Upload checkpoints/config.json with huggingface_hub
Browse files- checkpoints/config.json +29 -0
checkpoints/config.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": {
|
| 3 |
+
"vocab_size": 4096,
|
| 4 |
+
"max_seq_len": 128,
|
| 5 |
+
"d_model": 384,
|
| 6 |
+
"n_layers": 6,
|
| 7 |
+
"n_heads": 6,
|
| 8 |
+
"ffn_hidden": 768,
|
| 9 |
+
"dropout": 0.1,
|
| 10 |
+
"pad_id": 0,
|
| 11 |
+
"bos_id": 1,
|
| 12 |
+
"eos_id": 2
|
| 13 |
+
},
|
| 14 |
+
"train": {
|
| 15 |
+
"batch_size": 32,
|
| 16 |
+
"learning_rate": 0.0003,
|
| 17 |
+
"min_lr": 3e-05,
|
| 18 |
+
"weight_decay": 0.1,
|
| 19 |
+
"warmup_steps": 200,
|
| 20 |
+
"max_steps": 10000,
|
| 21 |
+
"eval_interval": 200,
|
| 22 |
+
"save_interval": 500,
|
| 23 |
+
"grad_clip": 1.0,
|
| 24 |
+
"device": "auto",
|
| 25 |
+
"seed": 42,
|
| 26 |
+
"data_dir": "data",
|
| 27 |
+
"output_dir": "checkpoints"
|
| 28 |
+
}
|
| 29 |
+
}
|