Update README.md
Browse files
README.md
CHANGED
@@ -24,18 +24,18 @@ anh-bloomz-7b1-mt-cross-lingual model can be loaded and used via the following c
|
|
24 |
```python
|
25 |
import re
|
26 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
27 |
-
|
28 |
-
|
29 |
-
)
|
30 |
-
tokenizer = AutoTokenizer.from_pretrained(
|
31 |
-
|
32 |
-
)
|
33 |
whitespace_tokens_map = {'\n': '<n>', ' ': '<w>'}
|
34 |
-
text = "User:
|
35 |
for k, v in whitespace_tokens_map.items():
|
36 |
text = text.replace(k, v)
|
37 |
inputs = tokenizer(text, return_tensors="pt")
|
38 |
-
tokens = model.generate(**inputs
|
|
|
39 |
output = tokenizer.decode(tokens[0], skip_special_tokens=True)
|
40 |
for v in whitespace_tokens_map.values():
|
41 |
output = re.sub(rf"{v}\s+(\S+)", rf"{v}\1", output)
|
|
|
24 |
```python
|
25 |
import re
|
26 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
27 |
+
|
28 |
+
model_name = "laion/anh-bloomz-7b1-mt-cross-lingual"
|
29 |
+
model = AutoModelForCausalLM.from_pretrained(model_name)
|
30 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
31 |
+
|
|
|
32 |
whitespace_tokens_map = {'\n': '<n>', ' ': '<w>'}
|
33 |
+
text = "User: Apakah kita akan bisa menyembuhkan penyakit kanker? Jawab dalam bahasa China.\n"
|
34 |
for k, v in whitespace_tokens_map.items():
|
35 |
text = text.replace(k, v)
|
36 |
inputs = tokenizer(text, return_tensors="pt")
|
37 |
+
tokens = model.generate(**inputs, max_new_tokens=200, do_sample=True, top_k=40, top_p=0.9, temperature=0.2,
|
38 |
+
repetition_penalty=1.2,num_return_sequences=1)
|
39 |
output = tokenizer.decode(tokens[0], skip_special_tokens=True)
|
40 |
for v in whitespace_tokens_map.values():
|
41 |
output = re.sub(rf"{v}\s+(\S+)", rf"{v}\1", output)
|