Upload Qwen2ForSequenceClassification
#2
by
michaelfeil
- opened
- config.json +11 -11
- model.safetensors +1 -1
config.json
CHANGED
@@ -1,18 +1,8 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "
|
3 |
"architectures": [
|
4 |
"Qwen2ForSequenceClassification"
|
5 |
],
|
6 |
-
"attention_dropout": 0.0,
|
7 |
-
"bos_token_id": 151643,
|
8 |
-
"eos_token_id": 151643,
|
9 |
-
"hidden_act": "silu",
|
10 |
-
"hidden_size": 896,
|
11 |
-
"initializer_range": 0.02,
|
12 |
-
"intermediate_size": 4864,
|
13 |
-
"max_position_embeddings": 131072,
|
14 |
-
"max_window_layers": 24,
|
15 |
-
"model_type": "qwen2",
|
16 |
"id2label": {
|
17 |
"0": "dummy_0",
|
18 |
"1": "dummy_1"
|
@@ -21,6 +11,16 @@
|
|
21 |
"dummy_0": 0,
|
22 |
"dummy_1": 1
|
23 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
"num_attention_heads": 14,
|
25 |
"num_hidden_layers": 24,
|
26 |
"num_key_value_heads": 2,
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "Qwen/Qwen2.5-0.5B-Instruct",
|
3 |
"architectures": [
|
4 |
"Qwen2ForSequenceClassification"
|
5 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6 |
"id2label": {
|
7 |
"0": "dummy_0",
|
8 |
"1": "dummy_1"
|
|
|
11 |
"dummy_0": 0,
|
12 |
"dummy_1": 1
|
13 |
},
|
14 |
+
"attention_dropout": 0.0,
|
15 |
+
"bos_token_id": 151643,
|
16 |
+
"eos_token_id": 151645,
|
17 |
+
"hidden_act": "silu",
|
18 |
+
"hidden_size": 896,
|
19 |
+
"initializer_range": 0.02,
|
20 |
+
"intermediate_size": 4864,
|
21 |
+
"max_position_embeddings": 32768,
|
22 |
+
"max_window_layers": 21,
|
23 |
+
"model_type": "qwen2",
|
24 |
"num_attention_heads": 14,
|
25 |
"num_hidden_layers": 24,
|
26 |
"num_key_value_heads": 2,
|
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1976170728
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:791acb14672e7ad4d1b6eaf48acf2891517cf0c48a56831a9a295343642c5814
|
3 |
size 1976170728
|