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 sha256:62b2d61f5fdecbbf3015f88aba528f5c8a06ff4ea59f6dfead221ad39324a8dd
+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:d66130dd754887a465bda9235035ea9911826b1e9794764d65edb9a9914284c8
-size 224537202
+oid sha256:1b0bc1e35ae9b098fcf5452ff46ce2554c38302fca135d1505894551bf9f29f6
+size 286585234

checkpoint-1250/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ef026f555a46c42d792322d35411643777570b1ca6b1014cfa7a30abaf4e64c8
+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:d2fdb54fb6aad546e6936eec86cf48cc2ae80401c6cff9226dc16a5d0aca9d01
+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:f3dfc21a9f40d65c985f229cae17b538db05f0589364cb9200df08ddce783b5e
+oid sha256:c282be22e3e90af7037a4441c7fe54d3ba862cd36b36d2dc1572c68b11b27e15
 size 15024

checkpoint-1250/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5ffbc1456aa6e9a5e3dac019657cda183a5796343ccc9a690ae38f4a76e576c5
+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": 0.6251392960548401,
       "learning_rate": 5e-05,
-      "loss": 2.427,
+      "loss": 2.1914,
       "step": 100
     },
     {
       "epoch": 0.8,
-      "grad_norm": 0.6590011715888977,
       "learning_rate": 5e-05,
-      "loss": 2.2748,
+      "loss": 1.798,
       "step": 200
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.5142539682539683,
-      "eval_loss": 2.336690902709961,
-      "eval_runtime": 5.4733,
-      "eval_samples_per_second": 91.353,
-      "eval_steps_per_second": 11.51,
+      "eval_accuracy": 0.6067341772151899,
+      "eval_loss": 1.8213441371917725,
+      "eval_runtime": 4.8116,
+      "eval_samples_per_second": 103.916,
+      "eval_steps_per_second": 13.093,
       "step": 250
     },
     {
       "epoch": 1.0,
-      "eval_exact_match": 16.8,
-      "eval_f1": 24.850246583481884,
+      "eval_exact_match": 25.4,
+      "eval_f1": 35.50992875545509,
       "step": 250
     },
     {
       "epoch": 1.2,
-      "grad_norm": 0.8194239139556885,
       "learning_rate": 5e-05,
-      "loss": 2.2322,
+      "loss": 1.7534,
       "step": 300
     },
     {
       "epoch": 1.6,
-      "grad_norm": 0.9867294430732727,
       "learning_rate": 5e-05,
-      "loss": 2.1904,
+      "loss": 1.7163,
       "step": 400
     },
     {
       "epoch": 2.0,
-      "grad_norm": 0.9660400152206421,
       "learning_rate": 5e-05,
-      "loss": 2.1657,
+      "loss": 1.7,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5168253968253969,
-      "eval_loss": 2.323688507080078,
-      "eval_runtime": 5.1683,
-      "eval_samples_per_second": 96.744,
-      "eval_steps_per_second": 12.19,
+      "eval_accuracy": 0.6077215189873417,
+      "eval_loss": 1.8046445846557617,
+      "eval_runtime": 5.1228,
+      "eval_samples_per_second": 97.603,
+      "eval_steps_per_second": 12.298,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 17.2,
-      "eval_f1": 24.620036075036083,
+      "eval_exact_match": 25.8,
+      "eval_f1": 35.76454372424963,
       "step": 500
     },
     {
       "epoch": 2.4,
-      "grad_norm": 1.281128168106079,
       "learning_rate": 5e-05,
-      "loss": 2.0251,
+      "loss": 1.5854,
       "step": 600
     },
     {
       "epoch": 2.8,
-      "grad_norm": 1.3365116119384766,
       "learning_rate": 5e-05,
-      "loss": 2.0158,
+      "loss": 1.5869,
       "step": 700
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.5156507936507937,
-      "eval_loss": 2.3560824394226074,
-      "eval_runtime": 5.4675,
-      "eval_samples_per_second": 91.449,
-      "eval_steps_per_second": 11.523,
+      "eval_accuracy": 0.6071392405063291,
+      "eval_loss": 1.829284906387329,
+      "eval_runtime": 4.5692,
+      "eval_samples_per_second": 109.429,
+      "eval_steps_per_second": 13.788,
       "step": 750
     },
     {
       "epoch": 3.0,
-      "eval_exact_match": 19.0,
-      "eval_f1": 27.242790688843332,
+      "eval_exact_match": 24.4,
+      "eval_f1": 34.46831890331891,
       "step": 750
     },
     {
       "epoch": 3.2,
-      "grad_norm": 1.7840828895568848,
       "learning_rate": 5e-05,
-      "loss": 1.9394,
+      "loss": 1.5039,
       "step": 800
     },
     {
       "epoch": 3.6,
-      "grad_norm": 1.8030190467834473,
       "learning_rate": 5e-05,
-      "loss": 1.8502,
+      "loss": 1.4165,
       "step": 900
     },
     {
       "epoch": 4.0,
-      "grad_norm": 1.9682222604751587,
       "learning_rate": 5e-05,
-      "loss": 1.8682,
+      "loss": 1.4349,
       "step": 1000
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.5131746031746032,
-      "eval_loss": 2.4071946144104004,
-      "eval_runtime": 5.2825,
-      "eval_samples_per_second": 94.653,
-      "eval_steps_per_second": 11.926,
+      "eval_accuracy": 0.6042784810126582,
+      "eval_loss": 1.89736008644104,
+      "eval_runtime": 4.5698,
+      "eval_samples_per_second": 109.415,
+      "eval_steps_per_second": 13.786,
       "step": 1000
     },
     {
       "epoch": 4.0,
-      "eval_exact_match": 17.6,
-      "eval_f1": 26.010988455988468,
+      "eval_exact_match": 22.8,
+      "eval_f1": 32.991320346320364,
       "step": 1000
     },
     {
       "epoch": 4.4,
-      "grad_norm": 2.138505220413208,
       "learning_rate": 5e-05,
-      "loss": 1.6795,
+      "loss": 1.2624,
       "step": 1100
     },
     {
       "epoch": 4.8,
-      "grad_norm": 2.3212926387786865,
       "learning_rate": 5e-05,
-      "loss": 1.7036,
+      "loss": 1.3111,
       "step": 1200
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.5099365079365079,
-      "eval_loss": 2.494770050048828,
-      "eval_runtime": 5.1799,
-      "eval_samples_per_second": 96.528,
-      "eval_steps_per_second": 12.162,
+      "eval_accuracy": 0.601493670886076,
+      "eval_loss": 1.9769378900527954,
+      "eval_runtime": 4.7168,
+      "eval_samples_per_second": 106.005,
+      "eval_steps_per_second": 13.357,
       "step": 1250
     },
     {
       "epoch": 5.0,
-      "eval_exact_match": 17.2,
-      "eval_f1": 25.84752525252526,
+      "eval_exact_match": 22.6,
+      "eval_f1": 32.46891774891776,
       "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": 3.233846448147661e+16,
-  "train_batch_size": 1,
+  "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:ede3565a5658e9b8743d6efdd527fd23709623a130469c81f1191d07c33fbcb1
-size 5304
+oid sha256:ddd08619aa8955df0f518864a252643438869f6968a9f390d3026fbeb48e7ddf
+size 4728