Spaces:
Sleeping
Sleeping
Update app.py
Browse filesFugaku-LLM-13B-instruct-gguf
app.py
CHANGED
@@ -16,7 +16,7 @@ MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
|
|
16 |
DESCRIPTION = """\
|
17 |
# magarpr/Fugaku-LLM-Fugaku-LLM-13B - ISDX CET
|
18 |
|
19 |
-
This Space demonstrates model [Fugaku-LLM
|
20 |
|
21 |
|
22 |
|
@@ -35,7 +35,7 @@ if not torch.cuda.is_available():
|
|
35 |
if torch.cuda.is_available():
|
36 |
# Login to Hugging Face
|
37 |
login(token=os.getenv("HUGGINGFACE_HUB_TOKEN"))
|
38 |
-
model_id = "Fugaku-LLM/Fugaku-LLM-13B-instruct"
|
39 |
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", load_in_4bit=True)
|
40 |
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
41 |
tokenizer.use_default_system_prompt = False
|
|
|
16 |
DESCRIPTION = """\
|
17 |
# magarpr/Fugaku-LLM-Fugaku-LLM-13B - ISDX CET
|
18 |
|
19 |
+
This Space demonstrates model [Fugaku-LLM/Fugaku-LLM-13B-instruct-gguf) , 13B parameters.
|
20 |
|
21 |
|
22 |
|
|
|
35 |
if torch.cuda.is_available():
|
36 |
# Login to Hugging Face
|
37 |
login(token=os.getenv("HUGGINGFACE_HUB_TOKEN"))
|
38 |
+
model_id = "Fugaku-LLM/Fugaku-LLM-13B-instruct-gguf"
|
39 |
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", load_in_4bit=True)
|
40 |
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
41 |
tokenizer.use_default_system_prompt = False
|