Made Config json from EasyTransformerConfig processing
Browse files- config.json +32 -0
config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"n_layers": 2,
|
| 3 |
+
"d_model": 736,
|
| 4 |
+
"n_ctx": 1024,
|
| 5 |
+
"d_head": 64,
|
| 6 |
+
"n_heads": 11,
|
| 7 |
+
"d_mlp": 2944,
|
| 8 |
+
"act_fn": "solu_ln",
|
| 9 |
+
"d_vocab": 50278,
|
| 10 |
+
"eps": 1e-05,
|
| 11 |
+
"use_attn_result": false,
|
| 12 |
+
"use_attn_scale": true,
|
| 13 |
+
"use_local_attn": false,
|
| 14 |
+
"model_family": null,
|
| 15 |
+
"checkpoint": null,
|
| 16 |
+
"tokenizer_name": "EleutherAI/gpt-neox-20b",
|
| 17 |
+
"window_size": null,
|
| 18 |
+
"attn_types": null,
|
| 19 |
+
"init_mode": "gpt2",
|
| 20 |
+
"normalization_type": "LNPre",
|
| 21 |
+
"device": "cuda",
|
| 22 |
+
"attention_dir": "causal",
|
| 23 |
+
"attn_only": false,
|
| 24 |
+
"seed": 42,
|
| 25 |
+
"initializer_range": 0.02948839123097943,
|
| 26 |
+
"init_weights": true,
|
| 27 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 28 |
+
"positional_embedding_type": "standard",
|
| 29 |
+
"final_rms": true,
|
| 30 |
+
"d_vocab_out": 50278,
|
| 31 |
+
"parallel_attn_mlp": false,
|
| 32 |
+
}
|