Training in progress, epoch 14, checkpoint

checkpoint-2812/README.md

@@ -0,0 +1,9 @@
+---
+library_name: peft
+---
+## Training procedure
+
+### Framework versions
+
+
+- PEFT 0.5.0

checkpoint-2812/adapter_config.json

@@ -0,0 +1,25 @@
+{
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen1.5-4B",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "lora_alpha": 16,
+  "lora_dropout": 0.05,
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 16,
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "k_proj",
+    "v_proj",
+    "down_proj",
+    "gate_proj",
+    "up_proj"
+  ],
+  "task_type": "CAUSAL_LM"
+}

checkpoint-2812/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cba77c8aff685483dcfe41f1e9d44e02db77bd31778c759046913fea7c5821fc
+size 112130416

checkpoint-2812/added_tokens.json

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

checkpoint-2812/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed64979889d56ebc92913b6bedcf5371871b72ccb1e9493e2f69e1174936c1c5
+size 224537202

checkpoint-2812/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d989c0390eb74b25d91cdb8e17d434f8d3f8953898a73d7b1d729aeed975731
+size 15024

checkpoint-2812/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f45aeeb8ec465787582679dcddcf2954654c6e5f775e6c6f5f731e5f43e324ae
+size 15024

checkpoint-2812/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec4248a7659af7176ddadd6e136e125b5375ba85aa24593cd6af6c20427faae6
+size 15024

checkpoint-2812/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82e8bea9badb98c92315ac2c04051ebcd96a6cd0a068d3e34e6c4df411ebe957
+size 15024

checkpoint-2812/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e71be2303e15d26babc549166690acc89078d04fd61f2740c9977dbde9d1740
+size 1064

checkpoint-2812/special_tokens_map.json

@@ -0,0 +1,14 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<|endoftext|>"
+}

checkpoint-2812/tokenizer_config.json

@@ -0,0 +1,43 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "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 %}",
+  "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
+}

checkpoint-2812/trainer_state.json

@@ -0,0 +1,514 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 14.997333333333334,
+  "eval_steps": 500,
+  "global_step": 2812,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.5333333333333333,
+      "grad_norm": 0.275871217250824,
+      "learning_rate": 3e-05,
+      "loss": 1.8545,
+      "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,
+      "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,
+      "eval_recite_exact_match": 0.0,
+      "step": 187
+    },
+    {
+      "epoch": 1.0666666666666667,
+      "grad_norm": 0.32765188813209534,
+      "learning_rate": 3e-05,
+      "loss": 1.7408,
+      "step": 200
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 0.3532409369945526,
+      "learning_rate": 3e-05,
+      "loss": 1.7063,
+      "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,
+      "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_recite_exact_match": 0.0,
+      "step": 375
+    },
+    {
+      "epoch": 2.1333333333333333,
+      "grad_norm": 0.3777780532836914,
+      "learning_rate": 3e-05,
+      "loss": 1.7094,
+      "step": 400
+    },
+    {
+      "epoch": 2.6666666666666665,
+      "grad_norm": 0.3971036374568939,
+      "learning_rate": 3e-05,
+      "loss": 1.6702,
+      "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,
+      "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,
+      "step": 562
+    },
+    {
+      "epoch": 3.2,
+      "grad_norm": 0.4782978594303131,
+      "learning_rate": 3e-05,
+      "loss": 1.6574,
+      "step": 600
+    },
+    {
+      "epoch": 3.7333333333333334,
+      "grad_norm": 0.52582186460495,
+      "learning_rate": 3e-05,
+      "loss": 1.6356,
+      "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,
+      "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,
+      "step": 750
+    },
+    {
+      "epoch": 4.266666666666667,
+      "grad_norm": 0.6418241262435913,
+      "learning_rate": 3e-05,
+      "loss": 1.6005,
+      "step": 800
+    },
+    {
+      "epoch": 4.8,
+      "grad_norm": 0.6805307269096375,
+      "learning_rate": 3e-05,
+      "loss": 1.5775,
+      "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,
+      "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,
+      "step": 937
+    },
+    {
+      "epoch": 5.333333333333333,
+      "grad_norm": 0.8362048864364624,
+      "learning_rate": 3e-05,
+      "loss": 1.5173,
+      "step": 1000
+    },
+    {
+      "epoch": 5.866666666666667,
+      "grad_norm": 0.8521270155906677,
+      "learning_rate": 3e-05,
+      "loss": 1.529,
+      "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,
+      "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,
+      "step": 1125
+    },
+    {
+      "epoch": 6.4,
+      "grad_norm": 0.9372308254241943,
+      "learning_rate": 3e-05,
+      "loss": 1.4613,
+      "step": 1200
+    },
+    {
+      "epoch": 6.933333333333334,
+      "grad_norm": 1.0236693620681763,
+      "learning_rate": 3e-05,
+      "loss": 1.457,
+      "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,
+      "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,
+      "step": 1312
+    },
+    {
+      "epoch": 7.466666666666667,
+      "grad_norm": 1.0773049592971802,
+      "learning_rate": 3e-05,
+      "loss": 1.3996,
+      "step": 1400
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 1.0705862045288086,
+      "learning_rate": 3e-05,
+      "loss": 1.3907,
+      "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,
+      "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,
+      "step": 1500
+    },
+    {
+      "epoch": 8.533333333333333,
+      "grad_norm": 1.2424529790878296,
+      "learning_rate": 3e-05,
+      "loss": 1.3309,
+      "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,
+      "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,
+      "step": 1687
+    },
+    {
+      "epoch": 9.066666666666666,
+      "grad_norm": 1.535229206085205,
+      "learning_rate": 3e-05,
+      "loss": 1.3336,
+      "step": 1700
+    },
+    {
+      "epoch": 9.6,
+      "grad_norm": 1.4936225414276123,
+      "learning_rate": 3e-05,
+      "loss": 1.2841,
+      "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,
+      "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,
+      "step": 1875
+    },
+    {
+      "epoch": 10.133333333333333,
+      "grad_norm": 1.6687729358673096,
+      "learning_rate": 3e-05,
+      "loss": 1.2562,
+      "step": 1900
+    },
+    {
+      "epoch": 10.666666666666666,
+      "grad_norm": 1.437538504600525,
+      "learning_rate": 3e-05,
+      "loss": 1.2308,
+      "step": 2000
+    },
+    {
+      "epoch": 10.997333333333334,
+      "eval_accuracy": 0.6000717488789238,
+      "eval_loss": 1.8528188467025757,
+      "eval_runtime": 6.6644,
+      "eval_samples_per_second": 75.026,
+      "eval_steps_per_second": 9.453,
+      "step": 2062
+    },
+    {
+      "epoch": 10.997333333333334,
+      "eval_exact_match": 12.6,
+      "eval_f1": 20.865213675213692,
+      "eval_qa_bleu": 7.784433449171618,
+      "eval_qa_exact_match": 0.108,
+      "eval_recite_bleu": 12.305413068460787,
+      "eval_recite_exact_match": 0.008,
+      "step": 2062
+    },
+    {
+      "epoch": 11.2,
+      "grad_norm": 1.708260416984558,
+      "learning_rate": 3e-05,
+      "loss": 1.2011,
+      "step": 2100
+    },
+    {
+      "epoch": 11.733333333333333,
+      "grad_norm": 1.7026830911636353,
+      "learning_rate": 3e-05,
+      "loss": 1.1722,
+      "step": 2200
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.5987892376681614,
+      "eval_loss": 1.8851206302642822,
+      "eval_runtime": 5.5764,
+      "eval_samples_per_second": 89.663,
+      "eval_steps_per_second": 11.298,
+      "step": 2250
+    },
+    {
+      "epoch": 12.0,
+      "eval_exact_match": 13.0,
+      "eval_f1": 21.17809523809525,
+      "eval_qa_bleu": 11.19937598521782,
+      "eval_qa_exact_match": 0.096,
+      "eval_recite_bleu": 12.500534642121808,
+      "eval_recite_exact_match": 0.01,
+      "step": 2250
+    },
+    {
+      "epoch": 12.266666666666667,
+      "grad_norm": 1.6281794309616089,
+      "learning_rate": 3e-05,
+      "loss": 1.151,
+      "step": 2300
+    },
+    {
+      "epoch": 12.8,
+      "grad_norm": 1.686490535736084,
+      "learning_rate": 3e-05,
+      "loss": 1.1354,
+      "step": 2400
+    },
+    {
+      "epoch": 12.997333333333334,
+      "eval_accuracy": 0.5979910313901345,
+      "eval_loss": 1.9114351272583008,
+      "eval_runtime": 5.5718,
+      "eval_samples_per_second": 89.737,
+      "eval_steps_per_second": 11.307,
+      "step": 2437
+    },
+    {
+      "epoch": 12.997333333333334,
+      "eval_exact_match": 13.8,
+      "eval_f1": 21.856666666666673,
+      "eval_qa_bleu": 10.206087008658956,
+      "eval_qa_exact_match": 0.116,
+      "eval_recite_bleu": 12.669262328097592,
+      "eval_recite_exact_match": 0.008,
+      "step": 2437
+    },
+    {
+      "epoch": 13.333333333333334,
+      "grad_norm": 1.8666125535964966,
+      "learning_rate": 3e-05,
+      "loss": 1.097,
+      "step": 2500
+    },
+    {
+      "epoch": 13.866666666666667,
+      "grad_norm": 1.9091932773590088,
+      "learning_rate": 3e-05,
+      "loss": 1.0793,
+      "step": 2600
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": 0.5960986547085202,
+      "eval_loss": 1.9584920406341553,
+      "eval_runtime": 6.3056,
+      "eval_samples_per_second": 79.294,
+      "eval_steps_per_second": 9.991,
+      "step": 2625
+    },
+    {
+      "epoch": 14.0,
+      "eval_exact_match": 10.6,
+      "eval_f1": 18.697936507936518,
+      "eval_qa_bleu": 6.4581555766931436,
+      "eval_qa_exact_match": 0.082,
+      "eval_recite_bleu": 12.843526576123885,
+      "eval_recite_exact_match": 0.01,
+      "step": 2625
+    },
+    {
+      "epoch": 14.4,
+      "grad_norm": 2.042712688446045,
+      "learning_rate": 3e-05,
+      "loss": 1.0376,
+      "step": 2700
+    },
+    {
+      "epoch": 14.933333333333334,
+      "grad_norm": 1.8875784873962402,
+      "learning_rate": 3e-05,
+      "loss": 1.037,
+      "step": 2800
+    },
+    {
+      "epoch": 14.997333333333334,
+      "eval_accuracy": 0.5948161434977578,
+      "eval_loss": 1.9966908693313599,
+      "eval_runtime": 6.0322,
+      "eval_samples_per_second": 82.888,
+      "eval_steps_per_second": 10.444,
+      "step": 2812
+    },
+    {
+      "epoch": 14.997333333333334,
+      "eval_exact_match": 12.8,
+      "eval_f1": 20.438917748917753,
+      "eval_qa_bleu": 10.19364322169534,
+      "eval_qa_exact_match": 0.104,
+      "eval_recite_bleu": 12.69742056629334,
+      "eval_recite_exact_match": 0.012,
+      "step": 2812
+    }
+  ],
+  "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": 3.1036655948057805e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-2812/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e9c21c8367f69c99b21b20aa3dbf722814d1b222850e23dacbe471b235b2dd4
+size 5304