Example config.json (?)
#8
by
zakhar-kogan
- opened
- config.json +36 -0
config.json
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"MambaForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"bos_token_id": 0,
|
| 6 |
+
"conv_kernel": 4,
|
| 7 |
+
"eos_token_id": 0,
|
| 8 |
+
"expand": 2,
|
| 9 |
+
"fused_add_norm": true,
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 2560,
|
| 12 |
+
"initializer_range": 0.1,
|
| 13 |
+
"intermediate_size": 5120,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "mamba",
|
| 16 |
+
"n_layer": 64,
|
| 17 |
+
"num_hidden_layers": 64,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"pad_vocab_size_multiple": 8,
|
| 20 |
+
"rescale_prenorm_residual": false,
|
| 21 |
+
"residual_in_fp32": true,
|
| 22 |
+
"rms_norm": true,
|
| 23 |
+
"state_size": 16,
|
| 24 |
+
"time_step_floor": 0.0001,
|
| 25 |
+
"time_step_init_scheme": "random",
|
| 26 |
+
"time_step_max": 0.1,
|
| 27 |
+
"time_step_min": 0.001,
|
| 28 |
+
"time_step_rank": 160,
|
| 29 |
+
"time_step_scale": 1.0,
|
| 30 |
+
"torch_dtype": "float32",
|
| 31 |
+
"transformers_version": "4.39.0.dev0",
|
| 32 |
+
"use_bias": false,
|
| 33 |
+
"use_cache": true,
|
| 34 |
+
"use_conv_bias": true,
|
| 35 |
+
"vocab_size": 50280
|
| 36 |
+
}
|