[WIP] Optimized q4f16 ONNX export
#3
by
Xenova
HF staff
- opened
- config.json +3 -0
- onnx/model_q4f16.onnx +2 -2
config.json
CHANGED
|
@@ -25,6 +25,9 @@
|
|
| 25 |
"tie_word_embeddings": true,
|
| 26 |
"torch_dtype": "bfloat16",
|
| 27 |
"transformers_version": "4.42.3",
|
|
|
|
|
|
|
|
|
|
| 28 |
"use_cache": true,
|
| 29 |
"vocab_size": 49152
|
| 30 |
}
|
|
|
|
| 25 |
"tie_word_embeddings": true,
|
| 26 |
"torch_dtype": "bfloat16",
|
| 27 |
"transformers_version": "4.42.3",
|
| 28 |
+
"transformers.js_config": {
|
| 29 |
+
"kv_cache_dtype": "float16"
|
| 30 |
+
},
|
| 31 |
"use_cache": true,
|
| 32 |
"vocab_size": 49152
|
| 33 |
}
|
onnx/model_q4f16.onnx
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:bc95f62ea740d675d75c0f263ecf467c950f4002d18428dce832cb2fd5705b9e
|
| 3 |
+
size 298430898
|