GGUF fp32/fp16 conversion to checkpoint #134
Open
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
Summary: Only works for fp32 and fp16 types so that means, It isn't providing much value right now because `convert_hf_checkpoint.py` can directly generate an equivalent .pth checkpoint file without gguf format indirection. In the future, we will support running the quantized version of the graph. Test Plan: 0. Setup pip install gguf git@github.com:ggerganov/llama.cpp.git 1. Preparation: convert existing hf model to fp16 `python llama.cpp/convert.py [HF-dir] --outtype f16`` which will generate [HF-dir]/ggml-model-f16.gguf 2. Convert GGUF file to a checkpoint `python scripts/convert_from_gguf.py --gguf_file [HF-dir]/ggml-model-f16.gguf --checkpoint_file [HF-dir]/model_gguf.pth` 3. Validate that it works: `python generate.py --checkpoint_path [HF-dir]/model_gguf.pth --device=cpu --prompt "Hello, my name is" --max_new_tokens 20`
facebook-github-bot
added
the
CLA Signed
|
Why import GGUF when one can do decode in place using native PyTorch, see https://github.com/malfet/llm_experiments/blob/74a935344fbce5680dbd2dafc7dfd95231303444/run_llama.py#L447 |
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Outdated suggestions cannot be applied.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.
Summary:
Only works for fp32 and fp16 types so that means it isn't providing much value right now.
convert_hf_checkpoint.pycan already directly generate an equivalent .pth checkpoint file without gguf format indirection.However this PR just creates the foundation and validation that the basic fp32 and fp16 works fine. In the future, we will support running the quantized version of the gguf graph in eager.
Test Plan:
pip install ggufgit clone git@github.com:ggerganov/llama.cpp.gitpython scripts/download.py --repo_id [HF-dir]python llama.cpp/convert.py [HF-dir] --outtype f16`` which will generate [HF-dir]/ggml-model-f16.ggufpython scripts/convert_from_gguf.py --gguf_file [HF-dir]/ggml-model-f16.gguf --checkpoint_file [HF-dir]/model_gguf.pthpython generate.py --checkpoint_path [HF-dir]/model_gguf.pth --device=cpu --prompt "Hello, my name is" --max_new_tokens 20