Training in progress, epoch 10, checkpoint

checkpoint-1875/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-1875/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fcf5bd32a64a136052e2c18776d5067c6de7d1ee645292f76d07552c7809e690
+size 143269386

checkpoint-1875/added_tokens.json

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

checkpoint-1875/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1040061b8814349eab55049f8f301509b4a042d999f356e062106e4ddce71f04
-size 224537202
+oid sha256:27db7fb03ac6a4ce45fae019660d31bb16e716b662305a9c5df6f38daac744c6
+size 286585234

checkpoint-1875/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d6dbcb0c7fb4a79d52fa3e70bd52ed1019b547168ff6c7ae9f8df489fbfe0a80
+oid sha256:a72d9e8580ef33ad0869337100d05c6253790ee8e202a392fa8c908ed628e86c
 size 15024

checkpoint-1875/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a4a006c1e1cd2878100312bd542a4cdaaecf7185bff24f4fc9f3752f1fecb959
+oid sha256:f1ecfc27c9a2b7f7ba9d1dae8c355bd7eb790644a70189746dac49a1f7de2d64
 size 15024

checkpoint-1875/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8ff8768b15de07008bd9cf52340bdaa64dbfcbf6fc99e0c40d033aa1d893ec66
+oid sha256:b98e5e256a50b30839540efc512b9b7aafcca6c2375000ea21907f364e363d42
 size 15024

checkpoint-1875/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9b1f3fd04e3f7b4f50f28678fe627b7dcab3d610acfda2b1d411f93fad4cf6a9
+oid sha256:a92839ff1fe3e5bd96af1505e89c77f7f715015b28c84cde713be99caba57c2c
 size 15024

checkpoint-1875/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-1875/tokenizer.model

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

checkpoint-1875/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-1875/trainer_state.json

@@ -9,341 +9,309 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.5333333333333333,
-      "grad_norm": 0.275871217250824,
+      "epoch": 0.53,
       "learning_rate": 3e-05,
-      "loss": 1.8545,
+      "loss": 1.4597,
       "step": 100
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_accuracy": 0.6037399103139014,
-      "eval_loss": 1.7123007774353027,
-      "eval_runtime": 6.2976,
-      "eval_samples_per_second": 79.395,
-      "eval_steps_per_second": 10.004,
+      "epoch": 1.0,
+      "eval_accuracy": 0.6618980392156862,
+      "eval_loss": 1.2718806266784668,
+      "eval_runtime": 7.6194,
+      "eval_samples_per_second": 65.622,
+      "eval_steps_per_second": 8.268,
       "step": 187
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_exact_match": 14.4,
-      "eval_f1": 22.843015873015876,
-      "eval_qa_bleu": 9.471299126615676,
-      "eval_qa_exact_match": 0.112,
-      "eval_recite_bleu": 12.267251092702137,
+      "epoch": 1.0,
+      "eval_exact_match": 22.4,
+      "eval_f1": 30.941064425770318,
+      "eval_qa_bleu": 7.097403515221558,
+      "eval_qa_exact_match": 0.196,
+      "eval_recite_bleu": 13.258373740102225,
       "eval_recite_exact_match": 0.0,
       "step": 187
     },
     {
-      "epoch": 1.0666666666666667,
-      "grad_norm": 0.32765188813209534,
+      "epoch": 1.07,
       "learning_rate": 3e-05,
-      "loss": 1.7408,
+      "loss": 1.2905,
       "step": 200
     },
     {
       "epoch": 1.6,
-      "grad_norm": 0.3532409369945526,
       "learning_rate": 3e-05,
-      "loss": 1.7063,
+      "loss": 1.2157,
       "step": 300
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.6058565022421525,
-      "eval_loss": 1.69453763961792,
-      "eval_runtime": 6.5928,
-      "eval_samples_per_second": 75.841,
-      "eval_steps_per_second": 9.556,
+      "eval_accuracy": 0.6670039215686274,
+      "eval_loss": 1.206207036972046,
+      "eval_runtime": 7.6793,
+      "eval_samples_per_second": 65.11,
+      "eval_steps_per_second": 8.204,
       "step": 375
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 14.4,
-      "eval_f1": 23.29731601731603,
-      "eval_qa_bleu": 9.4408132575913,
-      "eval_qa_exact_match": 0.11,
-      "eval_recite_bleu": 13.027794116434837,
+      "eval_exact_match": 3.2,
+      "eval_f1": 3.96976236704394,
+      "eval_qa_bleu": 0.20484366294597936,
+      "eval_qa_exact_match": 0.026,
+      "eval_recite_bleu": 2.5778714769111824e-05,
       "eval_recite_exact_match": 0.0,
       "step": 375
     },
     {
-      "epoch": 2.1333333333333333,
-      "grad_norm": 0.3777780532836914,
+      "epoch": 2.13,
       "learning_rate": 3e-05,
-      "loss": 1.7094,
+      "loss": 1.1852,
       "step": 400
     },
     {
-      "epoch": 2.6666666666666665,
-      "grad_norm": 0.3971036374568939,
+      "epoch": 2.67,
       "learning_rate": 3e-05,
-      "loss": 1.6702,
+      "loss": 1.1861,
       "step": 500
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_accuracy": 0.6065919282511211,
-      "eval_loss": 1.6851236820220947,
-      "eval_runtime": 6.3104,
-      "eval_samples_per_second": 79.234,
-      "eval_steps_per_second": 9.984,
+      "epoch": 3.0,
+      "eval_accuracy": 0.667513725490196,
+      "eval_loss": 1.19828462600708,
+      "eval_runtime": 7.1425,
+      "eval_samples_per_second": 70.004,
+      "eval_steps_per_second": 8.82,
       "step": 562
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_exact_match": 14.0,
-      "eval_f1": 23.306825396825417,
-      "eval_qa_bleu": 8.683248465366908,
-      "eval_qa_exact_match": 0.11,
-      "eval_recite_bleu": 12.39336052958371,
-      "eval_recite_exact_match": 0.006,
+      "epoch": 3.0,
+      "eval_exact_match": 4.4,
+      "eval_f1": 6.333660342551587,
+      "eval_qa_bleu": 0.522391008398391,
+      "eval_qa_exact_match": 0.028,
+      "eval_recite_bleu": 0.027569975851374308,
+      "eval_recite_exact_match": 0.0,
       "step": 562
     },
     {
       "epoch": 3.2,
-      "grad_norm": 0.4782978594303131,
       "learning_rate": 3e-05,
-      "loss": 1.6574,
+      "loss": 1.1722,
       "step": 600
     },
     {
-      "epoch": 3.7333333333333334,
-      "grad_norm": 0.52582186460495,
+      "epoch": 3.73,
       "learning_rate": 3e-05,
-      "loss": 1.6356,
+      "loss": 1.1445,
       "step": 700
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.6074977578475337,
-      "eval_loss": 1.682567834854126,
-      "eval_runtime": 6.65,
-      "eval_samples_per_second": 75.188,
-      "eval_steps_per_second": 9.474,
+      "eval_accuracy": 0.6672,
+      "eval_loss": 1.197709560394287,
+      "eval_runtime": 7.7427,
+      "eval_samples_per_second": 64.577,
+      "eval_steps_per_second": 8.137,
       "step": 750
     },
     {
       "epoch": 4.0,
-      "eval_exact_match": 15.0,
-      "eval_f1": 23.366825396825416,
-      "eval_qa_bleu": 9.05994637010322,
-      "eval_qa_exact_match": 0.114,
-      "eval_recite_bleu": 13.060415058384661,
-      "eval_recite_exact_match": 0.006,
+      "eval_exact_match": 7.4,
+      "eval_f1": 10.950016966795278,
+      "eval_qa_bleu": 0.6700083846206139,
+      "eval_qa_exact_match": 0.052,
+      "eval_recite_bleu": 1.259776140465531,
+      "eval_recite_exact_match": 0.0,
       "step": 750
     },
     {
-      "epoch": 4.266666666666667,
-      "grad_norm": 0.6418241262435913,
+      "epoch": 4.27,
       "learning_rate": 3e-05,
-      "loss": 1.6005,
+      "loss": 1.1018,
       "step": 800
     },
     {
       "epoch": 4.8,
-      "grad_norm": 0.6805307269096375,
       "learning_rate": 3e-05,
-      "loss": 1.5775,
+      "loss": 1.1,
       "step": 900
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_accuracy": 0.6071031390134529,
-      "eval_loss": 1.69106125831604,
-      "eval_runtime": 5.6598,
-      "eval_samples_per_second": 88.343,
-      "eval_steps_per_second": 11.131,
+      "epoch": 5.0,
+      "eval_accuracy": 0.6664941176470588,
+      "eval_loss": 1.2078713178634644,
+      "eval_runtime": 7.7721,
+      "eval_samples_per_second": 64.333,
+      "eval_steps_per_second": 8.106,
       "step": 937
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_exact_match": 15.2,
-      "eval_f1": 23.38539682539684,
-      "eval_qa_bleu": 11.854591678140883,
-      "eval_qa_exact_match": 0.112,
-      "eval_recite_bleu": 12.694062454678122,
-      "eval_recite_exact_match": 0.004,
+      "epoch": 5.0,
+      "eval_exact_match": 6.6,
+      "eval_f1": 10.12048028049545,
+      "eval_qa_bleu": 0.8314625442591895,
+      "eval_qa_exact_match": 0.05,
+      "eval_recite_bleu": 0.8030246920772521,
+      "eval_recite_exact_match": 0.0,
       "step": 937
     },
     {
-      "epoch": 5.333333333333333,
-      "grad_norm": 0.8362048864364624,
+      "epoch": 5.33,
       "learning_rate": 3e-05,
-      "loss": 1.5173,
+      "loss": 1.0666,
       "step": 1000
     },
     {
-      "epoch": 5.866666666666667,
-      "grad_norm": 0.8521270155906677,
+      "epoch": 5.87,
       "learning_rate": 3e-05,
-      "loss": 1.529,
+      "loss": 1.0439,
       "step": 1100
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.6069417040358744,
-      "eval_loss": 1.7066757678985596,
-      "eval_runtime": 6.2922,
-      "eval_samples_per_second": 79.464,
-      "eval_steps_per_second": 10.012,
+      "eval_accuracy": 0.6649725490196079,
+      "eval_loss": 1.22383713722229,
+      "eval_runtime": 7.6639,
+      "eval_samples_per_second": 65.241,
+      "eval_steps_per_second": 8.22,
       "step": 1125
     },
     {
       "epoch": 6.0,
-      "eval_exact_match": 15.2,
-      "eval_f1": 24.04682539682541,
-      "eval_qa_bleu": 12.150293484705452,
-      "eval_qa_exact_match": 0.106,
-      "eval_recite_bleu": 12.732083161926754,
-      "eval_recite_exact_match": 0.002,
+      "eval_exact_match": 9.0,
+      "eval_f1": 14.19791549030679,
+      "eval_qa_bleu": 2.7960558519683327,
+      "eval_qa_exact_match": 0.07,
+      "eval_recite_bleu": 2.082318255561628,
+      "eval_recite_exact_match": 0.0,
       "step": 1125
     },
     {
       "epoch": 6.4,
-      "grad_norm": 0.9372308254241943,
       "learning_rate": 3e-05,
-      "loss": 1.4613,
+      "loss": 1.0029,
       "step": 1200
     },
     {
-      "epoch": 6.933333333333334,
-      "grad_norm": 1.0236693620681763,
+      "epoch": 6.93,
       "learning_rate": 3e-05,
-      "loss": 1.457,
+      "loss": 0.9888,
       "step": 1300
     },
     {
-      "epoch": 6.997333333333334,
-      "eval_accuracy": 0.6055515695067265,
-      "eval_loss": 1.72792649269104,
-      "eval_runtime": 6.3278,
-      "eval_samples_per_second": 79.016,
-      "eval_steps_per_second": 9.956,
+      "epoch": 7.0,
+      "eval_accuracy": 0.6638117647058823,
+      "eval_loss": 1.2457345724105835,
+      "eval_runtime": 6.7577,
+      "eval_samples_per_second": 73.99,
+      "eval_steps_per_second": 9.323,
       "step": 1312
     },
     {
-      "epoch": 6.997333333333334,
-      "eval_exact_match": 15.0,
-      "eval_f1": 23.28492063492065,
-      "eval_qa_bleu": 9.849658422779212,
-      "eval_qa_exact_match": 0.114,
-      "eval_recite_bleu": 12.3997433702838,
-      "eval_recite_exact_match": 0.006,
+      "epoch": 7.0,
+      "eval_exact_match": 10.6,
+      "eval_f1": 15.30586844512479,
+      "eval_qa_bleu": 0.94556412718174,
+      "eval_qa_exact_match": 0.086,
+      "eval_recite_bleu": 3.1558565534273777,
+      "eval_recite_exact_match": 0.0,
       "step": 1312
     },
     {
-      "epoch": 7.466666666666667,
-      "grad_norm": 1.0773049592971802,
+      "epoch": 7.47,
       "learning_rate": 3e-05,
-      "loss": 1.3996,
+      "loss": 0.9453,
       "step": 1400
     },
     {
       "epoch": 8.0,
-      "grad_norm": 1.0705862045288086,
       "learning_rate": 3e-05,
-      "loss": 1.3907,
+      "loss": 0.9364,
       "step": 1500
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.6046188340807175,
-      "eval_loss": 1.7511730194091797,
-      "eval_runtime": 5.5895,
-      "eval_samples_per_second": 89.453,
-      "eval_steps_per_second": 11.271,
+      "eval_accuracy": 0.6615921568627451,
+      "eval_loss": 1.2816376686096191,
+      "eval_runtime": 7.5902,
+      "eval_samples_per_second": 65.875,
+      "eval_steps_per_second": 8.3,
       "step": 1500
     },
     {
       "epoch": 8.0,
-      "eval_exact_match": 13.0,
-      "eval_f1": 19.93238095238096,
-      "eval_qa_bleu": 11.93942669425287,
-      "eval_qa_exact_match": 0.098,
-      "eval_recite_bleu": 11.949414001720045,
-      "eval_recite_exact_match": 0.006,
+      "eval_exact_match": 9.2,
+      "eval_f1": 12.989213707726288,
+      "eval_qa_bleu": 1.9797940311992752,
+      "eval_qa_exact_match": 0.074,
+      "eval_recite_bleu": 2.551621610315755,
+      "eval_recite_exact_match": 0.0,
       "step": 1500
     },
     {
-      "epoch": 8.533333333333333,
-      "grad_norm": 1.2424529790878296,
+      "epoch": 8.53,
       "learning_rate": 3e-05,
-      "loss": 1.3309,
+      "loss": 0.8889,
       "step": 1600
     },
     {
-      "epoch": 8.997333333333334,
-      "eval_accuracy": 0.6025022421524664,
-      "eval_loss": 1.7774019241333008,
-      "eval_runtime": 6.4013,
-      "eval_samples_per_second": 78.109,
-      "eval_steps_per_second": 9.842,
+      "epoch": 9.0,
+      "eval_accuracy": 0.6606196078431372,
+      "eval_loss": 1.3036144971847534,
+      "eval_runtime": 7.4845,
+      "eval_samples_per_second": 66.805,
+      "eval_steps_per_second": 8.417,
       "step": 1687
     },
     {
-      "epoch": 8.997333333333334,
-      "eval_exact_match": 13.6,
-      "eval_f1": 21.089523809523815,
-      "eval_qa_bleu": 10.094627644241111,
-      "eval_qa_exact_match": 0.094,
-      "eval_recite_bleu": 12.197681191541987,
-      "eval_recite_exact_match": 0.006,
+      "epoch": 9.0,
+      "eval_exact_match": 14.2,
+      "eval_f1": 19.66897045942666,
+      "eval_qa_bleu": 3.3530752528407004,
+      "eval_qa_exact_match": 0.112,
+      "eval_recite_bleu": 5.3293362575452425,
+      "eval_recite_exact_match": 0.002,
       "step": 1687
     },
     {
-      "epoch": 9.066666666666666,
-      "grad_norm": 1.535229206085205,
+      "epoch": 9.07,
       "learning_rate": 3e-05,
-      "loss": 1.3336,
+      "loss": 0.8853,
       "step": 1700
     },
     {
       "epoch": 9.6,
-      "grad_norm": 1.4936225414276123,
       "learning_rate": 3e-05,
-      "loss": 1.2841,
+      "loss": 0.8373,
       "step": 1800
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.6012645739910314,
-      "eval_loss": 1.804250717163086,
-      "eval_runtime": 6.385,
-      "eval_samples_per_second": 78.309,
-      "eval_steps_per_second": 9.867,
+      "eval_accuracy": 0.6587450980392157,
+      "eval_loss": 1.3337457180023193,
+      "eval_runtime": 6.6361,
+      "eval_samples_per_second": 75.345,
+      "eval_steps_per_second": 9.493,
       "step": 1875
     },
     {
       "epoch": 10.0,
-      "eval_exact_match": 13.2,
-      "eval_f1": 21.172832722832737,
-      "eval_qa_bleu": 9.841075001407239,
-      "eval_qa_exact_match": 0.096,
-      "eval_recite_bleu": 11.691398764835105,
-      "eval_recite_exact_match": 0.004,
+      "eval_exact_match": 14.4,
+      "eval_f1": 21.20886352105531,
+      "eval_qa_bleu": 2.50339041197013,
+      "eval_qa_exact_match": 0.106,
+      "eval_recite_bleu": 6.648771889251858,
+      "eval_recite_exact_match": 0.0,
       "step": 1875
     }
   ],
   "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": 2.0683971187913523e+17,
-  "train_batch_size": 2,
+  "total_flos": 4.3444450722893005e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-1875/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9e9c21c8367f69c99b21b20aa3dbf722814d1b222850e23dacbe471b235b2dd4
-size 5304
+oid sha256:9347580f7d2a7b3957c29c0f67ee2342d50aa99cfc14c2f59eba6127909f6932
+size 4728