning_rate": 0.0003,
-      "loss": 0.5959,
+      "loss": 0.4997,
       "step": 1000
     },
     {
-      "epoch": 5.866666666666667,
-      "grad_norm": 0.9258321523666382,
+      "epoch": 5.87,
       "learning_rate": 0.0003,
-      "loss": 0.5833,
+      "loss": 0.4917,
       "step": 1100
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.5611794871794872,
-      "eval_loss": 2.101524829864502,
-      "eval_runtime": 5.4208,
-      "eval_samples_per_second": 92.238,
-      "eval_steps_per_second": 11.622,
+      "eval_accuracy": 0.6064102564102564,
+      "eval_loss": 1.6917529106140137,
+      "eval_runtime": 4.8299,
+      "eval_samples_per_second": 103.522,
+      "eval_steps_per_second": 13.044,
       "step": 1125
     },
     {
       "epoch": 6.0,
-      "eval_exact_match": 17.0,
-      "eval_f1": 25.444285714285726,
+      "eval_exact_match": 24.2,
+      "eval_f1": 36.030317460317455,
       "step": 1125
     },
     {
       "epoch": 6.4,
-      "grad_norm": 0.6029295921325684,
       "learning_rate": 0.0003,
-      "loss": 0.545,
+      "loss": 0.4605,
       "step": 1200
     },
     {
-      "epoch": 6.933333333333334,
-      "grad_norm": 0.5719736814498901,
+      "epoch": 6.93,
       "learning_rate": 0.0003,
-      "loss": 0.5567,
+      "loss": 0.4746,
       "step": 1300
     },
     {
-      "epoch": 6.997333333333334,
-      "eval_accuracy": 0.5607179487179487,
-      "eval_loss": 2.1645195484161377,
-      "eval_runtime": 5.1974,
-      "eval_samples_per_second": 96.202,
-      "eval_steps_per_second": 12.121,
+      "epoch": 7.0,
+      "eval_accuracy": 0.6048205128205129,
+      "eval_loss": 1.7547425031661987,
+      "eval_runtime": 4.2021,
+      "eval_samples_per_second": 118.988,
+      "eval_steps_per_second": 14.993,
       "step": 1312
     },
     {
-      "epoch": 6.997333333333334,
-      "eval_exact_match": 15.0,
-      "eval_f1": 23.618095238095247,
+      "epoch": 7.0,
+      "eval_exact_match": 23.8,
+      "eval_f1": 36.49507936507937,
       "step": 1312
     },
     {
-      "epoch": 7.466666666666667,
-      "grad_norm": 0.49924588203430176,
+      "epoch": 7.47,
       "learning_rate": 0.0003,
-      "loss": 0.5205,
+      "loss": 0.4432,
       "step": 1400
     },
     {
       "epoch": 8.0,
-      "grad_norm": 0.5360875129699707,
       "learning_rate": 0.0003,
-      "loss": 0.5411,
+      "loss": 0.463,
       "step": 1500
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.5614358974358974,
-      "eval_loss": 2.2039756774902344,
-      "eval_runtime": 5.0936,
-      "eval_samples_per_second": 98.162,
-      "eval_steps_per_second": 12.368,
+      "eval_accuracy": 0.6056923076923076,
+      "eval_loss": 1.7707523107528687,
+      "eval_runtime": 4.821,
+      "eval_samples_per_second": 103.713,
+      "eval_steps_per_second": 13.068,
       "step": 1500
     },
     {
       "epoch": 8.0,
-      "eval_exact_match": 16.8,
-      "eval_f1": 25.132380952380966,
+      "eval_exact_match": 22.4,
+      "eval_f1": 35.834603174603174,
       "step": 1500
     },
     {
-      "epoch": 8.533333333333333,
-      "grad_norm": 0.44560280442237854,
+      "epoch": 8.53,
       "learning_rate": 0.0003,
-      "loss": 0.5028,
+      "loss": 0.4307,
       "step": 1600
     },
     {
-      "epoch": 8.997333333333334,
-      "eval_accuracy": 0.5608205128205128,
-      "eval_loss": 2.236473321914673,
-      "eval_runtime": 5.2122,
-      "eval_samples_per_second": 95.93,
-      "eval_steps_per_second": 12.087,
+      "epoch": 9.0,
+      "eval_accuracy": 0.6051282051282051,
+      "eval_loss": 1.7772654294967651,
+      "eval_runtime": 4.4303,
+      "eval_samples_per_second": 112.86,
+      "eval_steps_per_second": 14.22,
       "step": 1687
     },
     {
-      "epoch": 8.997333333333334,
-      "eval_exact_match": 16.0,
-      "eval_f1": 24.712380952380958,
+      "epoch": 9.0,
+      "eval_exact_match": 24.4,
+      "eval_f1": 36.40761904761904,
       "step": 1687
     },
     {
-      "epoch": 9.066666666666666,
-      "grad_norm": 0.42847225069999695,
+      "epoch": 9.07,
       "learning_rate": 0.0003,
-      "loss": 0.5246,
+      "loss": 0.4694,
       "step": 1700
     },
     {
       "epoch": 9.6,
-      "grad_norm": 0.4186452627182007,
       "learning_rate": 0.0003,
-      "loss": 0.5041,
+      "loss": 0.4323,
       "step": 1800
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.5604615384615385,
-      "eval_loss": 2.2862467765808105,
-      "eval_runtime": 5.8141,
-      "eval_samples_per_second": 85.998,
-      "eval_steps_per_second": 10.836,
+      "eval_accuracy": 0.5996923076923077,
+      "eval_loss": 1.7927325963974,
+      "eval_runtime": 4.4251,
+      "eval_samples_per_second": 112.993,
+      "eval_steps_per_second": 14.237,
       "step": 1875
     },
     {
       "epoch": 10.0,
-      "eval_exact_match": 14.4,
-      "eval_f1": 22.683333333333344,
+      "eval_exact_match": 23.2,
+      "eval_f1": 35.76095238095239,
       "step": 1875
     },
     {
-      "epoch": 10.133333333333333,
-      "grad_norm": 0.427721232175827,
+      "epoch": 10.13,
       "learning_rate": 0.0003,
-      "loss": 0.5083,
+      "loss": 0.4392,
       "step": 1900
     },
     {
-      "epoch": 10.666666666666666,
-      "grad_norm": 0.4298701584339142,
+      "epoch": 10.67,
       "learning_rate": 0.0003,
-      "loss": 0.4991,
+      "loss": 0.4301,
       "step": 2000
     },
     {
-      "epoch": 10.997333333333334,
-      "eval_accuracy": 0.5602564102564103,
-      "eval_loss": 2.285149574279785,
-      "eval_runtime": 5.1591,
-      "eval_samples_per_second": 96.916,
-      "eval_steps_per_second": 12.211,
+      "epoch": 11.0,
+      "eval_accuracy": 0.6046153846153847,
+      "eval_loss": 1.820131778717041,
+      "eval_runtime": 4.2028,
+      "eval_samples_per_second": 118.968,
+      "eval_steps_per_second": 14.99,
       "step": 2062
     },
     {
-      "epoch": 10.997333333333334,
-      "eval_exact_match": 14.4,
-      "eval_f1": 22.41619047619048,
+      "epoch": 11.0,
+      "eval_exact_match": 22.6,
+      "eval_f1": 36.275425685425695,
       "step": 2062
     }
   ],
   "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": 3.592008408183603e+16,
-  "train_batch_size": 1,
+  "total_flos": 7.696566627126477e+16,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-2062/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a48aaa718d03f9c3a1cf4e80dbff36330a05dd346d42a3b722f1ae2626aabe3c
-size 5304
+oid sha256:79a83a5d44cad230e338c4f4f6393fbd052223dad36807daa3baa278ceb8a412
+size 4728