Upload Qwen3ForSequenceClassification
Browse files- config.json +6 -2
- model.safetensors +1 -1
config.json
CHANGED
|
@@ -15,6 +15,10 @@
|
|
| 15 |
"label2id": {
|
| 16 |
"LABEL_0": 0
|
| 17 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
"max_position_embeddings": 32768,
|
| 19 |
"max_window_layers": 28,
|
| 20 |
"model_type": "qwen3",
|
|
@@ -24,10 +28,10 @@
|
|
| 24 |
"rms_norm_eps": 1e-06,
|
| 25 |
"rope_scaling": null,
|
| 26 |
"rope_theta": 10000.0,
|
| 27 |
-
"sliding_window":
|
| 28 |
"tie_word_embeddings": false,
|
| 29 |
"torch_dtype": "float32",
|
| 30 |
-
"transformers_version": "4.
|
| 31 |
"use_cache": true,
|
| 32 |
"use_sliding_window": false,
|
| 33 |
"vocab_size": 151669
|
|
|
|
| 15 |
"label2id": {
|
| 16 |
"LABEL_0": 0
|
| 17 |
},
|
| 18 |
+
"layer_types": [
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention"
|
| 21 |
+
],
|
| 22 |
"max_position_embeddings": 32768,
|
| 23 |
"max_window_layers": 28,
|
| 24 |
"model_type": "qwen3",
|
|
|
|
| 28 |
"rms_norm_eps": 1e-06,
|
| 29 |
"rope_scaling": null,
|
| 30 |
"rope_theta": 10000.0,
|
| 31 |
+
"sliding_window": null,
|
| 32 |
"tie_word_embeddings": false,
|
| 33 |
"torch_dtype": "float32",
|
| 34 |
+
"transformers_version": "4.55.0.dev0",
|
| 35 |
"use_cache": true,
|
| 36 |
"use_sliding_window": false,
|
| 37 |
"vocab_size": 151669
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4962680
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a0ecdf32244a8a38b46dded58f9572d07fa3cce579d2fae5051611a9de48114
|
| 3 |
size 4962680
|