Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4.1, MiniMax-H3, Gemma 4, FLUX and more.
Python agentaichatgptdeepseekfine-tuninggemmaimage-generationllamallmllmsopenaipythonqwenreinforcement-learningself-hostedstable-diffusiontext-to-speechttsuiunsloth
tests: stub the tensor parallel guard in the for_training extraction harness (#15498)
Daniel Han945bf3fbcd23
diff --git a/tests/test_for_training_generation_flag.py b/tests/test_for_training_generation_flag.pyindex db64bac659b..9481a5bbb76 100644--- a/tests/test_for_training_generation_flag.py+++ b/tests/test_for_training_generation_flag.py@@ -92,6 +92,8 @@ def _for_training(module, class_name):node.decorator_list = []compiled = ast.Module(body = shared + _helpers_used_by(method, tree) + [method], type_ignores = [])namespace = _Namespace(os = os, inspect = inspect, torch = _torch_or_stub(), functools = functools)+ # No tensor parallel world in these tests (for_training calls the TP guard first since #15368).+ namespace["_tensor_parallel_for_training"] = lambda model: Noneexec(compile(ast.fix_missing_locations(compiled), str(path), "exec"), namespace)return namespace["for_training"]