tocmo0nlord/axolotl
1
0
Fork 0
Code Issues Pull Requests Actions 3 Packages Projects Releases Wiki Activity

Various fixes for CI, save_only_model for RL, prevent packing multiprocessing deadlocks (#2661)

Browse Source 
* lean mistral ft tests, remove e2e torch 2.4.1 test

* make sure to pass save_only_model for RL

* more tests to make ci leaner, add cleanup to modal ci

* fix module for import in e2e tests

* use mp spawn to prevent deadlocks with packing

* make sure cleanup shell script is executable when cloned out
This commit is contained in:
Wing Lian
2025-05-12 10:51:18 -04:00
parent 27e3329273
commit cf69de2eb9
13 changed files with 190 additions and 93 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
tests/e2e/patched/test_4d_multipack_llama.py
View File

@@ -57,9 +57,9 @@ class Test4dMultipackLlama(unittest.TestCase):
"learning_rate": 0.00001,
"optimizer": "adamw_torch_fused",
"lr_scheduler": "cosine",
"max_steps": 20,
"save_steps": 10,
"eval_steps": 10,
"max_steps": 5,
"save_steps": 3,
"eval_steps": 4,
"fp16": True,
}
)
@@ -105,9 +105,9 @@ class Test4dMultipackLlama(unittest.TestCase):
"learning_rate": 0.00001,
"optimizer": "adamw_torch_fused",
"lr_scheduler": "cosine",
"max_steps": 20,
"save_steps": 10,
"eval_steps": 10,
"max_steps": 5,
"save_steps": 3,
"eval_steps": 4,
"fp16": True,
}
)

12
tests/e2e/patched/test_mistral_samplepack.py
View File

@@ -57,9 +57,9 @@ class TestMistral(unittest.TestCase):
"learning_rate": 0.00001,
"optimizer": "adamw_torch_fused",
"lr_scheduler": "cosine",
"max_steps": 20,
"save_steps": 10,
"eval_steps": 10,
"max_steps": 5,
"save_steps": 3,
"eval_steps": 4,
"bf16": "auto",
}
)
@@ -99,9 +99,9 @@ class TestMistral(unittest.TestCase):
"learning_rate": 0.00001,
"optimizer": "adamw_torch_fused",
"lr_scheduler": "cosine",
"max_steps": 20,
"save_steps": 10,
"eval_steps": 10,
"max_steps": 5,
"save_steps": 3,
"eval_steps": 4,
"bf16": "auto",
}
)

12
tests/e2e/patched/test_mixtral_samplepack.py
View File

@@ -54,9 +54,9 @@ class TestMixtral(unittest.TestCase):
"learning_rate": 0.00001,
"optimizer": "adamw_bnb_8bit",
"lr_scheduler": "cosine",
"max_steps": 20,
"save_steps": 10,
"eval_steps": 10,
"max_steps": 5,
"save_steps": 3,
"eval_steps": 4,
"bf16": "auto",
}
)
@@ -93,9 +93,9 @@ class TestMixtral(unittest.TestCase):
"learning_rate": 0.00001,
"optimizer": "adamw_bnb_8bit",
"lr_scheduler": "cosine",
"max_steps": 20,
"save_steps": 10,
"eval_steps": 10,
"max_steps": 5,
"save_steps": 3,
"eval_steps": 4,
"bf16": "auto",
}
)

12
tests/e2e/patched/test_phi_multipack.py
View File

@@ -56,9 +56,9 @@ class TestPhiMultipack(unittest.TestCase):
"learning_rate": 0.00001,
"optimizer": "adamw_bnb_8bit",
"lr_scheduler": "cosine",
"max_steps": 20,
"eval_steps": 10,
"save_steps": 10,
"max_steps": 5,
"eval_steps": 3,
"save_steps": 4,
"bf16": "auto",
}
)
@@ -108,9 +108,9 @@ class TestPhiMultipack(unittest.TestCase):
"learning_rate": 0.00001,
"optimizer": "adamw_bnb_8bit",
"lr_scheduler": "cosine",
"max_steps": 20,
"eval_steps": 10,
"save_steps": 10,
"max_steps": 5,
"eval_steps": 3,
"save_steps": 4,
"bf16": "auto",
}
)