Upload folder using huggingface_hub
Browse files
shared/checkpoints/checkpoint-568/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"KosyForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_dropout": 0.0,
|
| 6 |
+
"dtype": "float32",
|
| 7 |
+
"hidden_dropout": 0.0,
|
| 8 |
+
"hidden_size": 640,
|
| 9 |
+
"initializer_range": 0.02,
|
| 10 |
+
"intermediate_size": 1792,
|
| 11 |
+
"max_position_embeddings": 512,
|
| 12 |
+
"model_type": "kosy",
|
| 13 |
+
"neftune_alpha": 0.0,
|
| 14 |
+
"num_attention_heads": 10,
|
| 15 |
+
"num_hidden_layers": 32,
|
| 16 |
+
"num_key_value_heads": 5,
|
| 17 |
+
"qk_norm": true,
|
| 18 |
+
"rezero_init": 1.0,
|
| 19 |
+
"rms_norm_eps": 1e-06,
|
| 20 |
+
"rope_theta": 10000.0,
|
| 21 |
+
"transformers_version": "5.0.0",
|
| 22 |
+
"use_cache": false,
|
| 23 |
+
"vocab_size": 32000
|
| 24 |
+
}
|
shared/checkpoints/checkpoint-568/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|