Delete llm\models\small-extended-8k\config.json with huggingface_hub
Browse files
llm//models//small-extended-8k//config.json
DELETED
|
@@ -1,31 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"model_config": {
|
| 3 |
-
"model_name": "OpenLLM-Small-8k",
|
| 4 |
-
"model_size": "small",
|
| 5 |
-
"vocab_size": 32000,
|
| 6 |
-
"n_layer": 6,
|
| 7 |
-
"n_head": 8,
|
| 8 |
-
"n_embd": 512,
|
| 9 |
-
"block_size": 1024,
|
| 10 |
-
"dropout": 0.1,
|
| 11 |
-
"bias": true,
|
| 12 |
-
"training_info": {
|
| 13 |
-
"step": 8000,
|
| 14 |
-
"best_loss": 5.3,
|
| 15 |
-
"model_type": "gpt-small"
|
| 16 |
-
}
|
| 17 |
-
},
|
| 18 |
-
"tokenizer_config": {
|
| 19 |
-
"type": "sentencepiece",
|
| 20 |
-
"vocab_size": 32000,
|
| 21 |
-
"model_file": "data/tokenizer/tokenizer.model"
|
| 22 |
-
},
|
| 23 |
-
"training_config": {
|
| 24 |
-
"learning_rate": 0.0001,
|
| 25 |
-
"batch_size": 4,
|
| 26 |
-
"gradient_accumulation_steps": 4,
|
| 27 |
-
"max_steps": 8000,
|
| 28 |
-
"warmup_steps": 100,
|
| 29 |
-
"weight_decay": 0.01
|
| 30 |
-
}
|
| 31 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|