Upload DistillationWrapper
Browse files- config.json +2 -1
- pytorch_model.bin +3 -0
config.json
CHANGED
|
@@ -5,7 +5,7 @@
|
|
| 5 |
],
|
| 6 |
"attention_probs_dropout_prob": 0.1,
|
| 7 |
"classifier_dropout": null,
|
| 8 |
-
"dtype": "
|
| 9 |
"embedding_size": 128,
|
| 10 |
"expert_intermediate_size": 2624,
|
| 11 |
"group_depth": 4,
|
|
@@ -24,6 +24,7 @@
|
|
| 24 |
"num_expert_modules": 3,
|
| 25 |
"num_experts": 8,
|
| 26 |
"num_hidden_layers": 16,
|
|
|
|
| 27 |
"pad_token_id": 0,
|
| 28 |
"router_jitter_noise": 0.01,
|
| 29 |
"top_k": 2,
|
|
|
|
| 5 |
],
|
| 6 |
"attention_probs_dropout_prob": 0.1,
|
| 7 |
"classifier_dropout": null,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
"embedding_size": 128,
|
| 10 |
"expert_intermediate_size": 2624,
|
| 11 |
"group_depth": 4,
|
|
|
|
| 24 |
"num_expert_modules": 3,
|
| 25 |
"num_experts": 8,
|
| 26 |
"num_hidden_layers": 16,
|
| 27 |
+
"output_hidden_states": true,
|
| 28 |
"pad_token_id": 0,
|
| 29 |
"router_jitter_noise": 0.01,
|
| 30 |
"top_k": 2,
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6105d560faf130f812413c74584b275b33f44892a61a835f686ccb80d5b66a22
|
| 3 |
+
size 151589821
|