artificialguybr commited on
Commit
2c1bcda
1 Parent(s): 9f69e3b

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +3 -3
app.py CHANGED
@@ -9,12 +9,12 @@ import copy
9
  import secrets
10
  from pathlib import Path
11
 
12
- tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen-VL-Chat-Int4", trust_remote_code=True)
13
- config = AutoConfig.from_pretrained("Qwen/Qwen-VL-Chat-Int4", trust_remote_code=True)
14
  #config.quantization_config["use_exllama"] = True
15
  config.quantization_config["disable_exllama"] = False
16
  config.quantization_config["exllama_config"] = {"version":2}
17
- model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen-VL-Chat-Int4", device_map="auto", config=config).eval()
18
 
19
  BOX_TAG_PATTERN = r"<box>([\s\S]*?)</box>"
20
  PUNCTUATION = "!?。"#$%&'()*+,-/:;<=>@[\]^_`{|}~⦅⦆「」、、〃》「」『』【】〔〕〖〗〘〙〚〛〜〝〞〟〰〾〿–—‘’‛“”„‟…‧﹏."
 
9
  import secrets
10
  from pathlib import Path
11
 
12
+ tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen-VL-Chat-Int4")
13
+ config = AutoConfig.from_pretrained("Qwen/Qwen-VL-Chat-Int4", trust_remote_code=True, torch_dtype=torch.float16)
14
  #config.quantization_config["use_exllama"] = True
15
  config.quantization_config["disable_exllama"] = False
16
  config.quantization_config["exllama_config"] = {"version":2}
17
+ model = AutoModelForCausalLM.from_config(config, trust_remote_code=True, torch_dtype=torch.float16)
18
 
19
  BOX_TAG_PATTERN = r"<box>([\s\S]*?)</box>"
20
  PUNCTUATION = "!?。"#$%&'()*+,-/:;<=>@[\]^_`{|}~⦅⦆「」、、〃》「」『』【】〔〕〖〗〘〙〚〛〜〝〞〟〰〾〿–—‘’‛“”„‟…‧﹏."