x54-729
commited on
Commit
•
70d038d
1
Parent(s):
a9418b3
fix example prompt
Browse files
README.md
CHANGED
@@ -135,7 +135,7 @@ tokenizer = AutoTokenizer.from_pretrained("internlm/internlm2-base-20b", trust_r
|
|
135 |
# `torch_dtype=torch.float16` 可以令模型以 float16 精度加载,否则 transformers 会将模型加载为 float32,有可能导致显存不足
|
136 |
model = AutoModelForCausalLM.from_pretrained("internlm/internlm2-base-20b", torch_dtype=torch.float16, trust_remote_code=True).cuda()
|
137 |
model = model.eval()
|
138 |
-
inputs = tokenizer(["
|
139 |
for k,v in inputs.items():
|
140 |
inputs[k] = v.cuda()
|
141 |
gen_kwargs = {"max_length": 128, "top_p": 0.8, "temperature": 0.8, "do_sample": True, "repetition_penalty": 1.0}
|
|
|
135 |
# `torch_dtype=torch.float16` 可以令模型以 float16 精度加载,否则 transformers 会将模型加载为 float32,有可能导致显存不足
|
136 |
model = AutoModelForCausalLM.from_pretrained("internlm/internlm2-base-20b", torch_dtype=torch.float16, trust_remote_code=True).cuda()
|
137 |
model = model.eval()
|
138 |
+
inputs = tokenizer(["来到美丽的大自然"], return_tensors="pt")
|
139 |
for k,v in inputs.items():
|
140 |
inputs[k] = v.cuda()
|
141 |
gen_kwargs = {"max_length": 128, "top_p": 0.8, "temperature": 0.8, "do_sample": True, "repetition_penalty": 1.0}
|