Upload folder using huggingface_hub
Browse files- config.json +21 -0
- model.safetensors +3 -0
config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architecture": "RotorNative25M",
|
| 3 |
+
"n_harmonics": 40,
|
| 4 |
+
"n_layers": 20,
|
| 5 |
+
"projection_free": false,
|
| 6 |
+
"d_model": 560,
|
| 7 |
+
"n_groups": 7,
|
| 8 |
+
"d_osc": 80,
|
| 9 |
+
"vocab_size": 256,
|
| 10 |
+
"seq_len": 256,
|
| 11 |
+
"dt": 0.3,
|
| 12 |
+
"description": "Rotor 25M fine-tuned on salon CoT dialogues (historical figures + thinking tags)",
|
| 13 |
+
"training": {
|
| 14 |
+
"step": 2500,
|
| 15 |
+
"bpb": 1.7598,
|
| 16 |
+
"val_bpb": 1.5449,
|
| 17 |
+
"base": "rotor-25m-annealed",
|
| 18 |
+
"finetune": "salon_cot",
|
| 19 |
+
"optimizer": "RotationalAdamW"
|
| 20 |
+
}
|
| 21 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4250b357e8ae39bca56420b8338a0093495e1289cf691de238a18823d7559a87
|
| 3 |
+
size 102054400
|