Skip to content

Support download to host - amd llama2 #480

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
arjunsuresh merged 9 commits into from
Nov 7, 2024
Merged

Support download to host - amd llama2 #480

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
232a22c
initial commit - model download to host
anandhu-eng Nov 5, 2024
0564d6b
safe tensor path -> new env keys
anandhu-eng Nov 5, 2024
f3b6b6f
updated mounts for amd-llama2
anandhu-eng Nov 5, 2024
e743110
Updated env - llama2 model
anandhu-eng Nov 5, 2024
f10267f
handled model download in host
anandhu-eng Nov 5, 2024
c20c175
added compressed tensors module
anandhu-eng Nov 5, 2024
c60d748
Added compressed tensors support
anandhu-eng Nov 5, 2024
3228f88
bug fix
anandhu-eng Nov 5, 2024
b7a6e16
Update _cm.yaml
arjunsuresh Nov 7, 2024
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
9 changes: 8 additions & 1 deletion script/app-mlperf-inference-amd/_cm.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -243,9 +243,16 @@ variations:

llama2-70b_:
deps:
- tags: get,generic-python-lib,_package.compressed_tensors
names:
- compressed_tensors
- tags: get,preprocessed,dataset,openorca,_mlc,_validation
- tags: get,ml-model,llama2,_fp32,_pytorch
- tags: get,ml-model,llama2,_amd,_pytorch
skip_if_env:
CM_MLPERF_MODEL_LLAMA2_70B_DOWNLOAD_TO_HOST:
- 'yes'
CM_RUN_STATE_DOCKER:
- 'yes'
- tags: get,preprocessed,dataset,openorca,_mlc,_validation
- tags: download,file,_url.https://github.com/vllm-project/vllm/blob/38c4b7e863570a045308af814c72f4504297222e/tests/fp8_kv/llama2-70b-fp8-kv/kv_cache_scales.json
extra_cache_tags: llama2-scales,kv-cache
Expand Down
2 changes: 1 addition & 1 deletion script/app-mlperf-inference-amd/run-llama2.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ export VLLM_FP8_REDUCE_CONV=1
export HARNESS_DISABLE_VLLM_LOGS=1
export VLLM_LOGGING_LEVEL=ERROR

MODEL_PATH=${CM_ML_MODEL_LLAMA2_FILE_WITH_PATH:-/data/llm/llama2-70b-chat/}
MODEL_PATH=${LLAMA2_CHECKPOINT_PATH:-/data/llm/llama2-70b-chat/}
DATASET_PATH=${CM_DATASET_OPENORCA_PREPROCESSED_PATH:-/data/open_orca/open_orca_gpt4_tokenized_llama.sampled_24576.pkl.gz}
QUANTIZED_WEIGHTS_PATH=${CM_LLAMA2_FINAL_SAFE_TENSORS_PATH:-quantized/quark_share/modelzoo/llama2_70b_wfp8_afp8_ofp8_nomerge/json-safetensors/llama.safetensors}
QUANTIZATION_PARAM_PATH=${QUANTIZATION_PARAM_PATH:-/app/kv_cache_scales.json}
Expand Down
11 changes: 11 additions & 0 deletions script/app-mlperf-inference/_cm.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -810,6 +810,17 @@ variations:
CM_MLPERF_MODEL_LLAMA2_70B_DOWNLOAD_TO_HOST:
- 'yes'
tags: get,ml-model,llama2

llama2-70b_,amd:
docker:
image_name: mlperf-inference-amd-python-implementation-llama2-70b
mounts:
- "${{ CM_LLAMA2_FINAL_SAFE_TENSORS_PATH }}:${{ CM_LLAMA2_FINAL_SAFE_TENSORS_PATH }"
deps:
- enable_if_env:
CM_MLPERF_MODEL_LLAMA2_70B_DOWNLOAD_TO_HOST:
- 'yes'
tags: get,ml-model,llama2,_amd,_pytorch

mixtral-8x7b:
group:
Expand Down
6 changes: 5 additions & 1 deletion script/get-ml-model-llama2/_cm.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,8 @@
"new_env_keys": [
"CM_ML_MODEL_*",
"LLAMA2_CHECKPOINT_PATH",
"CM_NVIDIA_TP_SIZE"
"CM_NVIDIA_TP_SIZE",
"CM_LLAMA2_FINAL_SAFE_TENSORS_PATH"
],
"prehook_deps": [
{
Expand Down Expand Up @@ -188,6 +189,9 @@
},
{
"tags": "get,generic-python-lib,_torch"
},
{
"tags": "get,generic-python-lib,_package.compressed_tensors"
}
]
},
Expand Down