Upload 8 files
Browse files- config.json +32 -0
- generation_config.json +6 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +663 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "EleutherAI/pythia-410m-deduped",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"GPTNeoXForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": true,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"classifier_dropout": 0.1,
|
| 10 |
+
"eos_token_id": 0,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout": 0.0,
|
| 13 |
+
"hidden_size": 1024,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 4096,
|
| 16 |
+
"layer_norm_eps": 1e-05,
|
| 17 |
+
"max_position_embeddings": 2048,
|
| 18 |
+
"model_type": "gpt_neox",
|
| 19 |
+
"num_attention_heads": 16,
|
| 20 |
+
"num_hidden_layers": 24,
|
| 21 |
+
"partial_rotary_factor": 0.25,
|
| 22 |
+
"rope_scaling": null,
|
| 23 |
+
"rope_theta": 10000,
|
| 24 |
+
"rotary_emb_base": 10000,
|
| 25 |
+
"rotary_pct": 0.25,
|
| 26 |
+
"tie_word_embeddings": false,
|
| 27 |
+
"torch_dtype": "float32",
|
| 28 |
+
"transformers_version": "4.46.0",
|
| 29 |
+
"use_cache": true,
|
| 30 |
+
"use_parallel_residual": true,
|
| 31 |
+
"vocab_size": 50304
|
| 32 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": 0,
|
| 5 |
+
"transformers_version": "4.46.0"
|
| 6 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:40f8d3a44a03c4c70030df96495ad0365bae104999750c5cd8126aed47c33f09
|
| 3 |
+
size 1621370224
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:03c9c3e30a8a616e44af0ee4f3789b53b5ed00819ce012a44ac21cb6fe20649e
|
| 3 |
+
size 3242916026
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6c44141bbdc09951afac7e1120c543329e05011264f0246580d2c7b808a155c7
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:645644195de668b36b925cd41f32d2b819c1c9ed782f9e1972c7f32ff665063a
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,663 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.9607843137254903,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 9000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02178649237472767,
|
| 13 |
+
"grad_norm": 1.4119257926940918,
|
| 14 |
+
"learning_rate": 4.945533769063181e-05,
|
| 15 |
+
"loss": 0.6917,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.04357298474945534,
|
| 20 |
+
"grad_norm": 1.3989609479904175,
|
| 21 |
+
"learning_rate": 4.891067538126362e-05,
|
| 22 |
+
"loss": 0.4496,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.06535947712418301,
|
| 27 |
+
"grad_norm": 1.8628056049346924,
|
| 28 |
+
"learning_rate": 4.8366013071895424e-05,
|
| 29 |
+
"loss": 0.4774,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.08714596949891068,
|
| 34 |
+
"grad_norm": 1.6151583194732666,
|
| 35 |
+
"learning_rate": 4.7821350762527234e-05,
|
| 36 |
+
"loss": 0.4805,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.10893246187363835,
|
| 41 |
+
"grad_norm": 0.3191039264202118,
|
| 42 |
+
"learning_rate": 4.7276688453159044e-05,
|
| 43 |
+
"loss": 0.4375,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.13071895424836602,
|
| 48 |
+
"grad_norm": 1.743823766708374,
|
| 49 |
+
"learning_rate": 4.673202614379085e-05,
|
| 50 |
+
"loss": 0.4678,
|
| 51 |
+
"step": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.15250544662309368,
|
| 55 |
+
"grad_norm": 1.9511303901672363,
|
| 56 |
+
"learning_rate": 4.6187363834422656e-05,
|
| 57 |
+
"loss": 0.5285,
|
| 58 |
+
"step": 700
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.17429193899782136,
|
| 62 |
+
"grad_norm": 0.5090198516845703,
|
| 63 |
+
"learning_rate": 4.564270152505447e-05,
|
| 64 |
+
"loss": 0.4451,
|
| 65 |
+
"step": 800
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.19607843137254902,
|
| 69 |
+
"grad_norm": 2.6404006481170654,
|
| 70 |
+
"learning_rate": 4.5098039215686275e-05,
|
| 71 |
+
"loss": 0.4524,
|
| 72 |
+
"step": 900
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.2178649237472767,
|
| 76 |
+
"grad_norm": 2.596175193786621,
|
| 77 |
+
"learning_rate": 4.4553376906318085e-05,
|
| 78 |
+
"loss": 0.4284,
|
| 79 |
+
"step": 1000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.23965141612200436,
|
| 83 |
+
"grad_norm": 2.235574960708618,
|
| 84 |
+
"learning_rate": 4.400871459694989e-05,
|
| 85 |
+
"loss": 0.4234,
|
| 86 |
+
"step": 1100
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.26143790849673204,
|
| 90 |
+
"grad_norm": 1.786952018737793,
|
| 91 |
+
"learning_rate": 4.3464052287581704e-05,
|
| 92 |
+
"loss": 0.4881,
|
| 93 |
+
"step": 1200
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.28322440087145967,
|
| 97 |
+
"grad_norm": 1.1873655319213867,
|
| 98 |
+
"learning_rate": 4.291938997821351e-05,
|
| 99 |
+
"loss": 0.4714,
|
| 100 |
+
"step": 1300
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.30501089324618735,
|
| 104 |
+
"grad_norm": 1.3493982553482056,
|
| 105 |
+
"learning_rate": 4.2374727668845316e-05,
|
| 106 |
+
"loss": 0.4329,
|
| 107 |
+
"step": 1400
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.32679738562091504,
|
| 111 |
+
"grad_norm": 1.2561562061309814,
|
| 112 |
+
"learning_rate": 4.1830065359477126e-05,
|
| 113 |
+
"loss": 0.3972,
|
| 114 |
+
"step": 1500
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.3485838779956427,
|
| 118 |
+
"grad_norm": 3.1698968410491943,
|
| 119 |
+
"learning_rate": 4.1285403050108935e-05,
|
| 120 |
+
"loss": 0.4276,
|
| 121 |
+
"step": 1600
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.37037037037037035,
|
| 125 |
+
"grad_norm": 1.0001728534698486,
|
| 126 |
+
"learning_rate": 4.074074074074074e-05,
|
| 127 |
+
"loss": 0.5036,
|
| 128 |
+
"step": 1700
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.39215686274509803,
|
| 132 |
+
"grad_norm": 2.307466745376587,
|
| 133 |
+
"learning_rate": 4.0196078431372555e-05,
|
| 134 |
+
"loss": 0.437,
|
| 135 |
+
"step": 1800
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.4139433551198257,
|
| 139 |
+
"grad_norm": 1.7833284139633179,
|
| 140 |
+
"learning_rate": 3.965141612200436e-05,
|
| 141 |
+
"loss": 0.4415,
|
| 142 |
+
"step": 1900
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.4357298474945534,
|
| 146 |
+
"grad_norm": 1.4704581499099731,
|
| 147 |
+
"learning_rate": 3.910675381263617e-05,
|
| 148 |
+
"loss": 0.4752,
|
| 149 |
+
"step": 2000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.45751633986928103,
|
| 153 |
+
"grad_norm": 1.4827431440353394,
|
| 154 |
+
"learning_rate": 3.8562091503267977e-05,
|
| 155 |
+
"loss": 0.4268,
|
| 156 |
+
"step": 2100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.4793028322440087,
|
| 160 |
+
"grad_norm": 0.10033423453569412,
|
| 161 |
+
"learning_rate": 3.8017429193899786e-05,
|
| 162 |
+
"loss": 0.4641,
|
| 163 |
+
"step": 2200
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.5010893246187363,
|
| 167 |
+
"grad_norm": 2.280324935913086,
|
| 168 |
+
"learning_rate": 3.747276688453159e-05,
|
| 169 |
+
"loss": 0.4393,
|
| 170 |
+
"step": 2300
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.5228758169934641,
|
| 174 |
+
"grad_norm": 0.7182455658912659,
|
| 175 |
+
"learning_rate": 3.6928104575163405e-05,
|
| 176 |
+
"loss": 0.4143,
|
| 177 |
+
"step": 2400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.5446623093681917,
|
| 181 |
+
"grad_norm": 6.931342124938965,
|
| 182 |
+
"learning_rate": 3.638344226579521e-05,
|
| 183 |
+
"loss": 0.4742,
|
| 184 |
+
"step": 2500
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.5664488017429193,
|
| 188 |
+
"grad_norm": 1.6501855850219727,
|
| 189 |
+
"learning_rate": 3.583877995642702e-05,
|
| 190 |
+
"loss": 0.4183,
|
| 191 |
+
"step": 2600
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.5882352941176471,
|
| 195 |
+
"grad_norm": 15.619595527648926,
|
| 196 |
+
"learning_rate": 3.529411764705883e-05,
|
| 197 |
+
"loss": 0.4591,
|
| 198 |
+
"step": 2700
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.6100217864923747,
|
| 202 |
+
"grad_norm": 1.5793421268463135,
|
| 203 |
+
"learning_rate": 3.474945533769064e-05,
|
| 204 |
+
"loss": 0.4326,
|
| 205 |
+
"step": 2800
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.6318082788671024,
|
| 209 |
+
"grad_norm": 1.2501574754714966,
|
| 210 |
+
"learning_rate": 3.420479302832244e-05,
|
| 211 |
+
"loss": 0.4019,
|
| 212 |
+
"step": 2900
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.6535947712418301,
|
| 216 |
+
"grad_norm": 2.454075336456299,
|
| 217 |
+
"learning_rate": 3.366013071895425e-05,
|
| 218 |
+
"loss": 0.4306,
|
| 219 |
+
"step": 3000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.6753812636165577,
|
| 223 |
+
"grad_norm": 1.4100277423858643,
|
| 224 |
+
"learning_rate": 3.311546840958606e-05,
|
| 225 |
+
"loss": 0.4533,
|
| 226 |
+
"step": 3100
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.6971677559912854,
|
| 230 |
+
"grad_norm": 1.2618755102157593,
|
| 231 |
+
"learning_rate": 3.257080610021787e-05,
|
| 232 |
+
"loss": 0.4523,
|
| 233 |
+
"step": 3200
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.7189542483660131,
|
| 237 |
+
"grad_norm": 1.4713486433029175,
|
| 238 |
+
"learning_rate": 3.202614379084967e-05,
|
| 239 |
+
"loss": 0.4371,
|
| 240 |
+
"step": 3300
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.7407407407407407,
|
| 244 |
+
"grad_norm": 1.1793227195739746,
|
| 245 |
+
"learning_rate": 3.148148148148148e-05,
|
| 246 |
+
"loss": 0.4465,
|
| 247 |
+
"step": 3400
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.7625272331154684,
|
| 251 |
+
"grad_norm": 0.7598987221717834,
|
| 252 |
+
"learning_rate": 3.093681917211329e-05,
|
| 253 |
+
"loss": 0.4759,
|
| 254 |
+
"step": 3500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.7843137254901961,
|
| 258 |
+
"grad_norm": 1.8617286682128906,
|
| 259 |
+
"learning_rate": 3.0392156862745097e-05,
|
| 260 |
+
"loss": 0.4248,
|
| 261 |
+
"step": 3600
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.8061002178649237,
|
| 265 |
+
"grad_norm": 0.8327335715293884,
|
| 266 |
+
"learning_rate": 2.984749455337691e-05,
|
| 267 |
+
"loss": 0.4257,
|
| 268 |
+
"step": 3700
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.8278867102396514,
|
| 272 |
+
"grad_norm": 1.4007729291915894,
|
| 273 |
+
"learning_rate": 2.9302832244008716e-05,
|
| 274 |
+
"loss": 0.4066,
|
| 275 |
+
"step": 3800
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.8496732026143791,
|
| 279 |
+
"grad_norm": 1.50600266456604,
|
| 280 |
+
"learning_rate": 2.8758169934640522e-05,
|
| 281 |
+
"loss": 0.4225,
|
| 282 |
+
"step": 3900
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.8714596949891068,
|
| 286 |
+
"grad_norm": 1.622689127922058,
|
| 287 |
+
"learning_rate": 2.8213507625272335e-05,
|
| 288 |
+
"loss": 0.4262,
|
| 289 |
+
"step": 4000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.8932461873638344,
|
| 293 |
+
"grad_norm": 1.0574493408203125,
|
| 294 |
+
"learning_rate": 2.766884531590414e-05,
|
| 295 |
+
"loss": 0.4102,
|
| 296 |
+
"step": 4100
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.9150326797385621,
|
| 300 |
+
"grad_norm": 0.6209813356399536,
|
| 301 |
+
"learning_rate": 2.7124183006535947e-05,
|
| 302 |
+
"loss": 0.4231,
|
| 303 |
+
"step": 4200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.9368191721132898,
|
| 307 |
+
"grad_norm": 0.6364914178848267,
|
| 308 |
+
"learning_rate": 2.657952069716776e-05,
|
| 309 |
+
"loss": 0.4556,
|
| 310 |
+
"step": 4300
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.9586056644880174,
|
| 314 |
+
"grad_norm": 0.9522435665130615,
|
| 315 |
+
"learning_rate": 2.6034858387799566e-05,
|
| 316 |
+
"loss": 0.4344,
|
| 317 |
+
"step": 4400
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.9803921568627451,
|
| 321 |
+
"grad_norm": 0.8629395365715027,
|
| 322 |
+
"learning_rate": 2.5490196078431373e-05,
|
| 323 |
+
"loss": 0.4111,
|
| 324 |
+
"step": 4500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 1.0021786492374727,
|
| 328 |
+
"grad_norm": 1.2452311515808105,
|
| 329 |
+
"learning_rate": 2.4945533769063182e-05,
|
| 330 |
+
"loss": 0.3741,
|
| 331 |
+
"step": 4600
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 1.0239651416122004,
|
| 335 |
+
"grad_norm": 0.6457747220993042,
|
| 336 |
+
"learning_rate": 2.4400871459694992e-05,
|
| 337 |
+
"loss": 0.3617,
|
| 338 |
+
"step": 4700
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.0457516339869282,
|
| 342 |
+
"grad_norm": 3.9131181240081787,
|
| 343 |
+
"learning_rate": 2.38562091503268e-05,
|
| 344 |
+
"loss": 0.3263,
|
| 345 |
+
"step": 4800
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.0675381263616557,
|
| 349 |
+
"grad_norm": 1.347782015800476,
|
| 350 |
+
"learning_rate": 2.3311546840958608e-05,
|
| 351 |
+
"loss": 0.3248,
|
| 352 |
+
"step": 4900
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 1.0893246187363834,
|
| 356 |
+
"grad_norm": 15.944000244140625,
|
| 357 |
+
"learning_rate": 2.2766884531590417e-05,
|
| 358 |
+
"loss": 0.3329,
|
| 359 |
+
"step": 5000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.1111111111111112,
|
| 363 |
+
"grad_norm": 0.9070430994033813,
|
| 364 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 365 |
+
"loss": 0.3405,
|
| 366 |
+
"step": 5100
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.132897603485839,
|
| 370 |
+
"grad_norm": 1.5099300146102905,
|
| 371 |
+
"learning_rate": 2.1677559912854033e-05,
|
| 372 |
+
"loss": 0.3405,
|
| 373 |
+
"step": 5200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 1.1546840958605664,
|
| 377 |
+
"grad_norm": 1.4810432195663452,
|
| 378 |
+
"learning_rate": 2.113289760348584e-05,
|
| 379 |
+
"loss": 0.3575,
|
| 380 |
+
"step": 5300
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.1764705882352942,
|
| 384 |
+
"grad_norm": 1.5918389558792114,
|
| 385 |
+
"learning_rate": 2.058823529411765e-05,
|
| 386 |
+
"loss": 0.3284,
|
| 387 |
+
"step": 5400
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.1982570806100217,
|
| 391 |
+
"grad_norm": 1.1542829275131226,
|
| 392 |
+
"learning_rate": 2.0043572984749455e-05,
|
| 393 |
+
"loss": 0.3328,
|
| 394 |
+
"step": 5500
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 1.2200435729847494,
|
| 398 |
+
"grad_norm": 2.9146111011505127,
|
| 399 |
+
"learning_rate": 1.9498910675381264e-05,
|
| 400 |
+
"loss": 0.3358,
|
| 401 |
+
"step": 5600
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.2418300653594772,
|
| 405 |
+
"grad_norm": 1.1495596170425415,
|
| 406 |
+
"learning_rate": 1.895424836601307e-05,
|
| 407 |
+
"loss": 0.3581,
|
| 408 |
+
"step": 5700
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 1.263616557734205,
|
| 412 |
+
"grad_norm": 1.925133228302002,
|
| 413 |
+
"learning_rate": 1.840958605664488e-05,
|
| 414 |
+
"loss": 0.338,
|
| 415 |
+
"step": 5800
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 1.2854030501089324,
|
| 419 |
+
"grad_norm": 0.6670601963996887,
|
| 420 |
+
"learning_rate": 1.786492374727669e-05,
|
| 421 |
+
"loss": 0.3166,
|
| 422 |
+
"step": 5900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 1.3071895424836601,
|
| 426 |
+
"grad_norm": 1.3753750324249268,
|
| 427 |
+
"learning_rate": 1.7320261437908496e-05,
|
| 428 |
+
"loss": 0.3322,
|
| 429 |
+
"step": 6000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.3289760348583877,
|
| 433 |
+
"grad_norm": 0.8356995582580566,
|
| 434 |
+
"learning_rate": 1.6775599128540306e-05,
|
| 435 |
+
"loss": 0.3465,
|
| 436 |
+
"step": 6100
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 1.3507625272331154,
|
| 440 |
+
"grad_norm": 0.6578918695449829,
|
| 441 |
+
"learning_rate": 1.6230936819172112e-05,
|
| 442 |
+
"loss": 0.3449,
|
| 443 |
+
"step": 6200
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.3725490196078431,
|
| 447 |
+
"grad_norm": 0.8830516338348389,
|
| 448 |
+
"learning_rate": 1.568627450980392e-05,
|
| 449 |
+
"loss": 0.3152,
|
| 450 |
+
"step": 6300
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 1.3943355119825709,
|
| 454 |
+
"grad_norm": 2.0208992958068848,
|
| 455 |
+
"learning_rate": 1.5141612200435731e-05,
|
| 456 |
+
"loss": 0.3237,
|
| 457 |
+
"step": 6400
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.4161220043572984,
|
| 461 |
+
"grad_norm": 0.7672997117042542,
|
| 462 |
+
"learning_rate": 1.4596949891067537e-05,
|
| 463 |
+
"loss": 0.3317,
|
| 464 |
+
"step": 6500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.4379084967320261,
|
| 468 |
+
"grad_norm": 1.3644828796386719,
|
| 469 |
+
"learning_rate": 1.4052287581699347e-05,
|
| 470 |
+
"loss": 0.3341,
|
| 471 |
+
"step": 6600
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.4596949891067539,
|
| 475 |
+
"grad_norm": 1.7263342142105103,
|
| 476 |
+
"learning_rate": 1.3507625272331156e-05,
|
| 477 |
+
"loss": 0.2875,
|
| 478 |
+
"step": 6700
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 1.4814814814814814,
|
| 482 |
+
"grad_norm": 0.5791486501693726,
|
| 483 |
+
"learning_rate": 1.2962962962962962e-05,
|
| 484 |
+
"loss": 0.3583,
|
| 485 |
+
"step": 6800
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.5032679738562091,
|
| 489 |
+
"grad_norm": 1.689420223236084,
|
| 490 |
+
"learning_rate": 1.2418300653594772e-05,
|
| 491 |
+
"loss": 0.3011,
|
| 492 |
+
"step": 6900
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 1.5250544662309369,
|
| 496 |
+
"grad_norm": 0.42838066816329956,
|
| 497 |
+
"learning_rate": 1.187363834422658e-05,
|
| 498 |
+
"loss": 0.3153,
|
| 499 |
+
"step": 7000
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 1.5468409586056646,
|
| 503 |
+
"grad_norm": 1.6973720788955688,
|
| 504 |
+
"learning_rate": 1.1328976034858388e-05,
|
| 505 |
+
"loss": 0.3253,
|
| 506 |
+
"step": 7100
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.5686274509803921,
|
| 510 |
+
"grad_norm": 1.9545549154281616,
|
| 511 |
+
"learning_rate": 1.0784313725490197e-05,
|
| 512 |
+
"loss": 0.3348,
|
| 513 |
+
"step": 7200
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.5904139433551199,
|
| 517 |
+
"grad_norm": 1.2370498180389404,
|
| 518 |
+
"learning_rate": 1.0239651416122005e-05,
|
| 519 |
+
"loss": 0.3207,
|
| 520 |
+
"step": 7300
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 1.6122004357298474,
|
| 524 |
+
"grad_norm": 1.108176589012146,
|
| 525 |
+
"learning_rate": 9.694989106753813e-06,
|
| 526 |
+
"loss": 0.3145,
|
| 527 |
+
"step": 7400
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 1.6339869281045751,
|
| 531 |
+
"grad_norm": 1.849331259727478,
|
| 532 |
+
"learning_rate": 9.150326797385621e-06,
|
| 533 |
+
"loss": 0.2982,
|
| 534 |
+
"step": 7500
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 1.6557734204793029,
|
| 538 |
+
"grad_norm": 1.2936469316482544,
|
| 539 |
+
"learning_rate": 8.60566448801743e-06,
|
| 540 |
+
"loss": 0.3161,
|
| 541 |
+
"step": 7600
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 1.6775599128540306,
|
| 545 |
+
"grad_norm": 1.2373560667037964,
|
| 546 |
+
"learning_rate": 8.061002178649239e-06,
|
| 547 |
+
"loss": 0.3335,
|
| 548 |
+
"step": 7700
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 1.6993464052287581,
|
| 552 |
+
"grad_norm": 0.7337093949317932,
|
| 553 |
+
"learning_rate": 7.5163398692810456e-06,
|
| 554 |
+
"loss": 0.3027,
|
| 555 |
+
"step": 7800
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.7211328976034859,
|
| 559 |
+
"grad_norm": 0.2871895134449005,
|
| 560 |
+
"learning_rate": 6.971677559912855e-06,
|
| 561 |
+
"loss": 0.3048,
|
| 562 |
+
"step": 7900
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 1.7429193899782134,
|
| 566 |
+
"grad_norm": 1.113721489906311,
|
| 567 |
+
"learning_rate": 6.427015250544663e-06,
|
| 568 |
+
"loss": 0.3076,
|
| 569 |
+
"step": 8000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.7647058823529411,
|
| 573 |
+
"grad_norm": 1.1479638814926147,
|
| 574 |
+
"learning_rate": 5.882352941176471e-06,
|
| 575 |
+
"loss": 0.3513,
|
| 576 |
+
"step": 8100
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 1.7864923747276689,
|
| 580 |
+
"grad_norm": 0.8669132590293884,
|
| 581 |
+
"learning_rate": 5.33769063180828e-06,
|
| 582 |
+
"loss": 0.3347,
|
| 583 |
+
"step": 8200
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 1.8082788671023966,
|
| 587 |
+
"grad_norm": 1.534184455871582,
|
| 588 |
+
"learning_rate": 4.7930283224400875e-06,
|
| 589 |
+
"loss": 0.336,
|
| 590 |
+
"step": 8300
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 1.8300653594771243,
|
| 594 |
+
"grad_norm": 1.0394097566604614,
|
| 595 |
+
"learning_rate": 4.2483660130718954e-06,
|
| 596 |
+
"loss": 0.3152,
|
| 597 |
+
"step": 8400
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.8518518518518519,
|
| 601 |
+
"grad_norm": 0.9366344809532166,
|
| 602 |
+
"learning_rate": 3.7037037037037037e-06,
|
| 603 |
+
"loss": 0.3272,
|
| 604 |
+
"step": 8500
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 1.8736383442265794,
|
| 608 |
+
"grad_norm": 1.5486916303634644,
|
| 609 |
+
"learning_rate": 3.159041394335512e-06,
|
| 610 |
+
"loss": 0.3223,
|
| 611 |
+
"step": 8600
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 1.8954248366013071,
|
| 615 |
+
"grad_norm": 2.589451551437378,
|
| 616 |
+
"learning_rate": 2.6143790849673204e-06,
|
| 617 |
+
"loss": 0.3189,
|
| 618 |
+
"step": 8700
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 1.9172113289760349,
|
| 622 |
+
"grad_norm": 0.5650085806846619,
|
| 623 |
+
"learning_rate": 2.0697167755991287e-06,
|
| 624 |
+
"loss": 0.3132,
|
| 625 |
+
"step": 8800
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.9389978213507626,
|
| 629 |
+
"grad_norm": 0.18872463703155518,
|
| 630 |
+
"learning_rate": 1.525054466230937e-06,
|
| 631 |
+
"loss": 0.3219,
|
| 632 |
+
"step": 8900
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 1.9607843137254903,
|
| 636 |
+
"grad_norm": 1.094841718673706,
|
| 637 |
+
"learning_rate": 9.80392156862745e-07,
|
| 638 |
+
"loss": 0.3015,
|
| 639 |
+
"step": 9000
|
| 640 |
+
}
|
| 641 |
+
],
|
| 642 |
+
"logging_steps": 100,
|
| 643 |
+
"max_steps": 9180,
|
| 644 |
+
"num_input_tokens_seen": 0,
|
| 645 |
+
"num_train_epochs": 2,
|
| 646 |
+
"save_steps": 1000,
|
| 647 |
+
"stateful_callbacks": {
|
| 648 |
+
"TrainerControl": {
|
| 649 |
+
"args": {
|
| 650 |
+
"should_epoch_stop": false,
|
| 651 |
+
"should_evaluate": false,
|
| 652 |
+
"should_log": false,
|
| 653 |
+
"should_save": true,
|
| 654 |
+
"should_training_stop": false
|
| 655 |
+
},
|
| 656 |
+
"attributes": {}
|
| 657 |
+
}
|
| 658 |
+
},
|
| 659 |
+
"total_flos": 7.825775061368832e+16,
|
| 660 |
+
"train_batch_size": 8,
|
| 661 |
+
"trial_name": null,
|
| 662 |
+
"trial_params": null
|
| 663 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c06bcf3356c4e1b4be3ca7956c0a71c4149a5b309b56013e2dfcaf5a860bc787
|
| 3 |
+
size 5176
|