Training in progress, epoch 5, checkpoint

checkpoint-1250/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-1250/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d540dcee59fdea54882ca6227980a8408947037a5e1f813d868f538f192e080b
+size 143269386

checkpoint-1250/added_tokens.json

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

checkpoint-1250/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eb5227def08ce770c6b6ef3a6156896f8ca3a23853bb026a9e61b00b668eeec7
-size 224537202
+oid sha256:7f698eda21f149f5744bcba27dbfc279f0bba89730f283da0b0e8d33bd5d6d32
+size 286585234

checkpoint-1250/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c27a50faba4545c880cd93025592bf145b53239b5ac200ede874cffe184a9f6f
+oid sha256:76e4f5e11740b6ebd4866595a96a2d2f794d8c0bfae8a8a789c7d7634c23a325
 size 15024

checkpoint-1250/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f6256bc5a66e1db011a4eb97f0128029a1656b56674f464b3832fa40c9227f71
+oid sha256:e5b2be0a6f48478a29701d3d85a06ebb6f953d1e81395dc001bd42cd17037d7d
 size 15024

checkpoint-1250/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:66754ad5da9dc3d8a2f2f1d6eb74bd03574d65a7a31cb6da7bf0adfb30957b02
+oid sha256:c282be22e3e90af7037a4441c7fe54d3ba862cd36b36d2dc1572c68b11b27e15
 size 15024

checkpoint-1250/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e3b4ec9a3a825d941fda5b9f4bc04ba248e68bc60fba3f1b50846af07a52f305
+oid sha256:0f9b6da035f731d5300cb5640f1ffa2221c8682e04cb37114422dfb02fd51040
 size 15024

checkpoint-1250/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-1250/tokenizer.model

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

checkpoint-1250/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-1250/trainer_state.json

@@ -10,183 +10,157 @@
   "log_history": [
     {
       "epoch": 0.4,
-      "grad_norm": 0.6831097602844238,
       "learning_rate": 3e-05,
-      "loss": 2.487,
+      "loss": 2.3698,
       "step": 100
     },
     {
       "epoch": 0.8,
-      "grad_norm": 0.7073227763175964,
       "learning_rate": 3e-05,
-      "loss": 2.276,
+      "loss": 1.8255,
       "step": 200
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.5134285714285715,
-      "eval_loss": 2.3502681255340576,
-      "eval_runtime": 5.2185,
-      "eval_samples_per_second": 95.812,
-      "eval_steps_per_second": 12.072,
+      "eval_accuracy": 0.6054430379746836,
+      "eval_loss": 1.83919358253479,
+      "eval_runtime": 4.687,
+      "eval_samples_per_second": 106.679,
+      "eval_steps_per_second": 13.442,
       "step": 250
     },
     {
       "epoch": 1.0,
-      "eval_exact_match": 17.0,
-      "eval_f1": 24.54219480519481,
+      "eval_exact_match": 25.2,
+      "eval_f1": 34.656485532011864,
       "step": 250
     },
     {
       "epoch": 1.2,
-      "grad_norm": 0.820932149887085,
       "learning_rate": 3e-05,
-      "loss": 2.2479,
+      "loss": 1.7843,
       "step": 300
     },
     {
       "epoch": 1.6,
-      "grad_norm": 0.9243379831314087,
       "learning_rate": 3e-05,
-      "loss": 2.2237,
+      "loss": 1.7546,
       "step": 400
     },
     {
       "epoch": 2.0,
-      "grad_norm": 0.8742844462394714,
       "learning_rate": 3e-05,
-      "loss": 2.1973,
+      "loss": 1.7368,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5151746031746032,
-      "eval_loss": 2.330955743789673,
-      "eval_runtime": 5.5478,
-      "eval_samples_per_second": 90.125,
-      "eval_steps_per_second": 11.356,
+      "eval_accuracy": 0.6078227848101266,
+      "eval_loss": 1.81111478805542,
+      "eval_runtime": 4.6445,
+      "eval_samples_per_second": 107.654,
+      "eval_steps_per_second": 13.564,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 16.2,
-      "eval_f1": 24.397655122655127,
+      "eval_exact_match": 26.0,
+      "eval_f1": 35.86535000087634,
       "step": 500
     },
     {
       "epoch": 2.4,
-      "grad_norm": 1.034700632095337,
       "learning_rate": 3e-05,
-      "loss": 2.1295,
+      "loss": 1.6749,
       "step": 600
     },
     {
       "epoch": 2.8,
-      "grad_norm": 1.130967617034912,
       "learning_rate": 3e-05,
-      "loss": 2.1128,
+      "loss": 1.6689,
       "step": 700
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.516095238095238,
-      "eval_loss": 2.3357651233673096,
-      "eval_runtime": 5.8761,
-      "eval_samples_per_second": 85.091,
-      "eval_steps_per_second": 10.721,
+      "eval_accuracy": 0.607493670886076,
+      "eval_loss": 1.8103100061416626,
+      "eval_runtime": 4.7447,
+      "eval_samples_per_second": 105.381,
+      "eval_steps_per_second": 13.278,
       "step": 750
     },
     {
       "epoch": 3.0,
-      "eval_exact_match": 19.0,
-      "eval_f1": 26.202922077922086,
+      "eval_exact_match": 23.6,
+      "eval_f1": 33.977337662337675,
       "step": 750
     },
     {
       "epoch": 3.2,
-      "grad_norm": 1.4410409927368164,
       "learning_rate": 3e-05,
-      "loss": 2.0768,
+      "loss": 1.6205,
       "step": 800
     },
     {
       "epoch": 3.6,
-      "grad_norm": 1.5838419198989868,
       "learning_rate": 3e-05,
-      "loss": 2.0139,
+      "loss": 1.556,
       "step": 900
     },
     {
       "epoch": 4.0,
-      "grad_norm": 1.824843406677246,
       "learning_rate": 3e-05,
-      "loss": 2.0165,
+      "loss": 1.5555,
       "step": 1000
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.5151746031746032,
-      "eval_loss": 2.355579376220703,
-      "eval_runtime": 5.1124,
-      "eval_samples_per_second": 97.801,
-      "eval_steps_per_second": 12.323,
+      "eval_accuracy": 0.6067341772151899,
+      "eval_loss": 1.8414338827133179,
+      "eval_runtime": 5.1526,
+      "eval_samples_per_second": 97.039,
+      "eval_steps_per_second": 12.227,
       "step": 1000
     },
     {
       "epoch": 4.0,
-      "eval_exact_match": 19.0,
-      "eval_f1": 26.60538961038962,
+      "eval_exact_match": 24.6,
+      "eval_f1": 35.02336940836942,
       "step": 1000
     },
     {
       "epoch": 4.4,
-      "grad_norm": 1.988659381866455,
       "learning_rate": 3e-05,
-      "loss": 1.8919,
+      "loss": 1.4289,
       "step": 1100
     },
     {
       "epoch": 4.8,
-      "grad_norm": 2.1034774780273438,
       "learning_rate": 3e-05,
-      "loss": 1.9024,
+      "loss": 1.4559,
       "step": 1200
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.5138730158730159,
-      "eval_loss": 2.4047114849090576,
-      "eval_runtime": 5.507,
-      "eval_samples_per_second": 90.794,
-      "eval_steps_per_second": 11.44,
+      "eval_accuracy": 0.6037721518987341,
+      "eval_loss": 1.8992472887039185,
+      "eval_runtime": 5.1672,
+      "eval_samples_per_second": 96.764,
+      "eval_steps_per_second": 12.192,
       "step": 1250
     },
     {
       "epoch": 5.0,
-      "eval_exact_match": 19.2,
-      "eval_f1": 27.08811688311689,
+      "eval_exact_match": 23.0,
+      "eval_f1": 33.18553391053392,
       "step": 1250
     }
   ],
   "logging_steps": 100,
   "max_steps": 12500,
-  "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": 3.821795124497613e+16,
-  "train_batch_size": 2,
+  "total_flos": 7.98552666144768e+16,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-1250/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:76141cb7041782161f7695865881a4df3a03be3731ae75640f405be3a5a6f23d
-size 5304
+oid sha256:b708b99b821bdf687ebed6196095d1415f662e1088452054a7df6a8e9ebd3c3f
+size 4728