feihu.hf
commited on
Commit
•
5b5ffe0
1
Parent(s):
b7c57fd
add torch_dtype
Browse files
README.md
CHANGED
|
@@ -48,6 +48,7 @@ device = "cuda" # the device to load the model onto
|
|
| 48 |
|
| 49 |
model = AutoModelForCausalLM.from_pretrained(
|
| 50 |
"Qwen/Qwen1.5-72B-Chat-GPTQ-Int8",
|
|
|
|
| 51 |
device_map="auto"
|
| 52 |
)
|
| 53 |
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen1.5-72B-Chat-GPTQ-Int8")
|
|
|
|
| 48 |
|
| 49 |
model = AutoModelForCausalLM.from_pretrained(
|
| 50 |
"Qwen/Qwen1.5-72B-Chat-GPTQ-Int8",
|
| 51 |
+
torch_dtype="auto",
|
| 52 |
device_map="auto"
|
| 53 |
)
|
| 54 |
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen1.5-72B-Chat-GPTQ-Int8")
|