Trained with Unsloth - config
Browse files- config.json +9 -4
config.json
CHANGED
|
@@ -18,7 +18,7 @@
|
|
| 18 |
"conv_dim": 1024,
|
| 19 |
"conv_dim_out": 1024,
|
| 20 |
"conv_use_xavier_init": true,
|
| 21 |
-
"torch_dtype": "
|
| 22 |
"eos_token_id": 7,
|
| 23 |
"hidden_size": 1024,
|
| 24 |
"initializer_range": 0.02,
|
|
@@ -42,6 +42,7 @@
|
|
| 42 |
"conv"
|
| 43 |
],
|
| 44 |
"max_position_embeddings": 128000,
|
|
|
|
| 45 |
"model_type": "lfm2",
|
| 46 |
"norm_eps": 1e-05,
|
| 47 |
"num_attention_heads": 16,
|
|
@@ -49,9 +50,13 @@
|
|
| 49 |
"num_hidden_layers": 16,
|
| 50 |
"num_key_value_heads": 8,
|
| 51 |
"pad_token_id": 0,
|
| 52 |
-
"
|
| 53 |
-
|
| 54 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 55 |
"use_pos_enc": true,
|
| 56 |
"vocab_size": 65536
|
| 57 |
}
|
|
|
|
| 18 |
"conv_dim": 1024,
|
| 19 |
"conv_dim_out": 1024,
|
| 20 |
"conv_use_xavier_init": true,
|
| 21 |
+
"torch_dtype": "bfloat16",
|
| 22 |
"eos_token_id": 7,
|
| 23 |
"hidden_size": 1024,
|
| 24 |
"initializer_range": 0.02,
|
|
|
|
| 42 |
"conv"
|
| 43 |
],
|
| 44 |
"max_position_embeddings": 128000,
|
| 45 |
+
"model_name": "LiquidAI/LFM2-350M",
|
| 46 |
"model_type": "lfm2",
|
| 47 |
"norm_eps": 1e-05,
|
| 48 |
"num_attention_heads": 16,
|
|
|
|
| 50 |
"num_hidden_layers": 16,
|
| 51 |
"num_key_value_heads": 8,
|
| 52 |
"pad_token_id": 0,
|
| 53 |
+
"rope_parameters": {
|
| 54 |
+
"rope_theta": 1000000.0,
|
| 55 |
+
"rope_type": "default"
|
| 56 |
+
},
|
| 57 |
+
"tie_word_embeddings": true,
|
| 58 |
+
"unsloth_version": "2026.3.4",
|
| 59 |
+
"use_cache": false,
|
| 60 |
"use_pos_enc": true,
|
| 61 |
"vocab_size": 65536
|
| 62 |
}
|