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 sha256:61ebc9ca8e76d5466a3c92cc041eb5585d8e27f42d2b4c123e685b9043a95330
+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 sha256:2b7f40df421968c4b56c9a9db94492d84825b2e5924a8f942b71329933e0d699
-size 224537202
+oid sha256:2e15e02049f07bec9bca368850b81a609637f65c129d3d0c5d28dd23b476c534
+size 286585234

checkpoint-1125/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2efa17419a9c2033bab9e002b97865bc472d59b5f11279763df2c0f7e4830a9b
+oid sha256:ffd6fd0069008b4dddae69faec1a87e33fcb31de6e9f06951f230a4ae5a42bfc
 size 15024

checkpoint-1125/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3a4eadd3d8f78e1efe57e5554e8f342747d9b42f8dfcb53c922010b62007649b
+oid sha256:a54aede9f47e68c2f49c83857b80a9b8251d1161bc0e6818c870f3907ede1716
 size 15024

checkpoint-1125/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d112554b0fb2e400e53af7593f468c74ea1d7173d0341562967e3d9aa55af930
+oid sha256:a22eba947e417fc91f0bc3227538423dc23edf1cb0f58765ffdd83c4a85d2659
 size 15024

checkpoint-1125/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cfb79661665e720fd752d165da0166a9cabb2796881b287ba3a7ce0d7c0fe095
+oid sha256:80876b8083da09cf850bea29b6c0f140172c01a678498d18c3a31ccc17e888f3
 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

@@ -9,192 +9,167 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.5333333333333333,
-      "grad_norm": 0.7454617619514465,
+      "epoch": 0.53,
       "learning_rate": 5e-05,
-      "loss": 1.8726,
+      "loss": 1.7923,
       "step": 100
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_accuracy": 0.5713333333333334,
-      "eval_loss": 1.6331332921981812,
-      "eval_runtime": 5.4718,
-      "eval_samples_per_second": 91.377,
-      "eval_steps_per_second": 11.514,
+      "epoch": 1.0,
+      "eval_accuracy": 0.6128205128205129,
+      "eval_loss": 1.2805120944976807,
+      "eval_runtime": 4.4685,
+      "eval_samples_per_second": 111.895,
+      "eval_steps_per_second": 14.099,
       "step": 187
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_exact_match": 15.0,
-      "eval_f1": 25.64857142857145,
+      "epoch": 1.0,
+      "eval_exact_match": 26.2,
+      "eval_f1": 37.42854700854702,
       "step": 187
     },
     {
-      "epoch": 1.0666666666666667,
-      "grad_norm": 0.9573808908462524,
+      "epoch": 1.07,
       "learning_rate": 5e-05,
-      "loss": 1.66,
+      "loss": 1.306,
       "step": 200
     },
     {
       "epoch": 1.6,
-      "grad_norm": 1.0792433023452759,
       "learning_rate": 5e-05,
-      "loss": 1.5764,
+      "loss": 1.2488,
       "step": 300
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5741538461538461,
-      "eval_loss": 1.6118417978286743,
-      "eval_runtime": 5.1541,
-      "eval_samples_per_second": 97.01,
-      "eval_steps_per_second": 12.223,
+      "eval_accuracy": 0.6167692307692307,
+      "eval_loss": 1.2676523923873901,
+      "eval_runtime": 4.1151,
+      "eval_samples_per_second": 121.504,
+      "eval_steps_per_second": 15.309,
       "step": 375
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 16.8,
-      "eval_f1": 26.120952380952396,
+      "eval_exact_match": 26.0,
+      "eval_f1": 39.16730158730158,
       "step": 375
     },
     {
-      "epoch": 2.1333333333333333,
-      "grad_norm": 1.2928162813186646,
+      "epoch": 2.13,
       "learning_rate": 5e-05,
-      "loss": 1.5364,
+      "loss": 1.1905,
       "step": 400
     },
     {
-      "epoch": 2.6666666666666665,
-      "grad_norm": 1.4883633852005005,
+      "epoch": 2.67,
       "learning_rate": 5e-05,
-      "loss": 1.4154,
+      "loss": 1.1097,
       "step": 500
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_accuracy": 0.5733846153846154,
-      "eval_loss": 1.6421321630477905,
-      "eval_runtime": 5.23,
-      "eval_samples_per_second": 95.603,
-      "eval_steps_per_second": 12.046,
+      "epoch": 3.0,
+      "eval_accuracy": 0.6161538461538462,
+      "eval_loss": 1.2942619323730469,
+      "eval_runtime": 4.2002,
+      "eval_samples_per_second": 119.043,
+      "eval_steps_per_second": 14.999,
       "step": 562
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_exact_match": 19.4,
-      "eval_f1": 28.4652380952381,
+      "epoch": 3.0,
+      "eval_exact_match": 25.8,
+      "eval_f1": 38.45095238095238,
       "step": 562
     },
     {
       "epoch": 3.2,
-      "grad_norm": 1.7553108930587769,
       "learning_rate": 5e-05,
-      "loss": 1.3426,
+      "loss": 1.0433,
       "step": 600
     },
     {
-      "epoch": 3.7333333333333334,
-      "grad_norm": 2.0343973636627197,
+      "epoch": 3.73,
       "learning_rate": 5e-05,
-      "loss": 1.2297,
+      "loss": 0.9244,
       "step": 700
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.5687692307692308,
-      "eval_loss": 1.7009365558624268,
-      "eval_runtime": 5.4367,
-      "eval_samples_per_second": 91.968,
-      "eval_steps_per_second": 11.588,
+      "eval_accuracy": 0.6126153846153846,
+      "eval_loss": 1.359810709953308,
+      "eval_runtime": 4.0966,
+      "eval_samples_per_second": 122.052,
+      "eval_steps_per_second": 15.379,
       "step": 750
     },
     {
       "epoch": 4.0,
-      "eval_exact_match": 17.6,
-      "eval_f1": 26.490952380952383,
+      "eval_exact_match": 24.2,
+      "eval_f1": 37.012380952380965,
       "step": 750
     },
     {
-      "epoch": 4.266666666666667,
-      "grad_norm": 2.3617172241210938,
+      "epoch": 4.27,
       "learning_rate": 5e-05,
-      "loss": 1.1613,
+      "loss": 0.8571,
       "step": 800
     },
     {
       "epoch": 4.8,
-      "grad_norm": 2.357593297958374,
       "learning_rate": 5e-05,
-      "loss": 1.0779,
+      "loss": 0.7924,
       "step": 900
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_accuracy": 0.567025641025641,
-      "eval_loss": 1.791063666343689,
-      "eval_runtime": 6.0902,
-      "eval_samples_per_second": 82.099,
-      "eval_steps_per_second": 10.345,
+      "epoch": 5.0,
+      "eval_accuracy": 0.6088717948717949,
+      "eval_loss": 1.4713767766952515,
+      "eval_runtime": 4.1965,
+      "eval_samples_per_second": 119.148,
+      "eval_steps_per_second": 15.013,
       "step": 937
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_exact_match": 16.0,
-      "eval_f1": 23.740000000000013,
+      "epoch": 5.0,
+      "eval_exact_match": 22.4,
+      "eval_f1": 34.34000000000002,
       "step": 937
     },
     {
-      "epoch": 5.333333333333333,
-      "grad_norm": 2.7846109867095947,
+      "epoch": 5.33,
       "learning_rate": 5e-05,
-      "loss": 0.9807,
+      "loss": 0.721,
       "step": 1000
     },
     {
-      "epoch": 5.866666666666667,
-      "grad_norm": 2.713942050933838,
+      "epoch": 5.87,
       "learning_rate": 5e-05,
-      "loss": 0.9547,
+      "loss": 0.6864,
       "step": 1100
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.563076923076923,
-      "eval_loss": 1.8913958072662354,
-      "eval_runtime": 5.2006,
-      "eval_samples_per_second": 96.142,
-      "eval_steps_per_second": 12.114,
+      "eval_accuracy": 0.6044615384615385,
+      "eval_loss": 1.5761321783065796,
+      "eval_runtime": 4.4048,
+      "eval_samples_per_second": 113.513,
+      "eval_steps_per_second": 14.303,
       "step": 1125
     },
     {
       "epoch": 6.0,
-      "eval_exact_match": 15.6,
-      "eval_f1": 24.21666666666668,
+      "eval_exact_match": 22.8,
+      "eval_f1": 34.564761904761916,
       "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": 1.959277291896832e+16,
-  "train_batch_size": 1,
+  "total_flos": 4.199582481252352e+16,
   "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:dd2191b7045c2a2985850d3b9f1472eaf3359ca38b9bc68efae9970e68b817ab
-size 5304
+oid sha256:742db4eb1b1050153055c5b82de0ef39604f0c615b7ee389916bfc19979a9577
+size 4728