Update README.md
Browse files
README.md
CHANGED
@@ -96,7 +96,13 @@ from peft import PeftModel
|
|
96 |
model_id = "unsloth/gemma-2-9b-it-bnb-4bit"
|
97 |
peft_model_id = "webbigdata/C3TR-Adapter"
|
98 |
|
99 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
100 |
model = PeftModel.from_pretrained(model = model, model_id = peft_model_id)
|
101 |
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
102 |
tokenizer.pad_token = tokenizer.unk_token
|
@@ -136,7 +142,7 @@ The prompt format is original.
|
|
136 |
Version1とVersion2(システムプロンプト追加)とVersion3(特殊トークン追加)ではプロンプトフォーマットも変わっています。
|
137 |
The prompt format has changed between Version 1, Version2(add sustem prompts) and Version 3(add special tokens).
|
138 |
|
139 |
-
|
140 |
Extra spaces, line breaks, and omission of special tokens in the prompt template will cause the model to malfunction (output will be truncated or repeated, extra sentences will be added, etc.), so please make sure there are no errors in the template.
|
141 |
|
142 |
Instructionsは"Translate Japanese to English."(日英翻訳)と"Translate English to Japanese."(英日翻訳)の2種類です。
|
|
|
96 |
model_id = "unsloth/gemma-2-9b-it-bnb-4bit"
|
97 |
peft_model_id = "webbigdata/C3TR-Adapter"
|
98 |
|
99 |
+
# デバイスがbfloat16をサポートしているかどうかを確認
|
100 |
+
if torch.cuda.is_available() and torch.cuda.get_device_capability(0)[0] >= 8:
|
101 |
+
dtype = torch.bfloat16
|
102 |
+
else:
|
103 |
+
dtype = torch.float16
|
104 |
+
|
105 |
+
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=dtype, device_map="auto")
|
106 |
model = PeftModel.from_pretrained(model = model, model_id = peft_model_id)
|
107 |
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
108 |
tokenizer.pad_token = tokenizer.unk_token
|
|
|
142 |
Version1とVersion2(システムプロンプト追加)とVersion3(特殊トークン追加)ではプロンプトフォーマットも変わっています。
|
143 |
The prompt format has changed between Version 1, Version2(add sustem prompts) and Version 3(add special tokens).
|
144 |
|
145 |
+
プロンプトテンプレート内に余分な空白や改行、特殊トークンの漏れはモデルの誤動作(出力が途切れたり繰り返す、余分な文章が付加される等)に繋がるのでテンプレートにミスがないようにしてください
|
146 |
Extra spaces, line breaks, and omission of special tokens in the prompt template will cause the model to malfunction (output will be truncated or repeated, extra sentences will be added, etc.), so please make sure there are no errors in the template.
|
147 |
|
148 |
Instructionsは"Translate Japanese to English."(日英翻訳)と"Translate English to Japanese."(英日翻訳)の2種類です。
|