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: read _gpu_init.py as UTF-8 in the torchao-too-new test (#15957)
Daniel Han49582cddb0d7
diff --git a/tests/test_torchao_too_new_for_torch.py b/tests/test_torchao_too_new_for_torch.pyindex 602046424b2..f7dd8628c93 100644--- a/tests/test_torchao_too_new_for_torch.py+++ b/tests/test_torchao_too_new_for_torch.py@@ -136,7 +136,7 @@ def test_calling_twice_is_stable(isolated, capsys):def test_it_runs_after_the_torchao_shims_and_before_unsloth_zoo():- source = GPU_INIT.read_text()+ source = GPU_INIT.read_text(encoding = "utf-8")call = source.index("\ndisable_torchao_if_torch_too_old()\n")assert source.index("\nfix_torchao_torch_symbol_skew()\n") < callassert source.index("\nfix_torchao_pytree_register_constant()\n") < call