Afrizal Hasbi Azizy
commited on
Commit
•
2b4cb5e
1
Parent(s):
a39e3c8
Update README.md
Browse files
README.md
CHANGED
@@ -13,7 +13,7 @@ language:
|
|
13 |
<center>
|
14 |
<img src="https://imgur.com/9nG5J1T.png" alt="Kancil" width="600" height="300">
|
15 |
<p><em>Kancil is a fine-tuned version of Llama 3 8B using synthetic QA dataset generated with Llama 3 70B. Version zero of Kancil is the first generative Indonesian LLM gain functional instruction performance using solely synthetic data.</em></p>
|
16 |
-
|
17 |
</center>
|
18 |
|
19 |
### Introducing the Kancil family of open models
|
@@ -88,14 +88,13 @@ pass
|
|
88 |
FastLanguageModel.for_inference(model)
|
89 |
inputs = tokenizer(
|
90 |
[
|
91 |
-
|
92 |
-
prompt="
|
93 |
-
response="",
|
94 |
-
)
|
95 |
], return_tensors = "pt").to("cuda")
|
96 |
|
97 |
-
outputs = model.generate(**inputs, max_new_tokens =
|
98 |
-
print(tokenizer.batch_decode(outputs)[0])
|
99 |
```
|
100 |
|
101 |
**Note:** There was an issue with the dataset such that newline characters are printed as string literals. Sorry about that!
|
|
|
13 |
<center>
|
14 |
<img src="https://imgur.com/9nG5J1T.png" alt="Kancil" width="600" height="300">
|
15 |
<p><em>Kancil is a fine-tuned version of Llama 3 8B using synthetic QA dataset generated with Llama 3 70B. Version zero of Kancil is the first generative Indonesian LLM gain functional instruction performance using solely synthetic data.</em></p>
|
16 |
+
<p><em><a href="https://colab.research.google.com/drive/1526QJYfk32X1CqYKX7IA_FFcIHLXbOkx?usp=sharing" style="color: blue;">Go straight to the colab demo</a></em></p>
|
17 |
</center>
|
18 |
|
19 |
### Introducing the Kancil family of open models
|
|
|
88 |
FastLanguageModel.for_inference(model)
|
89 |
inputs = tokenizer(
|
90 |
[
|
91 |
+
prompt_template.format(
|
92 |
+
prompt="Bagaimana canting dan malam digunakan untuk menggambar pola batik?",
|
93 |
+
response="",)
|
|
|
94 |
], return_tensors = "pt").to("cuda")
|
95 |
|
96 |
+
outputs = model.generate(**inputs, max_new_tokens = 600, temperature=.8, use_cache = True)
|
97 |
+
print(tokenizer.batch_decode(outputs)[0].replace('\\n', '\n'))
|
98 |
```
|
99 |
|
100 |
**Note:** There was an issue with the dataset such that newline characters are printed as string literals. Sorry about that!
|