sealical commited on
Commit
da1d18b
·
verified ·
1 Parent(s): 58af987

Delete .ipynb_checkpoints

Browse files
.ipynb_checkpoints/config-checkpoint.json DELETED
@@ -1,211 +0,0 @@
1
- {
2
- "_commit_hash": null,
3
- "architectures": [
4
- "SkyworkChatModel"
5
- ],
6
- "auto_map": {
7
- "AutoConfig": "configuration_skywork_chat.SkyworkChatConfig",
8
- "AutoModel": "modeling_skywork_chat.SkyworkChatModel",
9
- "AutoModelForCausalLM": "modeling_skywork_chat.SkyworkChatModel"
10
- },
11
- "downsample_ratio": 0.5,
12
- "dynamic_image_size": true,
13
- "force_image_size": 448,
14
- "hidden_size": 5120,
15
- "llm_config": {
16
- "_attn_implementation_autoset": true,
17
- "_name_or_path": "",
18
- "add_cross_attention": false,
19
- "architectures": [
20
- "Qwen2ForCausalLM"
21
- ],
22
- "attention_dropout": 0.0,
23
- "attn_implementation": "flash_attention_2",
24
- "auto_map": {
25
- "AutoConfig": "configuration_skywork_chat.SkyworkChatConfig",
26
- "AutoModel": "modeling_skywork_chat.SkyworkChatModel",
27
- "AutoModelForCausalLM": "modeling_skywork_chat.SkyworkChatModel",
28
- "AutoModelForSequenceClassification": "modeling_skywork_lm2.SkyworkLM2ForSequenceClassification"
29
- },
30
- "bad_words_ids": null,
31
- "begin_suppress_tokens": null,
32
- "bos_token_id": 151643,
33
- "chunk_size_feed_forward": 0,
34
- "cross_attention_hidden_size": null,
35
- "decoder_start_token_id": null,
36
- "diversity_penalty": 0.0,
37
- "do_sample": false,
38
- "early_stopping": false,
39
- "encoder_no_repeat_ngram_size": 0,
40
- "eos_token_id": 151643,
41
- "exponential_decay_length_penalty": null,
42
- "finetuning_task": null,
43
- "forced_bos_token_id": null,
44
- "forced_eos_token_id": null,
45
- "hidden_act": "silu",
46
- "hidden_size": 5120,
47
- "id2label": {
48
- "0": "LABEL_0",
49
- "1": "LABEL_1"
50
- },
51
- "initializer_range": 0.02,
52
- "intermediate_size": 27648,
53
- "is_decoder": false,
54
- "is_encoder_decoder": false,
55
- "label2id": {
56
- "LABEL_0": 0,
57
- "LABEL_1": 1
58
- },
59
- "length_penalty": 1.0,
60
- "max_length": 20,
61
- "max_position_embeddings": 40960,
62
- "max_window_layers": 64,
63
- "min_length": 0,
64
- "model_type": "qwen2",
65
- "no_repeat_ngram_size": 0,
66
- "num_attention_heads": 40,
67
- "num_beam_groups": 1,
68
- "num_beams": 1,
69
- "num_hidden_layers": 64,
70
- "num_key_value_heads": 8,
71
- "num_return_sequences": 1,
72
- "output_attentions": false,
73
- "output_hidden_states": false,
74
- "output_scores": false,
75
- "pad_token_id": null,
76
- "prefix": null,
77
- "problem_type": null,
78
- "pruned_heads": {},
79
- "quantization_config": {
80
- "bits": 4,
81
- "group_size": 128,
82
- "quant_method": "awq",
83
- "version": "gemm",
84
- "zero_point": true
85
- },
86
- "remove_invalid_values": false,
87
- "repetition_penalty": 1.0,
88
- "return_dict": true,
89
- "return_dict_in_generate": false,
90
- "rms_norm_eps": 1e-05,
91
- "rope_scaling": null,
92
- "rope_theta": 1000000.0,
93
- "sep_token_id": null,
94
- "sliding_window": null,
95
- "suppress_tokens": null,
96
- "task_specific_params": null,
97
- "temperature": 1.0,
98
- "tf_legacy_loss": false,
99
- "tie_encoder_decoder": false,
100
- "tie_word_embeddings": false,
101
- "tokenizer_class": null,
102
- "top_k": 50,
103
- "top_p": 1.0,
104
- "torch_dtype": "bfloat16",
105
- "torchscript": false,
106
- "transformers_version": "4.50.0",
107
- "typical_p": 1.0,
108
- "use_bfloat16": false,
109
- "use_cache": false,
110
- "use_sliding_window": false,
111
- "vocab_size": 152064
112
- },
113
- "max_dynamic_patch": 12,
114
- "min_dynamic_patch": 1,
115
- "model_type": "skywork_chat",
116
- "pad2square": false,
117
- "ps_version": "v2",
118
- "select_layer": -1,
119
- "template": "skywork-r1v-chat",
120
- "tie_word_embeddings": false,
121
- "torch_dtype": "float16",
122
- "transformers_version": null,
123
- "use_backbone_lora": 0,
124
- "use_llm_lora": 0,
125
- "use_thumbnail": true,
126
- "vision_config": {
127
- "_attn_implementation_autoset": true,
128
- "_name_or_path": "",
129
- "add_cross_attention": false,
130
- "architectures": [
131
- "InternVisionModel"
132
- ],
133
- "attention_dropout": 0.0,
134
- "bad_words_ids": null,
135
- "begin_suppress_tokens": null,
136
- "bos_token_id": null,
137
- "chunk_size_feed_forward": 0,
138
- "cross_attention_hidden_size": null,
139
- "decoder_start_token_id": null,
140
- "diversity_penalty": 0.0,
141
- "do_sample": false,
142
- "drop_path_rate": 0.4,
143
- "dropout": 0.0,
144
- "early_stopping": false,
145
- "encoder_no_repeat_ngram_size": 0,
146
- "eos_token_id": null,
147
- "exponential_decay_length_penalty": null,
148
- "finetuning_task": null,
149
- "forced_bos_token_id": null,
150
- "forced_eos_token_id": null,
151
- "hidden_act": "gelu",
152
- "hidden_size": 3200,
153
- "id2label": {
154
- "0": "LABEL_0",
155
- "1": "LABEL_1"
156
- },
157
- "image_size": 448,
158
- "initializer_factor": 0.1,
159
- "initializer_range": 1e-10,
160
- "intermediate_size": 12800,
161
- "is_decoder": false,
162
- "is_encoder_decoder": false,
163
- "label2id": {
164
- "LABEL_0": 0,
165
- "LABEL_1": 1
166
- },
167
- "layer_norm_eps": 1e-06,
168
- "length_penalty": 1.0,
169
- "max_length": 20,
170
- "min_length": 0,
171
- "model_type": "",
172
- "no_repeat_ngram_size": 0,
173
- "norm_type": "rms_norm",
174
- "num_attention_heads": 25,
175
- "num_beam_groups": 1,
176
- "num_beams": 1,
177
- "num_channels": 3,
178
- "num_hidden_layers": 45,
179
- "num_return_sequences": 1,
180
- "output_attentions": false,
181
- "output_hidden_states": false,
182
- "output_scores": false,
183
- "pad_token_id": null,
184
- "patch_size": 14,
185
- "prefix": null,
186
- "problem_type": null,
187
- "pruned_heads": {},
188
- "qk_normalization": true,
189
- "qkv_bias": false,
190
- "remove_invalid_values": false,
191
- "repetition_penalty": 1.0,
192
- "return_dict": true,
193
- "return_dict_in_generate": false,
194
- "sep_token_id": null,
195
- "suppress_tokens": null,
196
- "task_specific_params": null,
197
- "temperature": 1.0,
198
- "tf_legacy_loss": false,
199
- "tie_encoder_decoder": false,
200
- "tie_word_embeddings": true,
201
- "tokenizer_class": null,
202
- "top_k": 50,
203
- "top_p": 1.0,
204
- "torch_dtype": "bfloat16",
205
- "torchscript": false,
206
- "transformers_version": "4.50.0",
207
- "typical_p": 1.0,
208
- "use_bfloat16": true,
209
- "use_flash_attn": true
210
- }
211
- }