Upload folder using huggingface_hub
Browse files- config.json +21 -0
- training_args.json +6 -0
config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"vocab_size": 128256,
|
| 3 |
+
"max_seq_len": 256,
|
| 4 |
+
"d_model": 3456,
|
| 5 |
+
"n_layers": 28,
|
| 6 |
+
"d_state": 128,
|
| 7 |
+
"mimo_rank": 4,
|
| 8 |
+
"mimo_start_fraction": 0.75,
|
| 9 |
+
"bc_norm": true,
|
| 10 |
+
"bc_bias": true,
|
| 11 |
+
"dt_min": 0.001,
|
| 12 |
+
"dt_max": 0.1,
|
| 13 |
+
"ffn_hidden": 9216,
|
| 14 |
+
"segment_len": 256,
|
| 15 |
+
"memory_depth": 1,
|
| 16 |
+
"fast_weight_rank": 48,
|
| 17 |
+
"persistent_slots": 4,
|
| 18 |
+
"retrieved_memory_slots": 2,
|
| 19 |
+
"fusion_block_every": 14,
|
| 20 |
+
"tie_embeddings": true
|
| 21 |
+
}
|
training_args.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"step": 5000,
|
| 3 |
+
"stage": 1,
|
| 4 |
+
"model_type": "mighty-giant",
|
| 5 |
+
"architecture": "mamba3-titans-hybrid"
|
| 6 |
+
}
|