Training in progress, epoch 10, checkpoint

checkpoint-3415/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-3415/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04c670f1f76521248fbc5c850a2caa9df11652fa433c6fab1f3d028a06bba946
+size 143269386

checkpoint-3415/added_tokens.json

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

checkpoint-3415/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ee78c1a29238d149b9c4050f075f6b7252a065c22dc3a7ddc9726248119c7643
-size 224537202
+oid sha256:ac59d39672e1ab289e522f631a90861d7b604915d1001b65455e6301c8351000
+size 286585234

checkpoint-3415/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0e484163f85d2f6135eb0a196e80e4fc9cee2ba7d087334212d17365c73cfe5a
+oid sha256:3fd2dc3faeb16e7ca25ab610fd7ee9dec02a355326d23415f632434eb4a1e69d
 size 15024

checkpoint-3415/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:35030fb71f632a328026f186811a789a80d5da5290bf98cd13525868cb60237a
+oid sha256:0c7cc84fba2c52f408b9eb573cd230db37c1e0fa3db3dcea22cc534fcebe1d0b
 size 15024

checkpoint-3415/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bfeffc751300cc0fcce62a27f2a5fac49d8b8fd66f36d343b964fecd42a3b8c2
+oid sha256:3fb3e74c9eb76eb7a05ea1e377f2dbc0e6d21bca119544e52b2f087d7ecc8737
 size 15024

checkpoint-3415/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:11135272c79e6370653667afef6b3109e5092c7ae64306ff5a555c49a3646453
+oid sha256:15ea3bbf2e86f4ae18fabcc4799d9109877d529e3efa3d3e9dc9dd2a2aec4fdd
 size 15024

checkpoint-3415/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-3415/tokenizer.model

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

checkpoint-3415/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-3415/trainer_state.json

@@ -9,423 +9,375 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.29282576866764276,
-      "grad_norm": 0.2425622195005417,
+      "epoch": 0.29,
       "learning_rate": 0.0005,
-      "loss": 1.972,
+      "loss": 1.426,
       "step": 100
     },
     {
-      "epoch": 0.5856515373352855,
-      "grad_norm": 0.24936461448669434,
+      "epoch": 0.59,
       "learning_rate": 0.0005,
-      "loss": 1.954,
+      "loss": 1.3981,
       "step": 200
     },
     {
-      "epoch": 0.8784773060029283,
-      "grad_norm": 0.28534355759620667,
+      "epoch": 0.88,
       "learning_rate": 0.0005,
-      "loss": 1.9636,
+      "loss": 1.3903,
       "step": 300
     },
     {
-      "epoch": 0.9985358711566618,
-      "eval_accuracy": 0.4142051282051282,
-      "eval_loss": 4.36731481552124,
-      "eval_runtime": 5.2053,
-      "eval_samples_per_second": 96.056,
-      "eval_steps_per_second": 12.103,
+      "epoch": 1.0,
+      "eval_accuracy": 0.4563589743589744,
+      "eval_loss": 3.9959237575531006,
+      "eval_runtime": 4.4486,
+      "eval_samples_per_second": 112.395,
+      "eval_steps_per_second": 14.162,
       "step": 341
     },
     {
-      "epoch": 0.9985358711566618,
-      "eval_bleu": 0.16271683142618718,
+      "epoch": 1.0,
+      "eval_bleu": 0.2575532874585898,
       "eval_exact_match": 0.0,
-      "eval_prefix_exact_match": 0.032,
+      "eval_prefix_exact_match": 0.028,
       "step": 341
     },
     {
-      "epoch": 1.171303074670571,
-      "grad_norm": 0.39824551343917847,
+      "epoch": 1.17,
       "learning_rate": 0.0005,
-      "loss": 1.8098,
+      "loss": 1.2759,
       "step": 400
     },
     {
-      "epoch": 1.4641288433382138,
-      "grad_norm": 0.4451245069503784,
+      "epoch": 1.46,
       "learning_rate": 0.0005,
-      "loss": 1.6717,
+      "loss": 1.1891,
       "step": 500
     },
     {
-      "epoch": 1.7569546120058566,
-      "grad_norm": 0.49089929461479187,
+      "epoch": 1.76,
       "learning_rate": 0.0005,
-      "loss": 1.7006,
+      "loss": 1.2,
       "step": 600
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.43338461538461537,
-      "eval_loss": 4.586850643157959,
-      "eval_runtime": 5.9085,
-      "eval_samples_per_second": 84.624,
-      "eval_steps_per_second": 10.663,
+      "eval_accuracy": 0.43964102564102564,
+      "eval_loss": 4.510287284851074,
+      "eval_runtime": 4.5647,
+      "eval_samples_per_second": 109.537,
+      "eval_steps_per_second": 13.802,
       "step": 683
     },
     {
       "epoch": 2.0,
-      "eval_bleu": 0.2955390058097942,
+      "eval_bleu": 0.2948780175737383,
       "eval_exact_match": 0.0,
-      "eval_prefix_exact_match": 0.036,
+      "eval_prefix_exact_match": 0.038,
       "step": 683
     },
     {
-      "epoch": 2.049780380673499,
-      "grad_norm": 0.6150477528572083,
+      "epoch": 2.05,
       "learning_rate": 0.0005,
-      "loss": 1.6383,
+      "loss": 1.1443,
       "step": 700
     },
     {
-      "epoch": 2.342606149341142,
-      "grad_norm": 0.5455761551856995,
+      "epoch": 2.34,
       "learning_rate": 0.0005,
-      "loss": 1.2533,
+      "loss": 0.8689,
       "step": 800
     },
     {
-      "epoch": 2.6354319180087846,
-      "grad_norm": 0.6252492666244507,
+      "epoch": 2.64,
       "learning_rate": 0.0005,
-      "loss": 1.3256,
+      "loss": 0.8937,
       "step": 900
     },
     {
-      "epoch": 2.9282576866764276,
-      "grad_norm": 0.6721358299255371,
+      "epoch": 2.93,
       "learning_rate": 0.0005,
-      "loss": 1.3696,
+      "loss": 0.9155,
       "step": 1000
     },
     {
-      "epoch": 2.998535871156662,
-      "eval_accuracy": 0.42538461538461536,
-      "eval_loss": 4.720473766326904,
-      "eval_runtime": 5.6052,
-      "eval_samples_per_second": 89.204,
-      "eval_steps_per_second": 11.24,
+      "epoch": 3.0,
+      "eval_accuracy": 0.42846153846153845,
+      "eval_loss": 4.8750715255737305,
+      "eval_runtime": 4.9742,
+      "eval_samples_per_second": 100.519,
+      "eval_steps_per_second": 12.665,
       "step": 1024
     },
     {
-      "epoch": 2.998535871156662,
-      "eval_bleu": 0.22572800415324107,
-      "eval_exact_match": 0.004,
-      "eval_prefix_exact_match": 0.044,
+      "epoch": 3.0,
+      "eval_bleu": 0.3312576952091117,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.042,
       "step": 1024
     },
     {
-      "epoch": 3.22108345534407,
-      "grad_norm": 0.7449648380279541,
+      "epoch": 3.22,
       "learning_rate": 0.0005,
-      "loss": 1.0137,
+      "loss": 0.6588,
       "step": 1100
     },
     {
-      "epoch": 3.513909224011713,
-      "grad_norm": 0.7608182430267334,
+      "epoch": 3.51,
       "learning_rate": 0.0005,
-      "loss": 0.9535,
+      "loss": 0.6077,
       "step": 1200
     },
     {
-      "epoch": 3.8067349926793557,
-      "grad_norm": 0.8320977687835693,
+      "epoch": 3.81,
       "learning_rate": 0.0005,
-      "loss": 1.0297,
+      "loss": 0.6446,
       "step": 1300
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.41717948717948716,
-      "eval_loss": 4.783276557922363,
-      "eval_runtime": 5.4398,
-      "eval_samples_per_second": 91.915,
-      "eval_steps_per_second": 11.581,
+      "eval_accuracy": 0.43256410256410255,
+      "eval_loss": 4.817827224731445,
+      "eval_runtime": 4.8888,
+      "eval_samples_per_second": 102.275,
+      "eval_steps_per_second": 12.887,
       "step": 1366
     },
     {
       "epoch": 4.0,
-      "eval_bleu": 0.28487206250718533,
-      "eval_exact_match": 0.008,
-      "eval_prefix_exact_match": 0.034,
+      "eval_bleu": 0.31288072457553184,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.06,
       "step": 1366
     },
     {
-      "epoch": 4.099560761346998,
-      "grad_norm": 0.7788870334625244,
+      "epoch": 4.1,
       "learning_rate": 0.0005,
-      "loss": 0.9222,
+      "loss": 0.5701,
       "step": 1400
     },
     {
-      "epoch": 4.392386530014641,
-      "grad_norm": 0.7981182336807251,
+      "epoch": 4.39,
       "learning_rate": 0.0005,
-      "loss": 0.6733,
+      "loss": 0.3936,
       "step": 1500
     },
     {
-      "epoch": 4.685212298682284,
-      "grad_norm": 0.938538670539856,
+      "epoch": 4.69,
       "learning_rate": 0.0005,
-      "loss": 0.7366,
+      "loss": 0.4278,
       "step": 1600
     },
     {
-      "epoch": 4.978038067349927,
-      "grad_norm": 0.9696006178855896,
+      "epoch": 4.98,
       "learning_rate": 0.0005,
-      "loss": 0.7991,
+      "loss": 0.455,
       "step": 1700
     },
     {
-      "epoch": 4.998535871156662,
-      "eval_accuracy": 0.4169230769230769,
-      "eval_loss": 4.9334001541137695,
-      "eval_runtime": 5.1107,
-      "eval_samples_per_second": 97.835,
-      "eval_steps_per_second": 12.327,
+      "epoch": 5.0,
+      "eval_accuracy": 0.4404102564102564,
+      "eval_loss": 4.9433770179748535,
+      "eval_runtime": 4.9389,
+      "eval_samples_per_second": 101.238,
+      "eval_steps_per_second": 12.756,
       "step": 1707
     },
     {
-      "epoch": 4.998535871156662,
-      "eval_bleu": 0.2463806011037282,
-      "eval_exact_match": 0.006,
-      "eval_prefix_exact_match": 0.034,
+      "epoch": 5.0,
+      "eval_bleu": 0.2870610839582904,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.056,
       "step": 1707
     },
     {
-      "epoch": 5.270863836017569,
-      "grad_norm": 0.8532861471176147,
+      "epoch": 5.27,
       "learning_rate": 0.0005,
-      "loss": 0.5011,
+      "loss": 0.268,
       "step": 1800
     },
     {
-      "epoch": 5.563689604685212,
-      "grad_norm": 0.8303696513175964,
+      "epoch": 5.56,
       "learning_rate": 0.0005,
-      "loss": 0.5186,
+      "loss": 0.2804,
       "step": 1900
     },
     {
-      "epoch": 5.856515373352855,
-      "grad_norm": 1.0248757600784302,
+      "epoch": 5.86,
       "learning_rate": 0.0005,
-      "loss": 0.5801,
+      "loss": 0.3104,
       "step": 2000
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.4133846153846154,
-      "eval_loss": 5.196371078491211,
-      "eval_runtime": 5.382,
-      "eval_samples_per_second": 92.902,
-      "eval_steps_per_second": 11.706,
+      "eval_accuracy": 0.43128205128205127,
+      "eval_loss": 5.222636699676514,
+      "eval_runtime": 4.1272,
+      "eval_samples_per_second": 121.147,
+      "eval_steps_per_second": 15.265,
       "step": 2049
     },
     {
       "epoch": 6.0,
-      "eval_bleu": 0.24094851152622576,
-      "eval_exact_match": 0.008,
-      "eval_prefix_exact_match": 0.024,
+      "eval_bleu": 0.29524211120433264,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.056,
       "step": 2049
     },
     {
-      "epoch": 6.149341142020498,
-      "grad_norm": 0.8996496200561523,
+      "epoch": 6.15,
       "learning_rate": 0.0005,
-      "loss": 0.4772,
+      "loss": 0.2554,
       "step": 2100
     },
     {
-      "epoch": 6.44216691068814,
-      "grad_norm": 0.9010413885116577,
+      "epoch": 6.44,
       "learning_rate": 0.0005,
-      "loss": 0.377,
+      "loss": 0.1992,
       "step": 2200
     },
     {
-      "epoch": 6.734992679355783,
-      "grad_norm": 0.9476897716522217,
+      "epoch": 6.73,
       "learning_rate": 0.0005,
-      "loss": 0.4228,
+      "loss": 0.2187,
       "step": 2300
     },
     {
-      "epoch": 6.998535871156662,
-      "eval_accuracy": 0.4101025641025641,
-      "eval_loss": 5.403963565826416,
-      "eval_runtime": 5.2963,
-      "eval_samples_per_second": 94.406,
-      "eval_steps_per_second": 11.895,
+      "epoch": 7.0,
+      "eval_accuracy": 0.4295897435897436,
+      "eval_loss": 5.163324356079102,
+      "eval_runtime": 4.1651,
+      "eval_samples_per_second": 120.046,
+      "eval_steps_per_second": 15.126,
       "step": 2390
     },
     {
-      "epoch": 6.998535871156662,
-      "eval_bleu": 0.20794741740418707,
+      "epoch": 7.0,
+      "eval_bleu": 0.20515604337519538,
       "eval_exact_match": 0.0,
-      "eval_prefix_exact_match": 0.024,
+      "eval_prefix_exact_match": 0.05,
       "step": 2390
     },
     {
-      "epoch": 7.027818448023426,
-      "grad_norm": 0.912433922290802,
+      "epoch": 7.03,
       "learning_rate": 0.0005,
-      "loss": 0.4456,
+      "loss": 0.2325,
       "step": 2400
     },
     {
-      "epoch": 7.320644216691068,
-      "grad_norm": 0.8055697679519653,
+      "epoch": 7.32,
       "learning_rate": 0.0005,
-      "loss": 0.2907,
+      "loss": 0.1573,
       "step": 2500
     },
     {
-      "epoch": 7.613469985358711,
-      "grad_norm": 0.9217767119407654,
+      "epoch": 7.61,
       "learning_rate": 0.0005,
-      "loss": 0.3285,
+      "loss": 0.1733,
       "step": 2600
     },
     {
-      "epoch": 7.906295754026354,
-      "grad_norm": 0.9473926424980164,
+      "epoch": 7.91,
       "learning_rate": 0.0005,
-      "loss": 0.3691,
+      "loss": 0.1903,
       "step": 2700
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.40984615384615386,
-      "eval_loss": 5.655345439910889,
-      "eval_runtime": 5.2073,
-      "eval_samples_per_second": 96.019,
-      "eval_steps_per_second": 12.098,
+      "eval_accuracy": 0.43774358974358973,
+      "eval_loss": 5.074286460876465,
+      "eval_runtime": 4.481,
+      "eval_samples_per_second": 111.581,
+      "eval_steps_per_second": 14.059,
       "step": 2732
     },
     {
       "epoch": 8.0,
-      "eval_bleu": 0.1614197770036361,
+      "eval_bleu": 0.21778472883103692,
       "eval_exact_match": 0.0,
-      "eval_prefix_exact_match": 0.028,
+      "eval_prefix_exact_match": 0.042,
       "step": 2732
     },
     {
-      "epoch": 8.199121522693996,
-      "grad_norm": 0.7216379642486572,
+      "epoch": 8.2,
       "learning_rate": 0.0005,
-      "loss": 0.296,
+      "loss": 0.159,
       "step": 2800
     },
     {
-      "epoch": 8.49194729136164,
-      "grad_norm": 0.8579537868499756,
+      "epoch": 8.49,
       "learning_rate": 0.0005,
-      "loss": 0.2711,
+      "loss": 0.1496,
       "step": 2900
     },
     {
-      "epoch": 8.784773060029282,
-      "grad_norm": 0.8787992000579834,
+      "epoch": 8.78,
       "learning_rate": 0.0005,
-      "loss": 0.3052,
+      "loss": 0.1639,
       "step": 3000
     },
     {
-      "epoch": 8.998535871156662,
-      "eval_accuracy": 0.41215384615384615,
-      "eval_loss": 5.5592756271362305,
-      "eval_runtime": 5.4045,
-      "eval_samples_per_second": 92.515,
-      "eval_steps_per_second": 11.657,
+      "epoch": 9.0,
+      "eval_accuracy": 0.4339487179487179,
+      "eval_loss": 5.249066352844238,
+      "eval_runtime": 4.2806,
+      "eval_samples_per_second": 116.805,
+      "eval_steps_per_second": 14.717,
       "step": 3073
     },
     {
-      "epoch": 8.998535871156662,
-      "eval_bleu": 0.24598687767703112,
-      "eval_exact_match": 0.002,
-      "eval_prefix_exact_match": 0.018,
+      "epoch": 9.0,
+      "eval_bleu": 0.2259404736504329,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.046,
       "step": 3073
     },
     {
-      "epoch": 9.077598828696924,
-      "grad_norm": 0.73407381772995,
+      "epoch": 9.08,
       "learning_rate": 0.0005,
-      "loss": 0.3065,
+      "loss": 0.1678,
       "step": 3100
     },
     {
-      "epoch": 9.370424597364568,
-      "grad_norm": 0.7941272854804993,
+      "epoch": 9.37,
       "learning_rate": 0.0005,
-      "loss": 0.2348,
+      "loss": 0.1376,
       "step": 3200
     },
     {
-      "epoch": 9.66325036603221,
-      "grad_norm": 0.8868365287780762,
+      "epoch": 9.66,
       "learning_rate": 0.0005,
-      "loss": 0.2645,
+      "loss": 0.1522,
       "step": 3300
     },
     {
-      "epoch": 9.956076134699854,
-      "grad_norm": 0.9728419184684753,
+      "epoch": 9.96,
       "learning_rate": 0.0005,
-      "loss": 0.2993,
+      "loss": 0.1685,
       "step": 3400
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.4115384615384615,
-      "eval_loss": 5.659451484680176,
-      "eval_runtime": 5.4123,
-      "eval_samples_per_second": 92.383,
-      "eval_steps_per_second": 11.64,
+      "eval_accuracy": 0.4355897435897436,
+      "eval_loss": 5.167659282684326,
+      "eval_runtime": 4.9323,
+      "eval_samples_per_second": 101.372,
+      "eval_steps_per_second": 12.773,
       "step": 3415
     },
     {
       "epoch": 10.0,
-      "eval_bleu": 0.23027305956234542,
-      "eval_exact_match": 0.004,
-      "eval_prefix_exact_match": 0.024,
+      "eval_bleu": 0.22532102244298752,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.042,
       "step": 3415
     }
   ],
   "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": 2.932484301159137e+17,
-  "train_batch_size": 1,
+  "total_flos": 6.763068176298148e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-3415/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e09109d7fc8f49ffd654f0e1286ef00d8840620753fec453a60a103c55187142
-size 5304
+oid sha256:90c35679e38c5a16df008bb0194b7282479bcd6cc7127dd7f5dab851cf151c22
+size 4728