Jasper and Stella: distillation of SOTA embedding models #67
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
DunZhang
reviewed
Dec 30, 2024
| return f'Instruct: {task_description}\nQuery: {query}' | ||
|
|
||
|
|
||
| # Each query must come with a one-sentence instruction that describes the task |
There was a problem hiding this comment.
What if the teacher model doesn't need a prompt?
You can use a more detailed comment.
There was a problem hiding this comment.
Thanks for the reminder, I will add it to the readme note.
| dic = set() | ||
| with open(train_data_path) as f: | ||
| for line in tqdm.tqdm(f): | ||
| data_dic=json.loads(line.strip()) |
There was a problem hiding this comment.
data_dict=json.loads(line)
json.loads does not need strip
Complete variable names will never be an error
|
|
||
|
|
||
| if 'pos' in data_dic: | ||
| for text_pos in data_dic['pos']: |
There was a problem hiding this comment.
for text_pos in data_dic.get('pos',[]):
| trust_remote_code=True, | ||
| device="cuda:7", | ||
| model_kwargs={ | ||
| "torch_dtype": torch.bfloat16, # fp16 容易计算出nan |
There was a problem hiding this comment.
remove Chinese character
remove this comment; the experience is not suitable for every model.
| return loss | ||
|
|
||
| def pair_inbatch_similarity_loss( | ||
| self, |
| student_embeddings, # [batch_size,dim] | ||
| teacher_similarity, # [batch_size,dim] | ||
| ): | ||
| loss_fct = nn.MSELoss() |
There was a problem hiding this comment.
每次计算loss都要初始化一个nn.MSELoss()吗？如果编译器没有自动优化这个，我建议用F.mse_loss
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
todo: