Skip to content

Format llama read me page #12782

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
facebook-github-bot merged 1 commit into from
Jul 25, 2025
+9 −8
Merged

Format llama read me page #12782

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 9 additions & 8 deletions examples/models/llama/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -168,7 +168,7 @@ LLAMA_CHECKPOINT=path/to/consolidated.00.pth
LLAMA_PARAMS=path/to/params.json

python -m extension.llm.export.export_llm \
--config examples/models/llamaconfig/llama_bf16.yaml
--config examples/models/llamaconfig/llama_bf16.yaml \
+base.model_class="llama3_2" \
+base.checkpoint="${LLAMA_CHECKPOINT:?}" \
+base.params="${LLAMA_PARAMS:?}" \
Expand All @@ -186,7 +186,7 @@ LLAMA_QUANTIZED_CHECKPOINT=path/to/spinquant/consolidated.00.pth.pth
LLAMA_PARAMS=path/to/spinquant/params.json

python -m extension.llm.export.export_llm \
--config examples/models/llama/config/llama_xnnpack_spinquant.yaml
--config examples/models/llama/config/llama_xnnpack_spinquant.yaml \
+base.model_class="llama3_2" \
+base.checkpoint="${LLAMA_QUANTIZED_CHECKPOINT:?}" \
+base.params="${LLAMA_PARAMS:?}"
Expand All @@ -203,7 +203,7 @@ LLAMA_QUANTIZED_CHECKPOINT=path/to/qlora/consolidated.00.pth.pth
LLAMA_PARAMS=path/to/qlora/params.json

python -m extension.llm.export.export_llm \
--config examples/models/llama/config/llama_xnnpack_qat.yaml
--config examples/models/llama/config/llama_xnnpack_qat.yaml \
+base.model_class="llama3_2" \
+base.checkpoint="${LLAMA_QUANTIZED_CHECKPOINT:?}" \
+base.params="${LLAMA_PARAMS:?}" \
Expand All @@ -219,15 +219,16 @@ You can export and run the original Llama 3 8B instruct model.
2. Export model and generate `.pte` file
```
python -m extension.llm.export.export_llm \
--config examples/models/llama/config/llama_q8da4w.yaml
+base.model_clas="llama3"
--config examples/models/llama/config/llama_q8da4w.yaml \
+base.model_class="llama3" \
+base.checkpoint=<consolidated.00.pth.pth> \
+base.params=<params.json>
```
Due to the larger vocabulary size of Llama 3, we recommend quantizing the embeddings with `quantization.embedding_quantize=\'4,32\'` as shown above to further reduce the model size.

Due to the larger vocabulary size of Llama 3, we recommend quantizing the embeddings with `quantization.embedding_quantize=\'4,32\'` as shown above to further reduce the model size.

If you're interested in deploying on non-CPU backends, [please refer the non-cpu-backend section](non_cpu_backends.md)

If you're interested in deploying on non-CPU backends, [please refer the non-cpu-backend section](non_cpu_backends.md)

## Step 3: Run on your computer to validate

Expand Down Expand Up @@ -450,7 +451,7 @@ python -m examples.models.llama.eval_llama \
-d <checkpoint dtype> \
--tasks mmlu \
--num_fewshot 5 \
--max_seq_len <max sequence length>
--max_seq_len <max sequence length> \
--max_context_len <max context length>
```

Expand Down
Loading