Upload with huggingface_hub

README.md

@@ -0,0 +1,15 @@
+---
+license: apache-2.0
+tags:
+- canine
+- pretrained-on-english-language
+---
+
+### How to use
+
+Here is how to use this model:
+
+```python
+from transformers import CanineModel
+model = CanineModel.from_pretrained('mushfiqur11/<repo name>')
+```

all_results.json

@@ -0,0 +1,20 @@
+{
+    "epoch": 36.06,
+    "eval_accuracy": 0.9529491987725878,
+    "eval_loss": 0.37525418400764465,
+    "eval_runtime": 6.5718,
+    "eval_samples": 1659,
+    "eval_samples_per_second": 252.444,
+    "eval_steps_per_second": 31.651,
+    "test_accuracy": 0.953655094552639,
+    "test_loss": 0.38122233748435974,
+    "test_runtime": 6.6939,
+    "test_samples": 1684,
+    "test_samples_per_second": 251.572,
+    "test_steps_per_second": 31.521,
+    "train_loss": 0.05964354458153248,
+    "train_runtime": 2290.905,
+    "train_samples": 13304,
+    "train_samples_per_second": 209.524,
+    "train_steps_per_second": 6.548
+}

config.json

@@ -0,0 +1,67 @@
+{
+  "_name_or_path": "google/canine-s",
+  "architectures": [
+    "CanineForTokenClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 57344,
+  "downsampling_rate": 4,
+  "eos_token_id": 57345,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "ADJ",
+    "1": "ADP",
+    "2": "ADV",
+    "3": "AUX",
+    "4": "CCONJ",
+    "5": "DET",
+    "6": "INTJ",
+    "7": "NOUN",
+    "8": "NUM",
+    "9": "PART",
+    "10": "PRON",
+    "11": "PROPN",
+    "12": "PUNCT",
+    "13": "SCONJ",
+    "14": "SYM",
+    "15": "VERB",
+    "16": "X"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "ADJ": 0,
+    "ADP": 1,
+    "ADV": 2,
+    "AUX": 3,
+    "CCONJ": 4,
+    "DET": 5,
+    "INTJ": 6,
+    "NOUN": 7,
+    "NUM": 8,
+    "PART": 9,
+    "PRON": 10,
+    "PROPN": 11,
+    "PUNCT": 12,
+    "SCONJ": 13,
+    "SYM": 14,
+    "VERB": 15,
+    "X": 16
+  },
+  "layer_norm_eps": 1e-12,
+  "local_transformer_stride": 128,
+  "max_position_embeddings": 16384,
+  "model_type": "canine",
+  "num_attention_heads": 12,
+  "num_hash_buckets": 16384,
+  "num_hash_functions": 8,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "torch_dtype": "float32",
+  "transformers_version": "4.17.0",
+  "type_vocab_size": 16,
+  "upsampling_kernel_size": 4,
+  "use_cache": true
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b877b2875761f1d9d78f6f70d13678e5185f3e588a3405f63b74f83b9217c0e3
+size 528619377

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": {"content": "", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "eos_token": {"content": "", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "sep_token": {"content": "", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "pad_token": {"content": "\u0000", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "cls_token": {"content": "", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "mask_token": {"content": "", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true}}

tokenizer_config.json

@@ -0,0 +1 @@
+{"bos_token": {"content": "", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "eos_token": {"content": "", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "sep_token": {"content": "", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "cls_token": {"content": "", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "pad_token": {"content": "\u0000", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "mask_token": {"content": "", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "add_prefix_space": false, "model_max_length": 2048, "special_tokens_map_file": "/home/mrahma45/.cache/huggingface/transformers/00ffef2919c64d09945b6c06b1c03c4d18bf55f88013390d4443a1eb8c4f8a10.ab71f530366fe02e2834427e7b90198bfd0d573bc4279bfafdb2b95fe2b46dde", "name_or_path": "google/canine-s", "tokenizer_class": "CanineTokenizer"}

trainer_state.json

@@ -0,0 +1,1195 @@
+{
+  "best_metric": 0.9529491987725878,
+  "best_model_checkpoint": "/scratch/mrahma45/pixel/finetuned_models/canine/canine-base-finetuned-pos-ud-Hindi-HDTB/checkpoint-14000",
+  "epoch": 36.05769230769231,
+  "global_step": 15000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.24,
+      "learning_rate": 8e-05,
+      "loss": 1.6829,
+      "step": 100
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.946308724832215e-05,
+      "loss": 0.6655,
+      "step": 200
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 7.89261744966443e-05,
+      "loss": 0.4831,
+      "step": 300
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.838926174496645e-05,
+      "loss": 0.405,
+      "step": 400
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.78523489932886e-05,
+      "loss": 0.3386,
+      "step": 500
+    },
+    {
+      "epoch": 1.2,
+      "eval_accuracy": 0.8978292987839527,
+      "eval_loss": 0.3005571663379669,
+      "eval_runtime": 6.5562,
+      "eval_samples_per_second": 253.044,
+      "eval_steps_per_second": 31.726,
+      "step": 500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 7.731543624161075e-05,
+      "loss": 0.3137,
+      "step": 600
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 7.677852348993288e-05,
+      "loss": 0.2831,
+      "step": 700
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 7.624161073825503e-05,
+      "loss": 0.2711,
+      "step": 800
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 7.570469798657718e-05,
+      "loss": 0.2314,
+      "step": 900
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 7.516778523489933e-05,
+      "loss": 0.2131,
+      "step": 1000
+    },
+    {
+      "epoch": 2.4,
+      "eval_accuracy": 0.9229173769746563,
+      "eval_loss": 0.25128552317619324,
+      "eval_runtime": 6.5615,
+      "eval_samples_per_second": 252.839,
+      "eval_steps_per_second": 31.7,
+      "step": 1000
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 7.463087248322148e-05,
+      "loss": 0.2018,
+      "step": 1100
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 7.409395973154362e-05,
+      "loss": 0.2022,
+      "step": 1200
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 7.355704697986577e-05,
+      "loss": 0.1719,
+      "step": 1300
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 7.302013422818792e-05,
+      "loss": 0.1577,
+      "step": 1400
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 7.248322147651007e-05,
+      "loss": 0.1561,
+      "step": 1500
+    },
+    {
+      "epoch": 3.61,
+      "eval_accuracy": 0.9333446982611661,
+      "eval_loss": 0.23343516886234283,
+      "eval_runtime": 6.5751,
+      "eval_samples_per_second": 252.317,
+      "eval_steps_per_second": 31.635,
+      "step": 1500
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 7.194630872483222e-05,
+      "loss": 0.1555,
+      "step": 1600
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 7.140939597315438e-05,
+      "loss": 0.1439,
+      "step": 1700
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 7.087248322147653e-05,
+      "loss": 0.1181,
+      "step": 1800
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 7.033557046979866e-05,
+      "loss": 0.1215,
+      "step": 1900
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 6.979865771812081e-05,
+      "loss": 0.1219,
+      "step": 2000
+    },
+    {
+      "epoch": 4.81,
+      "eval_accuracy": 0.9382316172292305,
+      "eval_loss": 0.2121623307466507,
+      "eval_runtime": 6.5443,
+      "eval_samples_per_second": 253.505,
+      "eval_steps_per_second": 31.784,
+      "step": 2000
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 6.926174496644296e-05,
+      "loss": 0.117,
+      "step": 2100
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 6.87248322147651e-05,
+      "loss": 0.0911,
+      "step": 2200
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 6.818791946308725e-05,
+      "loss": 0.0925,
+      "step": 2300
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 6.76510067114094e-05,
+      "loss": 0.0956,
+      "step": 2400
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 6.711409395973155e-05,
+      "loss": 0.0943,
+      "step": 2500
+    },
+    {
+      "epoch": 6.01,
+      "eval_accuracy": 0.9419252187748608,
+      "eval_loss": 0.2105204463005066,
+      "eval_runtime": 6.5647,
+      "eval_samples_per_second": 252.714,
+      "eval_steps_per_second": 31.684,
+      "step": 2500
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 6.65771812080537e-05,
+      "loss": 0.069,
+      "step": 2600
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 6.604026845637585e-05,
+      "loss": 0.072,
+      "step": 2700
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 6.5503355704698e-05,
+      "loss": 0.0766,
+      "step": 2800
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 6.496644295302014e-05,
+      "loss": 0.0774,
+      "step": 2900
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 6.442953020134228e-05,
+      "loss": 0.0584,
+      "step": 3000
+    },
+    {
+      "epoch": 7.21,
+      "eval_accuracy": 0.9427207637231504,
+      "eval_loss": 0.2496863603591919,
+      "eval_runtime": 6.5209,
+      "eval_samples_per_second": 254.413,
+      "eval_steps_per_second": 31.898,
+      "step": 3000
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 6.389261744966443e-05,
+      "loss": 0.0582,
+      "step": 3100
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 6.335570469798657e-05,
+      "loss": 0.0611,
+      "step": 3200
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 6.281879194630872e-05,
+      "loss": 0.063,
+      "step": 3300
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 6.228187919463087e-05,
+      "loss": 0.0509,
+      "step": 3400
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 6.174496644295302e-05,
+      "loss": 0.0501,
+      "step": 3500
+    },
+    {
+      "epoch": 8.41,
+      "eval_accuracy": 0.9460734174337992,
+      "eval_loss": 0.24461901187896729,
+      "eval_runtime": 6.5594,
+      "eval_samples_per_second": 252.919,
+      "eval_steps_per_second": 31.71,
+      "step": 3500
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 6.120805369127517e-05,
+      "loss": 0.0516,
+      "step": 3600
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 6.067114093959732e-05,
+      "loss": 0.0505,
+      "step": 3700
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 6.013422818791947e-05,
+      "loss": 0.0431,
+      "step": 3800
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 5.959731543624162e-05,
+      "loss": 0.0393,
+      "step": 3900
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 5.906040268456377e-05,
+      "loss": 0.0397,
+      "step": 4000
+    },
+    {
+      "epoch": 9.62,
+      "eval_accuracy": 0.9453062848050915,
+      "eval_loss": 0.2671561539173126,
+      "eval_runtime": 6.5397,
+      "eval_samples_per_second": 253.681,
+      "eval_steps_per_second": 31.806,
+      "step": 4000
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 5.8523489932885916e-05,
+      "loss": 0.0412,
+      "step": 4100
+    },
+    {
+      "epoch": 10.1,
+      "learning_rate": 5.798657718120806e-05,
+      "loss": 0.0394,
+      "step": 4200
+    },
+    {
+      "epoch": 10.34,
+      "learning_rate": 5.7449664429530206e-05,
+      "loss": 0.032,
+      "step": 4300
+    },
+    {
+      "epoch": 10.58,
+      "learning_rate": 5.6912751677852354e-05,
+      "loss": 0.0333,
+      "step": 4400
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 5.63758389261745e-05,
+      "loss": 0.0327,
+      "step": 4500
+    },
+    {
+      "epoch": 10.82,
+      "eval_accuracy": 0.9442834413001477,
+      "eval_loss": 0.2739590108394623,
+      "eval_runtime": 6.5794,
+      "eval_samples_per_second": 252.151,
+      "eval_steps_per_second": 31.614,
+      "step": 4500
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 5.583892617449665e-05,
+      "loss": 0.0358,
+      "step": 4600
+    },
+    {
+      "epoch": 11.3,
+      "learning_rate": 5.53020134228188e-05,
+      "loss": 0.0263,
+      "step": 4700
+    },
+    {
+      "epoch": 11.54,
+      "learning_rate": 5.476510067114095e-05,
+      "loss": 0.028,
+      "step": 4800
+    },
+    {
+      "epoch": 11.78,
+      "learning_rate": 5.4233557046979874e-05,
+      "loss": 0.0297,
+      "step": 4900
+    },
+    {
+      "epoch": 12.02,
+      "learning_rate": 5.3696644295302015e-05,
+      "loss": 0.0306,
+      "step": 5000
+    },
+    {
+      "epoch": 12.02,
+      "eval_accuracy": 0.9459313558358905,
+      "eval_loss": 0.2738962173461914,
+      "eval_runtime": 6.5965,
+      "eval_samples_per_second": 251.499,
+      "eval_steps_per_second": 31.532,
+      "step": 5000
+    },
+    {
+      "epoch": 12.26,
+      "learning_rate": 5.315973154362416e-05,
+      "loss": 0.0243,
+      "step": 5100
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 5.262281879194631e-05,
+      "loss": 0.0244,
+      "step": 5200
+    },
+    {
+      "epoch": 12.74,
+      "learning_rate": 5.208590604026846e-05,
+      "loss": 0.026,
+      "step": 5300
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 5.154899328859061e-05,
+      "loss": 0.0277,
+      "step": 5400
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 5.1012080536912756e-05,
+      "loss": 0.0211,
+      "step": 5500
+    },
+    {
+      "epoch": 13.22,
+      "eval_accuracy": 0.9471814978974884,
+      "eval_loss": 0.29070886969566345,
+      "eval_runtime": 6.5764,
+      "eval_samples_per_second": 252.266,
+      "eval_steps_per_second": 31.628,
+      "step": 5500
+    },
+    {
+      "epoch": 13.46,
+      "learning_rate": 5.0475167785234905e-05,
+      "loss": 0.0222,
+      "step": 5600
+    },
+    {
+      "epoch": 13.7,
+      "learning_rate": 4.993825503355705e-05,
+      "loss": 0.0224,
+      "step": 5700
+    },
+    {
+      "epoch": 13.94,
+      "learning_rate": 4.94013422818792e-05,
+      "loss": 0.0222,
+      "step": 5800
+    },
+    {
+      "epoch": 14.18,
+      "learning_rate": 4.886442953020135e-05,
+      "loss": 0.0169,
+      "step": 5900
+    },
+    {
+      "epoch": 14.42,
+      "learning_rate": 4.83275167785235e-05,
+      "loss": 0.0186,
+      "step": 6000
+    },
+    {
+      "epoch": 14.42,
+      "eval_accuracy": 0.9481759290828503,
+      "eval_loss": 0.2863553464412689,
+      "eval_runtime": 6.5577,
+      "eval_samples_per_second": 252.985,
+      "eval_steps_per_second": 31.718,
+      "step": 6000
+    },
+    {
+      "epoch": 14.66,
+      "learning_rate": 4.779060402684564e-05,
+      "loss": 0.0189,
+      "step": 6100
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 4.725369127516779e-05,
+      "loss": 0.0185,
+      "step": 6200
+    },
+    {
+      "epoch": 15.14,
+      "learning_rate": 4.6716778523489936e-05,
+      "loss": 0.0189,
+      "step": 6300
+    },
+    {
+      "epoch": 15.38,
+      "learning_rate": 4.6179865771812084e-05,
+      "loss": 0.0148,
+      "step": 6400
+    },
+    {
+      "epoch": 15.62,
+      "learning_rate": 4.564295302013423e-05,
+      "loss": 0.0166,
+      "step": 6500
+    },
+    {
+      "epoch": 15.62,
+      "eval_accuracy": 0.94672690078418,
+      "eval_loss": 0.31074145436286926,
+      "eval_runtime": 6.5741,
+      "eval_samples_per_second": 252.355,
+      "eval_steps_per_second": 31.639,
+      "step": 6500
+    },
+    {
+      "epoch": 15.87,
+      "learning_rate": 4.510604026845638e-05,
+      "loss": 0.017,
+      "step": 6600
+    },
+    {
+      "epoch": 16.11,
+      "learning_rate": 4.456912751677853e-05,
+      "loss": 0.0165,
+      "step": 6700
+    },
+    {
+      "epoch": 16.35,
+      "learning_rate": 4.403221476510068e-05,
+      "loss": 0.0142,
+      "step": 6800
+    },
+    {
+      "epoch": 16.59,
+      "learning_rate": 4.3495302013422825e-05,
+      "loss": 0.0133,
+      "step": 6900
+    },
+    {
+      "epoch": 16.83,
+      "learning_rate": 4.295838926174497e-05,
+      "loss": 0.0149,
+      "step": 7000
+    },
+    {
+      "epoch": 16.83,
+      "eval_accuracy": 0.948090692124105,
+      "eval_loss": 0.30825093388557434,
+      "eval_runtime": 6.5568,
+      "eval_samples_per_second": 253.02,
+      "eval_steps_per_second": 31.723,
+      "step": 7000
+    },
+    {
+      "epoch": 17.07,
+      "learning_rate": 4.242147651006712e-05,
+      "loss": 0.015,
+      "step": 7100
+    },
+    {
+      "epoch": 17.31,
+      "learning_rate": 4.188456375838927e-05,
+      "loss": 0.0124,
+      "step": 7200
+    },
+    {
+      "epoch": 17.55,
+      "learning_rate": 4.134765100671141e-05,
+      "loss": 0.0122,
+      "step": 7300
+    },
+    {
+      "epoch": 17.79,
+      "learning_rate": 4.081073825503356e-05,
+      "loss": 0.0119,
+      "step": 7400
+    },
+    {
+      "epoch": 18.03,
+      "learning_rate": 4.027382550335571e-05,
+      "loss": 0.0138,
+      "step": 7500
+    },
+    {
+      "epoch": 18.03,
+      "eval_accuracy": 0.9493124218661212,
+      "eval_loss": 0.3079260587692261,
+      "eval_runtime": 6.5474,
+      "eval_samples_per_second": 253.382,
+      "eval_steps_per_second": 31.768,
+      "step": 7500
+    },
+    {
+      "epoch": 18.27,
+      "learning_rate": 3.9736912751677856e-05,
+      "loss": 0.0111,
+      "step": 7600
+    },
+    {
+      "epoch": 18.51,
+      "learning_rate": 3.9200000000000004e-05,
+      "loss": 0.0119,
+      "step": 7700
+    },
+    {
+      "epoch": 18.75,
+      "learning_rate": 3.866308724832215e-05,
+      "loss": 0.01,
+      "step": 7800
+    },
+    {
+      "epoch": 18.99,
+      "learning_rate": 3.81261744966443e-05,
+      "loss": 0.0119,
+      "step": 7900
+    },
+    {
+      "epoch": 19.23,
+      "learning_rate": 3.758926174496645e-05,
+      "loss": 0.0107,
+      "step": 8000
+    },
+    {
+      "epoch": 19.23,
+      "eval_accuracy": 0.9478918058870326,
+      "eval_loss": 0.33301591873168945,
+      "eval_runtime": 6.5621,
+      "eval_samples_per_second": 252.814,
+      "eval_steps_per_second": 31.697,
+      "step": 8000
+    },
+    {
+      "epoch": 19.47,
+      "learning_rate": 3.705234899328859e-05,
+      "loss": 0.0103,
+      "step": 8100
+    },
+    {
+      "epoch": 19.71,
+      "learning_rate": 3.651543624161074e-05,
+      "loss": 0.009,
+      "step": 8200
+    },
+    {
+      "epoch": 19.95,
+      "learning_rate": 3.597852348993289e-05,
+      "loss": 0.0109,
+      "step": 8300
+    },
+    {
+      "epoch": 20.19,
+      "learning_rate": 3.5441610738255035e-05,
+      "loss": 0.0091,
+      "step": 8400
+    },
+    {
+      "epoch": 20.43,
+      "learning_rate": 3.490469798657718e-05,
+      "loss": 0.0094,
+      "step": 8500
+    },
+    {
+      "epoch": 20.43,
+      "eval_accuracy": 0.9502216160927378,
+      "eval_loss": 0.3207840025424957,
+      "eval_runtime": 6.5583,
+      "eval_samples_per_second": 252.961,
+      "eval_steps_per_second": 31.715,
+      "step": 8500
+    },
+    {
+      "epoch": 20.67,
+      "learning_rate": 3.436778523489933e-05,
+      "loss": 0.0096,
+      "step": 8600
+    },
+    {
+      "epoch": 20.91,
+      "learning_rate": 3.383087248322148e-05,
+      "loss": 0.0093,
+      "step": 8700
+    },
+    {
+      "epoch": 21.15,
+      "learning_rate": 3.329395973154363e-05,
+      "loss": 0.0079,
+      "step": 8800
+    },
+    {
+      "epoch": 21.39,
+      "learning_rate": 3.2757046979865776e-05,
+      "loss": 0.0083,
+      "step": 8900
+    },
+    {
+      "epoch": 21.63,
+      "learning_rate": 3.2220134228187925e-05,
+      "loss": 0.0079,
+      "step": 9000
+    },
+    {
+      "epoch": 21.63,
+      "eval_accuracy": 0.9491987725877941,
+      "eval_loss": 0.33665016293525696,
+      "eval_runtime": 6.5733,
+      "eval_samples_per_second": 252.384,
+      "eval_steps_per_second": 31.643,
+      "step": 9000
+    },
+    {
+      "epoch": 21.88,
+      "learning_rate": 3.168322147651007e-05,
+      "loss": 0.008,
+      "step": 9100
+    },
+    {
+      "epoch": 22.12,
+      "learning_rate": 3.114630872483222e-05,
+      "loss": 0.0086,
+      "step": 9200
+    },
+    {
+      "epoch": 22.36,
+      "learning_rate": 3.060939597315436e-05,
+      "loss": 0.0071,
+      "step": 9300
+    },
+    {
+      "epoch": 22.6,
+      "learning_rate": 3.0077852348993293e-05,
+      "loss": 0.0076,
+      "step": 9400
+    },
+    {
+      "epoch": 22.84,
+      "learning_rate": 2.9540939597315438e-05,
+      "loss": 0.0074,
+      "step": 9500
+    },
+    {
+      "epoch": 22.84,
+      "eval_accuracy": 0.9497954312990112,
+      "eval_loss": 0.3385924994945526,
+      "eval_runtime": 6.533,
+      "eval_samples_per_second": 253.942,
+      "eval_steps_per_second": 31.838,
+      "step": 9500
+    },
+    {
+      "epoch": 23.08,
+      "learning_rate": 2.9004026845637586e-05,
+      "loss": 0.0076,
+      "step": 9600
+    },
+    {
+      "epoch": 23.32,
+      "learning_rate": 2.8467114093959734e-05,
+      "loss": 0.0059,
+      "step": 9700
+    },
+    {
+      "epoch": 23.56,
+      "learning_rate": 2.7930201342281882e-05,
+      "loss": 0.0061,
+      "step": 9800
+    },
+    {
+      "epoch": 23.8,
+      "learning_rate": 2.739328859060403e-05,
+      "loss": 0.0054,
+      "step": 9900
+    },
+    {
+      "epoch": 24.04,
+      "learning_rate": 2.685637583892618e-05,
+      "loss": 0.0057,
+      "step": 10000
+    },
+    {
+      "epoch": 24.04,
+      "eval_accuracy": 0.9505909762473008,
+      "eval_loss": 0.35995256900787354,
+      "eval_runtime": 6.545,
+      "eval_samples_per_second": 253.477,
+      "eval_steps_per_second": 31.78,
+      "step": 10000
+    },
+    {
+      "epoch": 24.28,
+      "learning_rate": 2.6319463087248324e-05,
+      "loss": 0.0054,
+      "step": 10100
+    },
+    {
+      "epoch": 24.52,
+      "learning_rate": 2.5782550335570472e-05,
+      "loss": 0.0061,
+      "step": 10200
+    },
+    {
+      "epoch": 24.76,
+      "learning_rate": 2.524563758389262e-05,
+      "loss": 0.0057,
+      "step": 10300
+    },
+    {
+      "epoch": 25.0,
+      "learning_rate": 2.470872483221477e-05,
+      "loss": 0.006,
+      "step": 10400
+    },
+    {
+      "epoch": 25.24,
+      "learning_rate": 2.4171812080536917e-05,
+      "loss": 0.0052,
+      "step": 10500
+    },
+    {
+      "epoch": 25.24,
+      "eval_accuracy": 0.9513865211955904,
+      "eval_loss": 0.34160152077674866,
+      "eval_runtime": 6.5426,
+      "eval_samples_per_second": 253.569,
+      "eval_steps_per_second": 31.792,
+      "step": 10500
+    },
+    {
+      "epoch": 25.48,
+      "learning_rate": 2.3634899328859065e-05,
+      "loss": 0.0047,
+      "step": 10600
+    },
+    {
+      "epoch": 25.72,
+      "learning_rate": 2.3103355704697988e-05,
+      "loss": 0.0054,
+      "step": 10700
+    },
+    {
+      "epoch": 25.96,
+      "learning_rate": 2.2566442953020136e-05,
+      "loss": 0.0048,
+      "step": 10800
+    },
+    {
+      "epoch": 26.2,
+      "learning_rate": 2.202953020134228e-05,
+      "loss": 0.0054,
+      "step": 10900
+    },
+    {
+      "epoch": 26.44,
+      "learning_rate": 2.149261744966443e-05,
+      "loss": 0.0038,
+      "step": 11000
+    },
+    {
+      "epoch": 26.44,
+      "eval_accuracy": 0.950448914649392,
+      "eval_loss": 0.35940098762512207,
+      "eval_runtime": 6.5479,
+      "eval_samples_per_second": 253.364,
+      "eval_steps_per_second": 31.766,
+      "step": 11000
+    },
+    {
+      "epoch": 26.68,
+      "learning_rate": 2.0955704697986578e-05,
+      "loss": 0.0041,
+      "step": 11100
+    },
+    {
+      "epoch": 26.92,
+      "learning_rate": 2.0418791946308726e-05,
+      "loss": 0.0037,
+      "step": 11200
+    },
+    {
+      "epoch": 27.16,
+      "learning_rate": 1.9881879194630874e-05,
+      "loss": 0.004,
+      "step": 11300
+    },
+    {
+      "epoch": 27.4,
+      "learning_rate": 1.9344966442953022e-05,
+      "loss": 0.0047,
+      "step": 11400
+    },
+    {
+      "epoch": 27.64,
+      "learning_rate": 1.880805369127517e-05,
+      "loss": 0.0048,
+      "step": 11500
+    },
+    {
+      "epoch": 27.64,
+      "eval_accuracy": 0.9511023979997727,
+      "eval_loss": 0.35828909277915955,
+      "eval_runtime": 6.5576,
+      "eval_samples_per_second": 252.989,
+      "eval_steps_per_second": 31.719,
+      "step": 11500
+    },
+    {
+      "epoch": 27.88,
+      "learning_rate": 1.8271140939597316e-05,
+      "loss": 0.0036,
+      "step": 11600
+    },
+    {
+      "epoch": 28.12,
+      "learning_rate": 1.7734228187919464e-05,
+      "loss": 0.0031,
+      "step": 11700
+    },
+    {
+      "epoch": 28.37,
+      "learning_rate": 1.7197315436241612e-05,
+      "loss": 0.0032,
+      "step": 11800
+    },
+    {
+      "epoch": 28.61,
+      "learning_rate": 1.666040268456376e-05,
+      "loss": 0.0035,
+      "step": 11900
+    },
+    {
+      "epoch": 28.85,
+      "learning_rate": 1.612348993288591e-05,
+      "loss": 0.0033,
+      "step": 12000
+    },
+    {
+      "epoch": 28.85,
+      "eval_accuracy": 0.9508466871235368,
+      "eval_loss": 0.3843555748462677,
+      "eval_runtime": 6.5421,
+      "eval_samples_per_second": 253.588,
+      "eval_steps_per_second": 31.794,
+      "step": 12000
+    },
+    {
+      "epoch": 29.09,
+      "learning_rate": 1.5586577181208057e-05,
+      "loss": 0.0035,
+      "step": 12100
+    },
+    {
+      "epoch": 29.33,
+      "learning_rate": 1.5049664429530202e-05,
+      "loss": 0.0031,
+      "step": 12200
+    },
+    {
+      "epoch": 29.57,
+      "learning_rate": 1.451275167785235e-05,
+      "loss": 0.0029,
+      "step": 12300
+    },
+    {
+      "epoch": 29.81,
+      "learning_rate": 1.3981208053691275e-05,
+      "loss": 0.003,
+      "step": 12400
+    },
+    {
+      "epoch": 30.05,
+      "learning_rate": 1.3444295302013423e-05,
+      "loss": 0.0036,
+      "step": 12500
+    },
+    {
+      "epoch": 30.05,
+      "eval_accuracy": 0.951187634958518,
+      "eval_loss": 0.37571844458580017,
+      "eval_runtime": 6.5956,
+      "eval_samples_per_second": 251.531,
+      "eval_steps_per_second": 31.536,
+      "step": 12500
+    },
+    {
+      "epoch": 30.29,
+      "learning_rate": 1.2907382550335571e-05,
+      "loss": 0.0034,
+      "step": 12600
+    },
+    {
+      "epoch": 30.53,
+      "learning_rate": 1.2370469798657718e-05,
+      "loss": 0.0025,
+      "step": 12700
+    },
+    {
+      "epoch": 30.77,
+      "learning_rate": 1.1833557046979866e-05,
+      "loss": 0.0025,
+      "step": 12800
+    },
+    {
+      "epoch": 31.01,
+      "learning_rate": 1.1296644295302015e-05,
+      "loss": 0.0028,
+      "step": 12900
+    },
+    {
+      "epoch": 31.25,
+      "learning_rate": 1.0759731543624161e-05,
+      "loss": 0.0025,
+      "step": 13000
+    },
+    {
+      "epoch": 31.25,
+      "eval_accuracy": 0.9515285827934993,
+      "eval_loss": 0.38056984543800354,
+      "eval_runtime": 6.5762,
+      "eval_samples_per_second": 252.275,
+      "eval_steps_per_second": 31.629,
+      "step": 13000
+    },
+    {
+      "epoch": 31.49,
+      "learning_rate": 1.022281879194631e-05,
+      "loss": 0.0021,
+      "step": 13100
+    },
+    {
+      "epoch": 31.73,
+      "learning_rate": 9.685906040268458e-06,
+      "loss": 0.0022,
+      "step": 13200
+    },
+    {
+      "epoch": 31.97,
+      "learning_rate": 9.148993288590606e-06,
+      "loss": 0.0026,
+      "step": 13300
+    },
+    {
+      "epoch": 32.21,
+      "learning_rate": 8.612080536912752e-06,
+      "loss": 0.0019,
+      "step": 13400
+    },
+    {
+      "epoch": 32.45,
+      "learning_rate": 8.0751677852349e-06,
+      "loss": 0.0018,
+      "step": 13500
+    },
+    {
+      "epoch": 32.45,
+      "eval_accuracy": 0.9526082509376066,
+      "eval_loss": 0.3810262680053711,
+      "eval_runtime": 6.5701,
+      "eval_samples_per_second": 252.508,
+      "eval_steps_per_second": 31.659,
+      "step": 13500
+    },
+    {
+      "epoch": 32.69,
+      "learning_rate": 7.538255033557048e-06,
+      "loss": 0.0021,
+      "step": 13600
+    },
+    {
+      "epoch": 32.93,
+      "learning_rate": 7.001342281879195e-06,
+      "loss": 0.0022,
+      "step": 13700
+    },
+    {
+      "epoch": 33.17,
+      "learning_rate": 6.464429530201343e-06,
+      "loss": 0.0017,
+      "step": 13800
+    },
+    {
+      "epoch": 33.41,
+      "learning_rate": 5.92751677852349e-06,
+      "loss": 0.0022,
+      "step": 13900
+    },
+    {
+      "epoch": 33.65,
+      "learning_rate": 5.390604026845638e-06,
+      "loss": 0.0016,
+      "step": 14000
+    },
+    {
+      "epoch": 33.65,
+      "eval_accuracy": 0.9529491987725878,
+      "eval_loss": 0.37525418400764465,
+      "eval_runtime": 6.5513,
+      "eval_samples_per_second": 253.233,
+      "eval_steps_per_second": 31.75,
+      "step": 14000
+    },
+    {
+      "epoch": 33.89,
+      "learning_rate": 4.853691275167786e-06,
+      "loss": 0.0017,
+      "step": 14100
+    },
+    {
+      "epoch": 34.13,
+      "learning_rate": 4.316778523489933e-06,
+      "loss": 0.0018,
+      "step": 14200
+    },
+    {
+      "epoch": 34.38,
+      "learning_rate": 3.7798657718120806e-06,
+      "loss": 0.0016,
+      "step": 14300
+    },
+    {
+      "epoch": 34.62,
+      "learning_rate": 3.2429530201342284e-06,
+      "loss": 0.0019,
+      "step": 14400
+    },
+    {
+      "epoch": 34.86,
+      "learning_rate": 2.706040268456376e-06,
+      "loss": 0.0014,
+      "step": 14500
+    },
+    {
+      "epoch": 34.86,
+      "eval_accuracy": 0.95266507557677,
+      "eval_loss": 0.37661275267601013,
+      "eval_runtime": 6.5633,
+      "eval_samples_per_second": 252.77,
+      "eval_steps_per_second": 31.692,
+      "step": 14500
+    },
+    {
+      "epoch": 35.1,
+      "learning_rate": 2.1691275167785236e-06,
+      "loss": 0.0017,
+      "step": 14600
+    },
+    {
+      "epoch": 35.34,
+      "learning_rate": 1.6322147651006712e-06,
+      "loss": 0.0015,
+      "step": 14700
+    },
+    {
+      "epoch": 35.58,
+      "learning_rate": 1.0953020134228189e-06,
+      "loss": 0.0017,
+      "step": 14800
+    },
+    {
+      "epoch": 35.82,
+      "learning_rate": 5.583892617449665e-07,
+      "loss": 0.0013,
+      "step": 14900
+    },
+    {
+      "epoch": 36.06,
+      "learning_rate": 2.1476510067114093e-08,
+      "loss": 0.0011,
+      "step": 15000
+    },
+    {
+      "epoch": 36.06,
+      "eval_accuracy": 0.9524661893396977,
+      "eval_loss": 0.3779400885105133,
+      "eval_runtime": 6.5493,
+      "eval_samples_per_second": 253.31,
+      "eval_steps_per_second": 31.759,
+      "step": 15000
+    },
+    {
+      "epoch": 36.06,
+      "step": 15000,
+      "total_flos": 7.878113593860096e+16,
+      "train_loss": 0.05964354458153248,
+      "train_runtime": 2290.905,
+      "train_samples_per_second": 209.524,
+      "train_steps_per_second": 6.548
+    }
+  ],
+  "max_steps": 15000,
+  "num_train_epochs": 37,
+  "total_flos": 7.878113593860096e+16,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6c46f854e9ffe1e52b18f5e014a239c66245bdcde06b28318886f7efe39e766
+size 3387