Upload config.json
Browse files- config.json +41 -0
config.json
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"KrakenForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"auto_map": {
|
| 6 |
+
"AutoConfig": "configuration_kraken.KrakenConfig",
|
| 7 |
+
"AutoModel": "modeling_kraken.KrakenForCausalLM",
|
| 8 |
+
"AutoModelForCausalLM": "modeling_kraken.KrakenForCausalLM"
|
| 9 |
+
},
|
| 10 |
+
"config_dict": {
|
| 11 |
+
"class_indices": {
|
| 12 |
+
"LABEL_0": 0,
|
| 13 |
+
"LABEL_1": 1,
|
| 14 |
+
"LABEL_2": 2,
|
| 15 |
+
"LABEL_3": 3
|
| 16 |
+
},
|
| 17 |
+
"model_type": "kraken",
|
| 18 |
+
"models": {
|
| 19 |
+
"expert1": "VAGOsolutions/Llama-3-SauerkrautLM-8b-Instruct",
|
| 20 |
+
"expert2": "mii-community/zefiro-7b-dpo-ITA",
|
| 21 |
+
"expert3": "paulml/Hermes-2-Pro-French",
|
| 22 |
+
"expert4": "norallm/normistral-7b-warm-instruct"
|
| 23 |
+
},
|
| 24 |
+
"quantization": {
|
| 25 |
+
"expert1": null,
|
| 26 |
+
"expert2": null,
|
| 27 |
+
"expert3": null,
|
| 28 |
+
"expert4": null
|
| 29 |
+
},
|
| 30 |
+
"router": "kraken_router",
|
| 31 |
+
"tokenizers": {
|
| 32 |
+
"expert1": "VAGOsolutions/Llama-3-SauerkrautLM-8b-Instruct",
|
| 33 |
+
"expert2": "mii-community/zefiro-7b-dpo-ITA",
|
| 34 |
+
"expert3": "paulml/Hermes-2-Pro-French",
|
| 35 |
+
"expert4": "norallm/normistral-7b-warm-instruct"
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"model_type": "kraken",
|
| 39 |
+
"torch_dtype": "float32",
|
| 40 |
+
"transformers_version": "4.41.0"
|
| 41 |
+
}
|