github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-01-09 15:08:02 -05:00
Code Issues Packages Projects Releases Wiki Activity

reenable LLaMA-3 8B BEAM on NV (#5746)

Browse Source
This commit is contained in:
George Hotz
2024-07-26 16:56:41 -07:00
committed by GitHub
parent c6b2d96474
commit db1d093b29
1 changed files with 2 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
.github/workflows/benchmark.yml vendored
View File

@@ -182,9 +182,8 @@ jobs:
run: NV=1 python3 examples/llama.py --gen 1 --size 7B --shard 4 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_four_gpu.txt
- name: Run LLaMA 7B on 6 GPUs
run: NV=1 python3 examples/llama.py --gen 1 --size 7B --shard 6 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_six_gpu.txt
# TODO: this is flaky
# - name: Run LLaMA-3 8B BEAM
# run: NV=1 JITBEAM=2 IGNORE_BEAM_CACHE=1 python3 examples/llama3.py --model weights/LLaMA-3/8B-SF-DPO/ --benchmark | tee llama3_beam.txt
- name: Run LLaMA-3 8B BEAM
run: NV=1 JITBEAM=2 IGNORE_BEAM_CACHE=1 python3 examples/llama3.py --model weights/LLaMA-3/8B-SF-DPO/ --benchmark | tee llama3_beam.txt
- name: Run LLaMA-3 8B on 4 GPUs
run: NV=1 python3 examples/llama3.py --shard 4 --model weights/LLaMA-3/8B-SF-DPO/ --benchmark | tee llama3_four_gpu.txt
- name: Run LLaMA-3 8B on 6 GPUs