Upload folder using huggingface_hub
Browse files- checkpoint-700/config.json +17 -0
- checkpoint-700/generation_config.json +6 -0
- checkpoint-700/hybrid_tokenizer_mapping.json +0 -0
- checkpoint-700/model.safetensors +3 -0
- checkpoint-700/optimizer.pt +3 -0
- checkpoint-700/rng_state.pth +3 -0
- checkpoint-700/scaler.pt +3 -0
- checkpoint-700/scheduler.pt +3 -0
- checkpoint-700/tokenizer_config.json +9 -0
- checkpoint-700/trainer_state.json +524 -0
- checkpoint-700/training_args.bin +3 -0
checkpoint-700/config.json
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"FastHyperNet125mForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"dtype": "float32",
|
| 6 |
+
"hidden_size": 768,
|
| 7 |
+
"hyperbolic_dim": 16,
|
| 8 |
+
"initializer_range": 0.02,
|
| 9 |
+
"intermediate_size": 2048,
|
| 10 |
+
"max_position_embeddings": 512,
|
| 11 |
+
"model_type": "fasthypernet_125m_tagless",
|
| 12 |
+
"num_attention_heads": 12,
|
| 13 |
+
"num_hidden_layers": 16,
|
| 14 |
+
"transformers_version": "5.13.1",
|
| 15 |
+
"use_cache": false,
|
| 16 |
+
"vocab_size": 8192
|
| 17 |
+
}
|
checkpoint-700/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"output_attentions": false,
|
| 4 |
+
"output_hidden_states": false,
|
| 5 |
+
"transformers_version": "5.13.1"
|
| 6 |
+
}
|
checkpoint-700/hybrid_tokenizer_mapping.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-700/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3318b5250f07871df6c5703bb0a397a180275a1fd76195068f2f60cb0015dcb5
|
| 3 |
+
size 501759848
|
checkpoint-700/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:693ac18308a2b116ee46de8c0c878f2fe04008e34084066b2f3786edf84af12a
|
| 3 |
+
size 254410983
|
checkpoint-700/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8297cebcaeec237f8774fcd2eb91d4ac6afaa0a57577f3ca4e7743e4e3469c8e
|
| 3 |
+
size 14645
|
checkpoint-700/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8fa0daa924908d42885ba3da99d2ce3bb6fe52cb2eaf4c9c2bb2ab8b35376fa6
|
| 3 |
+
size 1383
|
checkpoint-700/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2cc627c1e45817f0ab36507e03fbf607a294489149559c44459d1a88e6c496f9
|
| 3 |
+
size 1465
|
checkpoint-700/tokenizer_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"tokenizer_class": "HFCompatibleHybridTokenizer",
|
| 3 |
+
"vocab_size": 8192,
|
| 4 |
+
"model_max_length": 512,
|
| 5 |
+
"bos_token": "<bos>",
|
| 6 |
+
"eos_token": "<eos>",
|
| 7 |
+
"unk_token": "<unk>",
|
| 8 |
+
"pad_token": "<pad>"
|
| 9 |
+
}
|
checkpoint-700/trainer_state.json
ADDED
|
@@ -0,0 +1,524 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.14698934327261273,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 700,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.002099847761037325,
|
| 14 |
+
"grad_norm": 19872.2890625,
|
| 15 |
+
"learning_rate": 3e-07,
|
| 16 |
+
"loss": 8.7939,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.00419969552207465,
|
| 21 |
+
"grad_norm": 1458.198974609375,
|
| 22 |
+
"learning_rate": 1.2999999999999998e-06,
|
| 23 |
+
"loss": 8.1117,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.006299543283111974,
|
| 28 |
+
"grad_norm": 297.6234130859375,
|
| 29 |
+
"learning_rate": 2.2999999999999996e-06,
|
| 30 |
+
"loss": 7.3977,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.0083993910441493,
|
| 35 |
+
"grad_norm": 426.92987060546875,
|
| 36 |
+
"learning_rate": 3.2999999999999993e-06,
|
| 37 |
+
"loss": 7.2069,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.010499238805186623,
|
| 42 |
+
"grad_norm": 474.7532653808594,
|
| 43 |
+
"learning_rate": 4.2999999999999995e-06,
|
| 44 |
+
"loss": 7.08,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.012599086566223949,
|
| 49 |
+
"grad_norm": 1569.4381103515625,
|
| 50 |
+
"learning_rate": 5.3e-06,
|
| 51 |
+
"loss": 7.0092,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.014698934327261274,
|
| 56 |
+
"grad_norm": 1407.21142578125,
|
| 57 |
+
"learning_rate": 6.3e-06,
|
| 58 |
+
"loss": 6.9929,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.0167987820882986,
|
| 63 |
+
"grad_norm": 500.49493408203125,
|
| 64 |
+
"learning_rate": 7.299999999999999e-06,
|
| 65 |
+
"loss": 6.9501,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.018898629849335925,
|
| 70 |
+
"grad_norm": 442.69537353515625,
|
| 71 |
+
"learning_rate": 8.299999999999998e-06,
|
| 72 |
+
"loss": 6.9001,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.020998477610373247,
|
| 77 |
+
"grad_norm": 1423.2950439453125,
|
| 78 |
+
"learning_rate": 9.299999999999999e-06,
|
| 79 |
+
"loss": 6.9037,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.023098325371410572,
|
| 84 |
+
"grad_norm": 448.7774963378906,
|
| 85 |
+
"learning_rate": 1.03e-05,
|
| 86 |
+
"loss": 6.8732,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.025198173132447897,
|
| 91 |
+
"grad_norm": 711.8992919921875,
|
| 92 |
+
"learning_rate": 1.1299999999999999e-05,
|
| 93 |
+
"loss": 6.8381,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.027298020893485223,
|
| 98 |
+
"grad_norm": 884.006591796875,
|
| 99 |
+
"learning_rate": 1.2299999999999999e-05,
|
| 100 |
+
"loss": 6.8462,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.029397868654522548,
|
| 105 |
+
"grad_norm": 96.5859375,
|
| 106 |
+
"learning_rate": 1.33e-05,
|
| 107 |
+
"loss": 6.8433,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.03149771641555987,
|
| 112 |
+
"grad_norm": 91.52490997314453,
|
| 113 |
+
"learning_rate": 1.43e-05,
|
| 114 |
+
"loss": 6.8171,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.0335975641765972,
|
| 119 |
+
"grad_norm": 354.9374084472656,
|
| 120 |
+
"learning_rate": 1.53e-05,
|
| 121 |
+
"loss": 6.7899,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.035697411937634524,
|
| 126 |
+
"grad_norm": 189.09506225585938,
|
| 127 |
+
"learning_rate": 1.6299999999999996e-05,
|
| 128 |
+
"loss": 6.7456,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.03779725969867185,
|
| 133 |
+
"grad_norm": 170.39376831054688,
|
| 134 |
+
"learning_rate": 1.7299999999999997e-05,
|
| 135 |
+
"loss": 6.7505,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.03989710745970917,
|
| 140 |
+
"grad_norm": 193.9166259765625,
|
| 141 |
+
"learning_rate": 1.8299999999999998e-05,
|
| 142 |
+
"loss": 6.752,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.04199695522074649,
|
| 147 |
+
"grad_norm": 336.08624267578125,
|
| 148 |
+
"learning_rate": 1.93e-05,
|
| 149 |
+
"loss": 6.7015,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.04409680298178382,
|
| 154 |
+
"grad_norm": 307.40557861328125,
|
| 155 |
+
"learning_rate": 2.03e-05,
|
| 156 |
+
"loss": 6.7024,
|
| 157 |
+
"step": 210
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.046196650742821144,
|
| 161 |
+
"grad_norm": 175.94061279296875,
|
| 162 |
+
"learning_rate": 2.1299999999999996e-05,
|
| 163 |
+
"loss": 6.6639,
|
| 164 |
+
"step": 220
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.04829649850385847,
|
| 168 |
+
"grad_norm": 496.3529052734375,
|
| 169 |
+
"learning_rate": 2.23e-05,
|
| 170 |
+
"loss": 6.6309,
|
| 171 |
+
"step": 230
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.050396346264895794,
|
| 175 |
+
"grad_norm": 112.63114929199219,
|
| 176 |
+
"learning_rate": 2.3299999999999997e-05,
|
| 177 |
+
"loss": 6.6146,
|
| 178 |
+
"step": 240
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.05249619402593312,
|
| 182 |
+
"grad_norm": 141.9170379638672,
|
| 183 |
+
"learning_rate": 2.4299999999999998e-05,
|
| 184 |
+
"loss": 6.6386,
|
| 185 |
+
"step": 250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.054596041786970445,
|
| 189 |
+
"grad_norm": 120.92185974121094,
|
| 190 |
+
"learning_rate": 2.53e-05,
|
| 191 |
+
"loss": 6.6323,
|
| 192 |
+
"step": 260
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.05669588954800777,
|
| 196 |
+
"grad_norm": 241.84129333496094,
|
| 197 |
+
"learning_rate": 2.63e-05,
|
| 198 |
+
"loss": 6.598,
|
| 199 |
+
"step": 270
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.058795737309045096,
|
| 203 |
+
"grad_norm": 115.57807159423828,
|
| 204 |
+
"learning_rate": 2.7299999999999996e-05,
|
| 205 |
+
"loss": 6.5549,
|
| 206 |
+
"step": 280
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.06089558507008242,
|
| 210 |
+
"grad_norm": 99.67350769042969,
|
| 211 |
+
"learning_rate": 2.83e-05,
|
| 212 |
+
"loss": 6.5484,
|
| 213 |
+
"step": 290
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.06299543283111975,
|
| 217 |
+
"grad_norm": 105.20110321044922,
|
| 218 |
+
"learning_rate": 2.9299999999999997e-05,
|
| 219 |
+
"loss": 6.5416,
|
| 220 |
+
"step": 300
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.06509528059215706,
|
| 224 |
+
"grad_norm": 207.86203002929688,
|
| 225 |
+
"learning_rate": 3.0299999999999998e-05,
|
| 226 |
+
"loss": 6.5463,
|
| 227 |
+
"step": 310
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.0671951283531944,
|
| 231 |
+
"grad_norm": 125.6773910522461,
|
| 232 |
+
"learning_rate": 3.1299999999999995e-05,
|
| 233 |
+
"loss": 6.4848,
|
| 234 |
+
"step": 320
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.06929497611423172,
|
| 238 |
+
"grad_norm": 121.31608581542969,
|
| 239 |
+
"learning_rate": 3.229999999999999e-05,
|
| 240 |
+
"loss": 6.4763,
|
| 241 |
+
"step": 330
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.07139482387526905,
|
| 245 |
+
"grad_norm": 86.57124328613281,
|
| 246 |
+
"learning_rate": 3.3299999999999996e-05,
|
| 247 |
+
"loss": 6.4579,
|
| 248 |
+
"step": 340
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.07349467163630637,
|
| 252 |
+
"grad_norm": 105.74996948242188,
|
| 253 |
+
"learning_rate": 3.4299999999999993e-05,
|
| 254 |
+
"loss": 6.408,
|
| 255 |
+
"step": 350
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.0755945193973437,
|
| 259 |
+
"grad_norm": 68.95454406738281,
|
| 260 |
+
"learning_rate": 3.53e-05,
|
| 261 |
+
"loss": 6.3928,
|
| 262 |
+
"step": 360
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.07769436715838102,
|
| 266 |
+
"grad_norm": 107.94305419921875,
|
| 267 |
+
"learning_rate": 3.6299999999999995e-05,
|
| 268 |
+
"loss": 6.3853,
|
| 269 |
+
"step": 370
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.07979421491941834,
|
| 273 |
+
"grad_norm": 136.39556884765625,
|
| 274 |
+
"learning_rate": 3.73e-05,
|
| 275 |
+
"loss": 6.3639,
|
| 276 |
+
"step": 380
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.08189406268045567,
|
| 280 |
+
"grad_norm": 152.05712890625,
|
| 281 |
+
"learning_rate": 3.83e-05,
|
| 282 |
+
"loss": 6.348,
|
| 283 |
+
"step": 390
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.08399391044149299,
|
| 287 |
+
"grad_norm": 139.25466918945312,
|
| 288 |
+
"learning_rate": 3.93e-05,
|
| 289 |
+
"loss": 6.3261,
|
| 290 |
+
"step": 400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.08609375820253032,
|
| 294 |
+
"grad_norm": 130.061279296875,
|
| 295 |
+
"learning_rate": 4.03e-05,
|
| 296 |
+
"loss": 6.2809,
|
| 297 |
+
"step": 410
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.08819360596356764,
|
| 301 |
+
"grad_norm": 115.39459991455078,
|
| 302 |
+
"learning_rate": 4.1299999999999994e-05,
|
| 303 |
+
"loss": 6.2646,
|
| 304 |
+
"step": 420
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.09029345372460497,
|
| 308 |
+
"grad_norm": 50.863616943359375,
|
| 309 |
+
"learning_rate": 4.229999999999999e-05,
|
| 310 |
+
"loss": 6.2808,
|
| 311 |
+
"step": 430
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.09239330148564229,
|
| 315 |
+
"grad_norm": 113.06661987304688,
|
| 316 |
+
"learning_rate": 4.3299999999999995e-05,
|
| 317 |
+
"loss": 6.2387,
|
| 318 |
+
"step": 440
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.09449314924667962,
|
| 322 |
+
"grad_norm": 202.89398193359375,
|
| 323 |
+
"learning_rate": 4.43e-05,
|
| 324 |
+
"loss": 6.2476,
|
| 325 |
+
"step": 450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.09659299700771694,
|
| 329 |
+
"grad_norm": 154.97865295410156,
|
| 330 |
+
"learning_rate": 4.5299999999999997e-05,
|
| 331 |
+
"loss": 6.2345,
|
| 332 |
+
"step": 460
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.09869284476875427,
|
| 336 |
+
"grad_norm": 72.81023406982422,
|
| 337 |
+
"learning_rate": 4.6299999999999994e-05,
|
| 338 |
+
"loss": 6.1687,
|
| 339 |
+
"step": 470
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.10079269252979159,
|
| 343 |
+
"grad_norm": 101.40583801269531,
|
| 344 |
+
"learning_rate": 4.73e-05,
|
| 345 |
+
"loss": 6.1705,
|
| 346 |
+
"step": 480
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.10289254029082892,
|
| 350 |
+
"grad_norm": 80.50221252441406,
|
| 351 |
+
"learning_rate": 4.8299999999999995e-05,
|
| 352 |
+
"loss": 6.1665,
|
| 353 |
+
"step": 490
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.10499238805186624,
|
| 357 |
+
"grad_norm": 91.6736068725586,
|
| 358 |
+
"learning_rate": 4.929999999999999e-05,
|
| 359 |
+
"loss": 6.1587,
|
| 360 |
+
"step": 500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.10709223581290356,
|
| 364 |
+
"grad_norm": 253.29408264160156,
|
| 365 |
+
"learning_rate": 5.0299999999999996e-05,
|
| 366 |
+
"loss": 6.1218,
|
| 367 |
+
"step": 510
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.10919208357394089,
|
| 371 |
+
"grad_norm": 117.43933868408203,
|
| 372 |
+
"learning_rate": 5.13e-05,
|
| 373 |
+
"loss": 6.1313,
|
| 374 |
+
"step": 520
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.11129193133497821,
|
| 378 |
+
"grad_norm": 69.32656860351562,
|
| 379 |
+
"learning_rate": 5.23e-05,
|
| 380 |
+
"loss": 6.1038,
|
| 381 |
+
"step": 530
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.11339177909601554,
|
| 385 |
+
"grad_norm": 85.16144561767578,
|
| 386 |
+
"learning_rate": 5.3299999999999995e-05,
|
| 387 |
+
"loss": 6.0745,
|
| 388 |
+
"step": 540
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.11549162685705286,
|
| 392 |
+
"grad_norm": 67.62905883789062,
|
| 393 |
+
"learning_rate": 5.429999999999999e-05,
|
| 394 |
+
"loss": 6.0786,
|
| 395 |
+
"step": 550
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.11759147461809019,
|
| 399 |
+
"grad_norm": 103.80160522460938,
|
| 400 |
+
"learning_rate": 5.529999999999999e-05,
|
| 401 |
+
"loss": 6.0472,
|
| 402 |
+
"step": 560
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.11969132237912751,
|
| 406 |
+
"grad_norm": 45.10822677612305,
|
| 407 |
+
"learning_rate": 5.63e-05,
|
| 408 |
+
"loss": 6.0311,
|
| 409 |
+
"step": 570
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.12179117014016484,
|
| 413 |
+
"grad_norm": 59.20137023925781,
|
| 414 |
+
"learning_rate": 5.73e-05,
|
| 415 |
+
"loss": 6.0219,
|
| 416 |
+
"step": 580
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.12389101790120216,
|
| 420 |
+
"grad_norm": 49.00022506713867,
|
| 421 |
+
"learning_rate": 5.8299999999999994e-05,
|
| 422 |
+
"loss": 6.0013,
|
| 423 |
+
"step": 590
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.1259908656622395,
|
| 427 |
+
"grad_norm": 50.88911437988281,
|
| 428 |
+
"learning_rate": 5.929999999999999e-05,
|
| 429 |
+
"loss": 5.981,
|
| 430 |
+
"step": 600
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.12809071342327683,
|
| 434 |
+
"grad_norm": 43.3376579284668,
|
| 435 |
+
"learning_rate": 6.0299999999999995e-05,
|
| 436 |
+
"loss": 5.9482,
|
| 437 |
+
"step": 610
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.13019056118431413,
|
| 441 |
+
"grad_norm": 54.567779541015625,
|
| 442 |
+
"learning_rate": 6.13e-05,
|
| 443 |
+
"loss": 5.9478,
|
| 444 |
+
"step": 620
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.13229040894535146,
|
| 448 |
+
"grad_norm": 51.82001495361328,
|
| 449 |
+
"learning_rate": 6.23e-05,
|
| 450 |
+
"loss": 5.9583,
|
| 451 |
+
"step": 630
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.1343902567063888,
|
| 455 |
+
"grad_norm": 65.38532257080078,
|
| 456 |
+
"learning_rate": 6.33e-05,
|
| 457 |
+
"loss": 5.9228,
|
| 458 |
+
"step": 640
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.1364901044674261,
|
| 462 |
+
"grad_norm": 47.24418640136719,
|
| 463 |
+
"learning_rate": 6.429999999999999e-05,
|
| 464 |
+
"loss": 5.8942,
|
| 465 |
+
"step": 650
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.13858995222846343,
|
| 469 |
+
"grad_norm": 39.78090286254883,
|
| 470 |
+
"learning_rate": 6.53e-05,
|
| 471 |
+
"loss": 5.8764,
|
| 472 |
+
"step": 660
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.14068979998950076,
|
| 476 |
+
"grad_norm": 59.94386672973633,
|
| 477 |
+
"learning_rate": 6.63e-05,
|
| 478 |
+
"loss": 5.8611,
|
| 479 |
+
"step": 670
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.1427896477505381,
|
| 483 |
+
"grad_norm": 50.66661071777344,
|
| 484 |
+
"learning_rate": 6.73e-05,
|
| 485 |
+
"loss": 5.8163,
|
| 486 |
+
"step": 680
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.1448894955115754,
|
| 490 |
+
"grad_norm": 59.74818801879883,
|
| 491 |
+
"learning_rate": 6.829999999999999e-05,
|
| 492 |
+
"loss": 5.8398,
|
| 493 |
+
"step": 690
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.14698934327261273,
|
| 497 |
+
"grad_norm": 53.87904739379883,
|
| 498 |
+
"learning_rate": 6.93e-05,
|
| 499 |
+
"loss": 5.7908,
|
| 500 |
+
"step": 700
|
| 501 |
+
}
|
| 502 |
+
],
|
| 503 |
+
"logging_steps": 10,
|
| 504 |
+
"max_steps": 100000,
|
| 505 |
+
"num_input_tokens_seen": 0,
|
| 506 |
+
"num_train_epochs": 21,
|
| 507 |
+
"save_steps": 100,
|
| 508 |
+
"stateful_callbacks": {
|
| 509 |
+
"TrainerControl": {
|
| 510 |
+
"args": {
|
| 511 |
+
"should_epoch_stop": false,
|
| 512 |
+
"should_evaluate": false,
|
| 513 |
+
"should_log": false,
|
| 514 |
+
"should_save": true,
|
| 515 |
+
"should_training_stop": false
|
| 516 |
+
},
|
| 517 |
+
"attributes": {}
|
| 518 |
+
}
|
| 519 |
+
},
|
| 520 |
+
"total_flos": 4.56474459181056e+16,
|
| 521 |
+
"train_batch_size": 16,
|
| 522 |
+
"trial_name": null,
|
| 523 |
+
"trial_params": null
|
| 524 |
+
}
|
checkpoint-700/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:000c64b2f44395822e8bf59265f8e6e5d085b71ae5ccbb8b496fbe6f256940c8
|
| 3 |
+
size 5201
|