liushaowei
commited on
Commit
·
b42671f
1
Parent(s):
579a3a3
update readme
Browse files
README.md
CHANGED
@@ -98,14 +98,14 @@ For our pretrained model (Moonlight-16B-A3B):
|
|
98 |
```python
|
99 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
100 |
|
101 |
-
|
102 |
model = AutoModelForCausalLM.from_pretrained(
|
103 |
-
|
104 |
torch_dtype="auto",
|
105 |
device_map="auto",
|
106 |
trust_remote_code=True,
|
107 |
)
|
108 |
-
tokenizer = AutoTokenizer.from_pretrained(
|
109 |
|
110 |
prompt = "1+1=2, 1+2="
|
111 |
inputs = tokenizer(prompt, return_tensors="pt", padding=True, truncation=True).to(model.device)
|
@@ -118,14 +118,14 @@ For our instruct model (Moonlight-16B-A3B-Instruct):
|
|
118 |
```python
|
119 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
120 |
|
121 |
-
|
122 |
model = AutoModelForCausalLM.from_pretrained(
|
123 |
-
|
124 |
torch_dtype="auto",
|
125 |
device_map="auto",
|
126 |
trust_remote_code=True
|
127 |
)
|
128 |
-
tokenizer = AutoTokenizer.from_pretrained(
|
129 |
|
130 |
prompt = "Give me a short introduction to large language model."
|
131 |
messages = [
|
|
|
98 |
```python
|
99 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
100 |
|
101 |
+
model_name = "moonshotai/Moonlight-16B-A3B"
|
102 |
model = AutoModelForCausalLM.from_pretrained(
|
103 |
+
model_name,
|
104 |
torch_dtype="auto",
|
105 |
device_map="auto",
|
106 |
trust_remote_code=True,
|
107 |
)
|
108 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
|
109 |
|
110 |
prompt = "1+1=2, 1+2="
|
111 |
inputs = tokenizer(prompt, return_tensors="pt", padding=True, truncation=True).to(model.device)
|
|
|
118 |
```python
|
119 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
120 |
|
121 |
+
model_name = "moonshotai/Moonlight-16B-A3B-Instruct"
|
122 |
model = AutoModelForCausalLM.from_pretrained(
|
123 |
+
model_name,
|
124 |
torch_dtype="auto",
|
125 |
device_map="auto",
|
126 |
trust_remote_code=True
|
127 |
)
|
128 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
|
129 |
|
130 |
prompt = "Give me a short introduction to large language model."
|
131 |
messages = [
|