Training in progress, epoch 6, checkpoint

checkpoint-2049/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-2049/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cba262d9ac7ed8cce020be0dd4f65becfb668e0b0f412efed06fc74202d09815
+size 143269386

checkpoint-2049/added_tokens.json

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

checkpoint-2049/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0ac03a18458e29d23a156c4f581edb0ee5872c330000ad20e78bf297302f4657
-size 224537202
+oid sha256:58073175243c35e731f69c55fa2ed99bcebdcc1604d2747955e5c29403d121d9
+size 286585234

checkpoint-2049/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:22aba0bb1bbc3a0dd716cd0efd7e126541e724e71ad16c510c9ee1b44d18bdb9
+oid sha256:8d18d9a25c4730bc9d3f04c0d9003513d84e4b6f1243f12bf97f04570aec2e2a
 size 15024

checkpoint-2049/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:95c9e03ca05fd020fa3d4ca7e0609b8c5ff920ccc6be11af14be1a8bc7dd59ca
+oid sha256:67d1286470d900669b6a628c6f3697476227ab6b5fb2146f4fc19ef4066daf3b
 size 15024

checkpoint-2049/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6633f321421b51ccf347f166ddb394a559452948c39779a162d9f83df6aa058f
+oid sha256:481168f77ef387bdedefb53497a818d46c7abd5c90d876747c068aa5a5e44fa2
 size 15024

checkpoint-2049/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c1154f90f090cfab662060601eff48abe6e06090c19ddf2a9ae98c151cba85f4
+oid sha256:8db6609bd9f27b23164e34d3394fccd90937d2d778bccd4154fbf07b9d04a0dc
 size 15024

checkpoint-2049/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-2049/tokenizer.model

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

checkpoint-2049/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-2049/trainer_state.json

@@ -9,261 +9,227 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.29282576866764276,
-      "grad_norm": 0.2411503791809082,
+      "epoch": 0.29,
       "learning_rate": 0.0001,
-      "loss": 2.0009,
+      "loss": 1.4842,
       "step": 100
     },
     {
-      "epoch": 0.5856515373352855,
-      "grad_norm": 0.22286373376846313,
+      "epoch": 0.59,
       "learning_rate": 0.0001,
-      "loss": 1.9538,
+      "loss": 1.4016,
       "step": 200
     },
     {
-      "epoch": 0.8784773060029283,
-      "grad_norm": 0.24079065024852753,
+      "epoch": 0.88,
       "learning_rate": 0.0001,
-      "loss": 1.9571,
+      "loss": 1.3892,
       "step": 300
     },
     {
-      "epoch": 0.9985358711566618,
-      "eval_accuracy": 0.45379487179487177,
-      "eval_loss": 3.9512319564819336,
-      "eval_runtime": 5.0647,
-      "eval_samples_per_second": 98.722,
-      "eval_steps_per_second": 12.439,
+      "epoch": 1.0,
+      "eval_accuracy": 0.4543589743589744,
+      "eval_loss": 3.4055869579315186,
+      "eval_runtime": 4.1688,
+      "eval_samples_per_second": 119.938,
+      "eval_steps_per_second": 15.112,
       "step": 341
     },
     {
-      "epoch": 0.9985358711566618,
-      "eval_bleu": 0.28767194245703276,
-      "eval_exact_match": 0.03,
-      "eval_prefix_exact_match": 0.042,
+      "epoch": 1.0,
+      "eval_bleu": 0.09711501224106496,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.054,
       "step": 341
     },
     {
-      "epoch": 1.171303074670571,
-      "grad_norm": 0.33546799421310425,
+      "epoch": 1.17,
       "learning_rate": 0.0001,
-      "loss": 1.9255,
+      "loss": 1.3608,
       "step": 400
     },
     {
-      "epoch": 1.4641288433382138,
-      "grad_norm": 0.3515666723251343,
+      "epoch": 1.46,
       "learning_rate": 0.0001,
-      "loss": 1.8684,
+      "loss": 1.3456,
       "step": 500
     },
     {
-      "epoch": 1.7569546120058566,
-      "grad_norm": 0.35733896493911743,
+      "epoch": 1.76,
       "learning_rate": 0.0001,
-      "loss": 1.8819,
+      "loss": 1.3499,
       "step": 600
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.4482564102564103,
-      "eval_loss": 4.112804412841797,
-      "eval_runtime": 5.0848,
-      "eval_samples_per_second": 98.333,
-      "eval_steps_per_second": 12.39,
+      "eval_accuracy": 0.4576923076923077,
+      "eval_loss": 3.453113317489624,
+      "eval_runtime": 4.4031,
+      "eval_samples_per_second": 113.556,
+      "eval_steps_per_second": 14.308,
       "step": 683
     },
     {
       "epoch": 2.0,
-      "eval_bleu": 0.24521038067633488,
-      "eval_exact_match": 0.02,
-      "eval_prefix_exact_match": 0.044,
+      "eval_bleu": 0.1225593574040313,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.056,
       "step": 683
     },
     {
-      "epoch": 2.049780380673499,
-      "grad_norm": 0.3784458339214325,
+      "epoch": 2.05,
       "learning_rate": 0.0001,
-      "loss": 1.8613,
+      "loss": 1.3238,
       "step": 700
     },
     {
-      "epoch": 2.342606149341142,
-      "grad_norm": 0.46702849864959717,
+      "epoch": 2.34,
       "learning_rate": 0.0001,
-      "loss": 1.743,
+      "loss": 1.2486,
       "step": 800
     },
     {
-      "epoch": 2.6354319180087846,
-      "grad_norm": 0.5156407356262207,
+      "epoch": 2.64,
       "learning_rate": 0.0001,
-      "loss": 1.7653,
+      "loss": 1.2452,
       "step": 900
     },
     {
-      "epoch": 2.9282576866764276,
-      "grad_norm": 0.4755038619041443,
+      "epoch": 2.93,
       "learning_rate": 0.0001,
-      "loss": 1.7702,
+      "loss": 1.2427,
       "step": 1000
     },
     {
-      "epoch": 2.998535871156662,
-      "eval_accuracy": 0.446051282051282,
-      "eval_loss": 4.327652454376221,
-      "eval_runtime": 5.1641,
-      "eval_samples_per_second": 96.822,
-      "eval_steps_per_second": 12.2,
+      "epoch": 3.0,
+      "eval_accuracy": 0.4584102564102564,
+      "eval_loss": 3.671137809753418,
+      "eval_runtime": 4.4094,
+      "eval_samples_per_second": 113.393,
+      "eval_steps_per_second": 14.288,
       "step": 1024
     },
     {
-      "epoch": 2.998535871156662,
-      "eval_bleu": 0.2598440457703149,
-      "eval_exact_match": 0.018,
-      "eval_prefix_exact_match": 0.034,
+      "epoch": 3.0,
+      "eval_bleu": 0.13165889774889974,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.056,
       "step": 1024
     },
     {
-      "epoch": 3.22108345534407,
-      "grad_norm": 0.5288150906562805,
+      "epoch": 3.22,
       "learning_rate": 0.0001,
-      "loss": 1.6325,
+      "loss": 1.1507,
       "step": 1100
     },
     {
-      "epoch": 3.513909224011713,
-      "grad_norm": 0.5784010887145996,
+      "epoch": 3.51,
       "learning_rate": 0.0001,
-      "loss": 1.5896,
+      "loss": 1.1218,
       "step": 1200
     },
     {
-      "epoch": 3.8067349926793557,
-      "grad_norm": 0.6458429098129272,
+      "epoch": 3.81,
       "learning_rate": 0.0001,
-      "loss": 1.6163,
+      "loss": 1.1231,
       "step": 1300
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.44235897435897437,
-      "eval_loss": 4.584905624389648,
-      "eval_runtime": 4.9243,
-      "eval_samples_per_second": 101.538,
-      "eval_steps_per_second": 12.794,
+      "eval_accuracy": 0.45697435897435895,
+      "eval_loss": 3.7999510765075684,
+      "eval_runtime": 4.0886,
+      "eval_samples_per_second": 122.29,
+      "eval_steps_per_second": 15.409,
       "step": 1366
     },
     {
       "epoch": 4.0,
-      "eval_bleu": 0.27967222841565076,
-      "eval_exact_match": 0.01,
-      "eval_prefix_exact_match": 0.038,
+      "eval_bleu": 0.15413978034957934,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.054,
       "step": 1366
     },
     {
-      "epoch": 4.099560761346998,
-      "grad_norm": 0.732841432094574,
+      "epoch": 4.1,
       "learning_rate": 0.0001,
-      "loss": 1.5441,
+      "loss": 1.0823,
       "step": 1400
     },
     {
-      "epoch": 4.392386530014641,
-      "grad_norm": 0.7372816205024719,
+      "epoch": 4.39,
       "learning_rate": 0.0001,
-      "loss": 1.4142,
+      "loss": 0.997,
       "step": 1500
     },
     {
-      "epoch": 4.685212298682284,
-      "grad_norm": 0.8235049247741699,
+      "epoch": 4.69,
       "learning_rate": 0.0001,
-      "loss": 1.4198,
+      "loss": 1.0024,
       "step": 1600
     },
     {
-      "epoch": 4.978038067349927,
-      "grad_norm": 0.7765743732452393,
+      "epoch": 4.98,
       "learning_rate": 0.0001,
-      "loss": 1.4427,
+      "loss": 0.995,
       "step": 1700
     },
     {
-      "epoch": 4.998535871156662,
-      "eval_accuracy": 0.43861538461538463,
-      "eval_loss": 4.850330829620361,
-      "eval_runtime": 4.9395,
-      "eval_samples_per_second": 101.224,
-      "eval_steps_per_second": 12.754,
+      "epoch": 5.0,
+      "eval_accuracy": 0.4552307692307692,
+      "eval_loss": 3.953216314315796,
+      "eval_runtime": 4.9824,
+      "eval_samples_per_second": 100.354,
+      "eval_steps_per_second": 12.645,
       "step": 1707
     },
     {
-      "epoch": 4.998535871156662,
-      "eval_bleu": 0.2170008640509292,
-      "eval_exact_match": 0.006,
-      "eval_prefix_exact_match": 0.032,
+      "epoch": 5.0,
+      "eval_bleu": 0.1611344512465791,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.046,
       "step": 1707
     },
     {
-      "epoch": 5.270863836017569,
-      "grad_norm": 0.9119400978088379,
+      "epoch": 5.27,
       "learning_rate": 0.0001,
-      "loss": 1.2469,
+      "loss": 0.8495,
       "step": 1800
     },
     {
-      "epoch": 5.563689604685212,
-      "grad_norm": 0.8966185450553894,
+      "epoch": 5.56,
       "learning_rate": 0.0001,
-      "loss": 1.2294,
+      "loss": 0.854,
       "step": 1900
     },
     {
-      "epoch": 5.856515373352855,
-      "grad_norm": 0.9629669189453125,
+      "epoch": 5.86,
       "learning_rate": 0.0001,
-      "loss": 1.2498,
+      "loss": 0.8693,
       "step": 2000
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.4348717948717949,
-      "eval_loss": 5.09260892868042,
-      "eval_runtime": 4.9572,
-      "eval_samples_per_second": 100.863,
-      "eval_steps_per_second": 12.709,
+      "eval_accuracy": 0.45261538461538464,
+      "eval_loss": 4.0766072273254395,
+      "eval_runtime": 4.8095,
+      "eval_samples_per_second": 103.961,
+      "eval_steps_per_second": 13.099,
       "step": 2049
     },
     {
       "epoch": 6.0,
-      "eval_bleu": 0.20196479624851718,
-      "eval_exact_match": 0.002,
-      "eval_prefix_exact_match": 0.03,
+      "eval_bleu": 0.1489904300924264,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.044,
       "step": 2049
     }
   ],
   "logging_steps": 100,
   "max_steps": 17050,
-  "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": 1.7594902430036787e+17,
-  "train_batch_size": 1,
+  "total_flos": 4.058144497152819e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-2049/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7e63e74298c414f7364874e60281a380506bddfac466373e2426a740b4eb5594
-size 5304
+oid sha256:3c60050af374bb2df141f4673c8141d59d83a5bd892d35ffa56ce8af555309ff
+size 4728