Commit
•
3e2610c
1
Parent(s):
578258f
Push model using huggingface_hub.
Browse files- config.json +4 -15
- model.safetensors +2 -2
config.json
CHANGED
|
@@ -1,20 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"accumulate_grad_batches": 64,
|
| 3 |
"auxk": 256,
|
| 4 |
-
"
|
| 5 |
-
"batch_size": 1,
|
| 6 |
-
"dead_steps_threshold": null,
|
| 7 |
"dead_threshold": 0.001,
|
| 8 |
-
"dead_tokens_threshold": 10000000,
|
| 9 |
-
"expansion_factor": 64,
|
| 10 |
"k": 32,
|
| 11 |
-
"
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
"lr": 0.0001,
|
| 15 |
-
"max_length": 2048,
|
| 16 |
-
"model_name": "EleutherAI/pythia-160m-deduped",
|
| 17 |
-
"skip_special_tokens": true,
|
| 18 |
-
"standardize": true,
|
| 19 |
-
"tuned_lens": false
|
| 20 |
}
|
|
|
|
| 1 |
{
|
|
|
|
| 2 |
"auxk": 256,
|
| 3 |
+
"dead_steps_threshold": 76,
|
|
|
|
|
|
|
| 4 |
"dead_threshold": 0.001,
|
|
|
|
|
|
|
| 5 |
"k": 32,
|
| 6 |
+
"n_inputs": 768,
|
| 7 |
+
"n_latents": 49152,
|
| 8 |
+
"standardize": true
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47837fa0b51039b32e6da02f5b839d6a92c0f0e750f16cbfb27ba8271ad6b096
|
| 3 |
+
size 301993232
|