Attempt to run multigpu in PR CI for now to ensure it works (#1815) [skip ci]

* Attempt to run multigpu in PR CI for now to ensure it works

* fix yaml file

* forgot to include multigpu tests

* fix call to cicd.multigpu

* dump dictdefault to dict for yaml conversion

* use to_dict instead of casting

* 16bit-lora w flash attention, 8bit lora seems problematic

* add llama fsdp test

* more tests

* Add test for qlora + fsdp with prequant

* limit accelerate to 2 processes and disable broken qlora+fsdp+bnb test

* move multigpu tests to biweekly
This commit is contained in:
Wing Lian
2024-08-09 11:50:13 -04:00
committed by GitHub
parent 3e2b269d06
commit 54392ac8a6
7 changed files with 473 additions and 4 deletions

View File

@@ -1,6 +1,8 @@
"""
modal application to run axolotl gpu tests in Modal
"""
# pylint: disable=duplicate-code
import os
import pathlib
import tempfile
@@ -21,9 +23,9 @@ df_template = template_env.get_template("Dockerfile.jinja")
df_args = {
"AXOLOTL_EXTRAS": os.environ.get("AXOLOTL_EXTRAS", ""),
"AXOLOTL_ARGS": os.environ.get("AXOLOTL_ARGS", ""),
"PYTORCH_VERSION": os.environ.get("PYTORCH_VERSION", "2.0.1"),
"BASE_TAG": os.environ.get("BASE_TAG", "main-base-py3.10-cu118-2.0.1"),
"CUDA": os.environ.get("CUDA", "118"),
"PYTORCH_VERSION": os.environ.get("PYTORCH_VERSION", "2.3.1"),
"BASE_TAG": os.environ.get("BASE_TAG", "main-base-py3.11-cu121-2.3.1"),
"CUDA": os.environ.get("CUDA", "121"),
"GITHUB_REF": os.environ.get("GITHUB_REF", "refs/heads/main"),
"GITHUB_SHA": os.environ.get("GITHUB_SHA", ""),
}