[Qeff.finetuning] Adding Full document for hf_based finetuning stack#732
Open
tchawada wants to merge 1 commit intoquic:ft_experimentalfrom
Open
[Qeff.finetuning] Adding Full document for hf_based finetuning stack#732tchawada wants to merge 1 commit intoquic:ft_experimentalfrom
tchawada wants to merge 1 commit intoquic:ft_experimentalfrom
Conversation
Signed-off-by: Tanisha Chawada <tchawada@qti.qualcomm.com>
tchawada
requested review from
ochougul,
quic-amitraj,
quic-hemagnih and
quic-rishinr
as code owners
| fsdp: "full_shard" | ||
| fsdp_config: "./configs/accelerate/fsdp_config.yaml" | ||
| fsdp_config: "./configs/accelerate/fsdp_tp_parallelism_config.yaml" | ||
| ``` |
Contributor
There was a problem hiding this comment.
We used to support PP+DDP also in earlier stack. Test and verify the same in new stack, and include it
|
|
||
| * **tokenizer\_name**: `default = "HuggingFaceTB/SmolLM-135M"` → Matches model name. | ||
| * **dataset\_type**: `default = "seq_completion"` → Used for sequence continuation tasks, where the language model learns to generate the correct output (completion) step by step, given an input (prompt). | ||
| * **dataset\_name**: `default = "knkarthick/samsum"` → Dataset name for training. |
Contributor
There was a problem hiding this comment.
We need to add additional num_samples to slect only a subset of the provided dataset, especially in testing scenario.
| ## Parallelism | ||
|
|
||
| The training script supports multiple parallelism strategies: | ||
|
|
Contributor
There was a problem hiding this comment.
Include the sample commands and steps to run all these different strategies
| python finetune.py --config configs/sample_config.yaml | ||
| ``` | ||
|
|
||
| **Distributed (TorchRun)** |
Contributor
There was a problem hiding this comment.
We need to add MULTI NODE FT based documentation also.
Contributor
There was a problem hiding this comment.
Please refer the existing stack finetune.md
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
2 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 adds a complete, professionally structured documentation set for the HF-Based QEfficient Finetune Module.
It includes installation steps, dataset preparation workflows, custom dataset support, Hugging Face dataset formatting, prompt template guidance, finetuning instructions, inference usage.