Training in progress, epoch 4, checkpoint

checkpoint-2116/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-2116/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2435c03f2190831213e9084151c6b42444d9c8557b45ea64687e82c0d57080c
+size 143269386

checkpoint-2116/added_tokens.json

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

checkpoint-2116/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:062508c1f76725c8420e0e84617dfa3c711cdb0b17a689e013ff2c326f3f9cf6
-size 224537202
+oid sha256:1fb4ec7b50b7e907baec1cc8e2e32238f7b01f26f38f372b091dbf855a82b9dc
+size 286585234

checkpoint-2116/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5bfe17eb51c7d00ebb0fc68dcab0eccbcd33b0a54364d18e3f67cca35c1c6465
+oid sha256:5482162cf79af5df816815a7998b53864f377712e6173e698a072ed90cbbd417
 size 15024

checkpoint-2116/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9c4f850fb2c216cd4c6beff06843e6c240165278df55525caaaf34f47249dc96
+oid sha256:725aceebd2200c4fbfe3cd51e2057af02d462280c324fd8bcc6c163a383c7f53
 size 15024

checkpoint-2116/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:99505752148a373771384d1aef065585c4d9e09da3eef320ba85511edad06257
+oid sha256:d0580f1bdd35d8911788c8c4cd2a6eb96109f4ab443e04a952700fea953994a4
 size 15024

checkpoint-2116/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ebe381ac361ad7aa1ac5649058f33f7e222f3ad820d37dc44e805aec2abdbb65
+oid sha256:e8ede547fad94a856ad03a42fa67de3c9904b13e45c66934c0e22d6caab956f7
 size 15024

checkpoint-2116/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-2116/tokenizer.model

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

checkpoint-2116/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-2116/trainer_state.json

@@ -9,232 +9,197 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.1890359168241966,
-      "grad_norm": 0.5107442140579224,
+      "epoch": 0.19,
       "learning_rate": 5e-05,
-      "loss": 2.0482,
+      "loss": 1.6355,
       "step": 100
     },
     {
-      "epoch": 0.3780718336483932,
-      "grad_norm": 0.5563424229621887,
+      "epoch": 0.38,
       "learning_rate": 5e-05,
-      "loss": 1.8835,
+      "loss": 1.4037,
       "step": 200
     },
     {
-      "epoch": 0.5671077504725898,
-      "grad_norm": 0.4735371172428131,
+      "epoch": 0.57,
       "learning_rate": 5e-05,
-      "loss": 1.8512,
+      "loss": 1.4183,
       "step": 300
     },
     {
-      "epoch": 0.7561436672967864,
-      "grad_norm": 0.4543495178222656,
+      "epoch": 0.76,
       "learning_rate": 5e-05,
-      "loss": 1.8489,
+      "loss": 1.3881,
       "step": 400
     },
     {
-      "epoch": 0.945179584120983,
-      "grad_norm": 0.5007112622261047,
+      "epoch": 0.95,
       "learning_rate": 5e-05,
-      "loss": 1.8439,
+      "loss": 1.3891,
       "step": 500
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.5733333333333334,
-      "eval_loss": 1.6194567680358887,
-      "eval_runtime": 5.4154,
-      "eval_samples_per_second": 92.33,
-      "eval_steps_per_second": 11.634,
+      "eval_accuracy": 0.6137948717948718,
+      "eval_loss": 1.3015451431274414,
+      "eval_runtime": 4.1821,
+      "eval_samples_per_second": 119.557,
+      "eval_steps_per_second": 15.064,
       "step": 529
     },
     {
       "epoch": 1.0,
-      "eval_exact_match": 17.0,
-      "eval_f1": 26.880952380952387,
+      "eval_exact_match": 23.2,
+      "eval_f1": 35.17014759120022,
       "step": 529
     },
     {
-      "epoch": 1.1342155009451795,
-      "grad_norm": 0.5181360244750977,
+      "epoch": 1.13,
       "learning_rate": 5e-05,
-      "loss": 1.8217,
+      "loss": 1.3808,
       "step": 600
     },
     {
-      "epoch": 1.3232514177693762,
-      "grad_norm": 1.1348918676376343,
+      "epoch": 1.32,
       "learning_rate": 5e-05,
-      "loss": 1.7866,
+      "loss": 1.348,
       "step": 700
     },
     {
-      "epoch": 1.5122873345935728,
-      "grad_norm": 0.661758303642273,
+      "epoch": 1.51,
       "learning_rate": 5e-05,
-      "loss": 1.7934,
+      "loss": 1.3627,
       "step": 800
     },
     {
-      "epoch": 1.7013232514177694,
-      "grad_norm": 0.5634772777557373,
+      "epoch": 1.7,
       "learning_rate": 5e-05,
-      "loss": 1.7731,
+      "loss": 1.3473,
       "step": 900
     },
     {
-      "epoch": 1.8903591682419658,
-      "grad_norm": 0.6698930263519287,
+      "epoch": 1.89,
       "learning_rate": 5e-05,
-      "loss": 1.7969,
+      "loss": 1.3633,
       "step": 1000
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5751794871794872,
-      "eval_loss": 1.603736162185669,
-      "eval_runtime": 5.0986,
-      "eval_samples_per_second": 98.066,
-      "eval_steps_per_second": 12.356,
+      "eval_accuracy": 0.6165641025641025,
+      "eval_loss": 1.2854759693145752,
+      "eval_runtime": 4.1064,
+      "eval_samples_per_second": 121.761,
+      "eval_steps_per_second": 15.342,
       "step": 1058
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 17.6,
-      "eval_f1": 27.218095238095238,
+      "eval_exact_match": 25.8,
+      "eval_f1": 37.509047619047635,
       "step": 1058
     },
     {
-      "epoch": 2.0793950850661624,
-      "grad_norm": 0.6265193223953247,
+      "epoch": 2.08,
       "learning_rate": 5e-05,
-      "loss": 1.7416,
+      "loss": 1.3358,
       "step": 1100
     },
     {
-      "epoch": 2.268431001890359,
-      "grad_norm": 0.812364935874939,
+      "epoch": 2.27,
       "learning_rate": 5e-05,
-      "loss": 1.6957,
+      "loss": 1.2917,
       "step": 1200
     },
     {
-      "epoch": 2.4574669187145557,
-      "grad_norm": 1.0972926616668701,
+      "epoch": 2.46,
       "learning_rate": 5e-05,
-      "loss": 1.7033,
+      "loss": 1.2948,
       "step": 1300
     },
     {
-      "epoch": 2.6465028355387523,
-      "grad_norm": 1.0831658840179443,
+      "epoch": 2.65,
       "learning_rate": 5e-05,
-      "loss": 1.6935,
+      "loss": 1.3144,
       "step": 1400
     },
     {
-      "epoch": 2.835538752362949,
-      "grad_norm": 1.0022494792938232,
+      "epoch": 2.84,
       "learning_rate": 5e-05,
-      "loss": 1.6799,
+      "loss": 1.2929,
       "step": 1500
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.573025641025641,
-      "eval_loss": 1.6428505182266235,
-      "eval_runtime": 5.1488,
-      "eval_samples_per_second": 97.111,
-      "eval_steps_per_second": 12.236,
+      "eval_accuracy": 0.6176923076923077,
+      "eval_loss": 1.2953603267669678,
+      "eval_runtime": 4.0992,
+      "eval_samples_per_second": 121.974,
+      "eval_steps_per_second": 15.369,
       "step": 1587
     },
     {
       "epoch": 3.0,
-      "eval_exact_match": 20.2,
-      "eval_f1": 27.93809523809523,
+      "eval_exact_match": 25.0,
+      "eval_f1": 37.08253968253969,
       "step": 1587
     },
     {
-      "epoch": 3.0245746691871456,
-      "grad_norm": 0.8240587711334229,
+      "epoch": 3.02,
       "learning_rate": 5e-05,
-      "loss": 1.6618,
+      "loss": 1.2906,
       "step": 1600
     },
     {
-      "epoch": 3.213610586011342,
-      "grad_norm": 1.2994645833969116,
+      "epoch": 3.21,
       "learning_rate": 5e-05,
-      "loss": 1.5736,
+      "loss": 1.2216,
       "step": 1700
     },
     {
-      "epoch": 3.402646502835539,
-      "grad_norm": 1.0891159772872925,
+      "epoch": 3.4,
       "learning_rate": 5e-05,
-      "loss": 1.5798,
+      "loss": 1.2459,
       "step": 1800
     },
     {
-      "epoch": 3.5916824196597354,
-      "grad_norm": 1.3826475143432617,
+      "epoch": 3.59,
       "learning_rate": 5e-05,
-      "loss": 1.5957,
+      "loss": 1.2474,
       "step": 1900
     },
     {
-      "epoch": 3.780718336483932,
-      "grad_norm": 0.9901586174964905,
+      "epoch": 3.78,
       "learning_rate": 5e-05,
-      "loss": 1.6026,
+      "loss": 1.2518,
       "step": 2000
     },
     {
-      "epoch": 3.9697542533081287,
-      "grad_norm": 1.0607813596725464,
+      "epoch": 3.97,
       "learning_rate": 5e-05,
-      "loss": 1.6106,
+      "loss": 1.2361,
       "step": 2100
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.5708717948717948,
-      "eval_loss": 1.7143596410751343,
-      "eval_runtime": 5.3942,
-      "eval_samples_per_second": 92.693,
-      "eval_steps_per_second": 11.679,
+      "eval_accuracy": 0.6045128205128205,
+      "eval_loss": 1.3488503694534302,
+      "eval_runtime": 4.2015,
+      "eval_samples_per_second": 119.004,
+      "eval_steps_per_second": 14.995,
       "step": 2116
     },
     {
       "epoch": 4.0,
-      "eval_exact_match": 19.0,
-      "eval_f1": 27.71428571428571,
+      "eval_exact_match": 25.0,
+      "eval_f1": 36.82126984126986,
       "step": 2116
     }
   ],
   "logging_steps": 100,
   "max_steps": 26450,
-  "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": 1.3034850571255808e+17,
-  "train_batch_size": 1,
+  "total_flos": 3.573289814620897e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-2116/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f01d18f717251d8791982fc93ed5352f2d247add98e1102e2a52cc0a32128bb7
-size 5304
+oid sha256:391c04887a1ac777804ab56ff52563bcf9c4927713b353f37947aac9be091742
+size 4728