Upload-Qwen3 4B Run 4.1D TEST (config/tokenizer): config.json
Browse files- config.json +31 -0
config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_attn_implementation_autoset": true,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"Qwen3ForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 151643,
|
| 9 |
+
"eos_token_id": 151643,
|
| 10 |
+
"head_dim": 128,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 2560,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 9728,
|
| 15 |
+
"max_position_embeddings": 32768,
|
| 16 |
+
"max_window_layers": 36,
|
| 17 |
+
"model_type": "qwen3",
|
| 18 |
+
"num_attention_heads": 32,
|
| 19 |
+
"num_hidden_layers": 36,
|
| 20 |
+
"num_key_value_heads": 8,
|
| 21 |
+
"rms_norm_eps": 1e-06,
|
| 22 |
+
"rope_scaling": null,
|
| 23 |
+
"rope_theta": 1000000,
|
| 24 |
+
"sliding_window": null,
|
| 25 |
+
"tie_word_embeddings": true,
|
| 26 |
+
"torch_dtype": "bfloat16",
|
| 27 |
+
"transformers_version": "4.52.0.dev0",
|
| 28 |
+
"use_cache": false,
|
| 29 |
+
"use_sliding_window": false,
|
| 30 |
+
"vocab_size": 151936
|
| 31 |
+
}
|