Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
@@ -15,6 +15,8 @@ from transformers import (
|
|
15 |
import subprocess
|
16 |
subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
|
17 |
|
|
|
|
|
18 |
DESCRIPTION = """\
|
19 |
# ORLM LLaMA-3-8B
|
20 |
|
@@ -24,6 +26,7 @@ Hello! I'm ORLM-LLaMA-3-8B, here to automate your optimization modeling tasks! C
|
|
24 |
MAX_MAX_NEW_TOKENS = 4096
|
25 |
DEFAULT_MAX_NEW_TOKENS = 4096
|
26 |
MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
|
|
|
27 |
|
28 |
# quantization_config = BitsAndBytesConfig(
|
29 |
# load_in_4bit=True,
|
@@ -32,19 +35,21 @@ MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
|
|
32 |
# bnb_4bit_quant_type= "nf4")
|
33 |
# quantization_config = BitsAndBytesConfig(load_in_8bit=True)
|
34 |
|
35 |
-
model_id =
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
)
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
|
|
|
|
48 |
def generate(
|
49 |
message: str,
|
50 |
chat_history: list[tuple[str, str]],
|
@@ -57,33 +62,33 @@ def generate(
|
|
57 |
if chat_history != []:
|
58 |
return "Sorry, I am an instruction-tuned model and currently do not support chatting. Please try clearing the chat history or refreshing the page to ask a new question."
|
59 |
|
60 |
-
tokenized_example = tokenizer(message, return_tensors='pt', max_length=MAX_INPUT_TOKEN_LENGTH, truncation=True)
|
61 |
-
input_ids = tokenized_example.input_ids
|
62 |
-
input_ids = input_ids.to(model.device)
|
63 |
-
|
64 |
-
streamer = TextIteratorStreamer(tokenizer, timeout=50.0, skip_prompt=True, skip_special_tokens=True)
|
65 |
-
generate_kwargs = dict(
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
)
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
|
88 |
|
89 |
chat_interface = gr.ChatInterface(
|
|
|
15 |
import subprocess
|
16 |
subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
|
17 |
|
18 |
+
from vllm import LLM, SamplingParams
|
19 |
+
|
20 |
DESCRIPTION = """\
|
21 |
# ORLM LLaMA-3-8B
|
22 |
|
|
|
26 |
MAX_MAX_NEW_TOKENS = 4096
|
27 |
DEFAULT_MAX_NEW_TOKENS = 4096
|
28 |
MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
|
29 |
+
model_id = "CardinalOperations/ORLM-LLaMA-3-8B"
|
30 |
|
31 |
# quantization_config = BitsAndBytesConfig(
|
32 |
# load_in_4bit=True,
|
|
|
35 |
# bnb_4bit_quant_type= "nf4")
|
36 |
# quantization_config = BitsAndBytesConfig(load_in_8bit=True)
|
37 |
|
38 |
+
# tokenizer = AutoTokenizer.from_pretrained(model_id, use_fast=True)
|
39 |
+
# model = AutoModelForCausalLM.from_pretrained(
|
40 |
+
# model_id,
|
41 |
+
# device_map="auto",
|
42 |
+
# torch_dtype=torch.bfloat16,
|
43 |
+
# attn_implementation="flash_attention_2",
|
44 |
+
# # quantization_config=quantization_config,
|
45 |
+
# )
|
46 |
+
# model.eval()
|
47 |
+
|
48 |
+
subprocess.run(f'huggingface-cli download {model_id} --local_dir ./local_model', shell=True)
|
49 |
+
model = LLM(model='./local_model', tensor_parallel_size=torch.cuda.device_count())
|
50 |
+
print("init model done.")
|
51 |
+
|
52 |
+
@spaces.GPU(duration=60)
|
53 |
def generate(
|
54 |
message: str,
|
55 |
chat_history: list[tuple[str, str]],
|
|
|
62 |
if chat_history != []:
|
63 |
return "Sorry, I am an instruction-tuned model and currently do not support chatting. Please try clearing the chat history or refreshing the page to ask a new question."
|
64 |
|
65 |
+
# tokenized_example = tokenizer(message, return_tensors='pt', max_length=MAX_INPUT_TOKEN_LENGTH, truncation=True)
|
66 |
+
# input_ids = tokenized_example.input_ids
|
67 |
+
# input_ids = input_ids.to(model.device)
|
68 |
+
|
69 |
+
# streamer = TextIteratorStreamer(tokenizer, timeout=50.0, skip_prompt=True, skip_special_tokens=True)
|
70 |
+
# generate_kwargs = dict(
|
71 |
+
# {"input_ids": input_ids},
|
72 |
+
# streamer=streamer,
|
73 |
+
# max_new_tokens=max_new_tokens,
|
74 |
+
# do_sample=False if temperature == 0.0 else True,
|
75 |
+
# top_p=top_p,
|
76 |
+
# top_k=top_k,
|
77 |
+
# temperature=temperature,
|
78 |
+
# num_beams=1,
|
79 |
+
# repetition_penalty=repetition_penalty,
|
80 |
+
# eos_token_id=[tok.eos_token_id],
|
81 |
+
# )
|
82 |
+
|
83 |
+
prompts = [message]
|
84 |
+
stop_tokens = ["</s>"]
|
85 |
+
if temperature == 0.0:
|
86 |
+
sampling_params = SamplingParams(n=topk, temperature=0, top_p=1, repetition_penalty=repetition_penalty, max_tokens=max_new_tokens, stop=stop_tokens)
|
87 |
+
else:
|
88 |
+
sampling_params = SamplingParams(n=topk, temperature=temperature, top_p=top_p, repetition_penalty=repetition_penalty, max_tokens=max_new_tokens, stop=stop_tokens)
|
89 |
+
generations = model.generate(prompts, sampling_params)
|
90 |
+
outputs = [g.outputs[0].text for g in generations]
|
91 |
+
return outputs[0]
|
92 |
|
93 |
|
94 |
chat_interface = gr.ChatInterface(
|