Add snip_momentum structured pruning example with 80% sparsity ratio#348
Merged
Conversation
Contributor
|
@ftian1, thanks for your patience. I have tested the method and it looks good. I will do some minor changes for the checkpoint saving. |
yaozhewei
added a commit
that referenced
this pull request
May 9, 2023
* add mask for generation, otherwise the generation is broken (#468) * Update model_utils.py (#471) * Support huggyllama/llama-7b with DSPipeline (#484) During token encoding for the huggyllama/llama-7b model, self.tokenizer.batch_encode_plus returns a token_type_ids kwarg that isn't recognized in the model's generate function. This is due to AutoTokenizer returning LlamaTokenizerFast instead of LlamaTokenizer as seen in the model tokenizer config: https://huggingface.co/huggyllama/llama-7b/blob/8416d3fefb0cb3ff5775a7b13c1692d10ff1aa16/tokenizer_config.json#L24 As a workaround, the DSPipeline will check to see if the tokenizer is LlamaTokenizerFast and only pass in input_tokens.input_ids to the self.model.generate(...) call if that is the case. * Add snip_momentum structured pruning example with 80% sparsity ratio (#348) --------- Co-authored-by: Jeff Rasley <jerasley@microsoft.com> Co-authored-by: Lev Kurilenko <113481193+lekurile@users.noreply.github.com> Co-authored-by: Tian, Feng <feng.tian@intel.com>
Contributor
Author
|
@xiaoxiawu-microsoft appreciate for your effort |
yaozhewei
added a commit
that referenced
this pull request
May 12, 2023
* add mask for generation, otherwise the generation is broken * . * merge master (#504) * add mask for generation, otherwise the generation is broken (#468) * Update model_utils.py (#471) * Support huggyllama/llama-7b with DSPipeline (#484) During token encoding for the huggyllama/llama-7b model, self.tokenizer.batch_encode_plus returns a token_type_ids kwarg that isn't recognized in the model's generate function. This is due to AutoTokenizer returning LlamaTokenizerFast instead of LlamaTokenizer as seen in the model tokenizer config: https://huggingface.co/huggyllama/llama-7b/blob/8416d3fefb0cb3ff5775a7b13c1692d10ff1aa16/tokenizer_config.json#L24 As a workaround, the DSPipeline will check to see if the tokenizer is LlamaTokenizerFast and only pass in input_tokens.input_ids to the self.model.generate(...) call if that is the case. * Add snip_momentum structured pruning example with 80% sparsity ratio (#348) --------- Co-authored-by: Jeff Rasley <jerasley@microsoft.com> Co-authored-by: Lev Kurilenko <113481193+lekurile@users.noreply.github.com> Co-authored-by: Tian, Feng <feng.tian@intel.com> * . * change eos * . * resolve tokenizer issue * add script * . --------- Co-authored-by: Jeff Rasley <jerasley@microsoft.com> Co-authored-by: Lev Kurilenko <113481193+lekurile@users.noreply.github.com> Co-authored-by: Tian, Feng <feng.tian@intel.com>
Syulin7
pushed a commit
to Syulin7/DeepSpeedExamples
that referenced
this pull request
May 15, 2023
leocnj
pushed a commit
to leocnj/DeepSpeedExamples
that referenced
this pull request
May 27, 2023
hwchen2017
pushed a commit
that referenced
this pull request
Jun 8, 2025
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
This PR is used to demonstrate the functionality of snip_momentum structured pruning algo implemented in here.
User can reproduce below result by running
source ./bash_script/pruning_sparse_snip_momentum.shwith the PR mentioned at above.