vrr/unsloth
2
0
Fork 0
mirror of https://github.com/unslothai/unsloth.git synced 2026-08-16 04:13:54 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 34
unsloth/tests
History
Exact
Daniel Han 8059be6bca
Bump install.sh / install.ps1 pin to unsloth>=2026.8.18 (#8836) 
* Bump install.sh / install.ps1 pin to unsloth>=2026.8.18

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2026-08-14 08:06:18 -07:00
..
fast_inference test: fast end-to-end GRPO fast_inference vLLM rollout test (#7136) 2026-07-26 00:22:48 -07:00
kaggle Bump install.sh / install.ps1 pin to unsloth>=2026.8.18 (#8836) 2026-08-14 08:06:18 -07:00
notebooks Make the Colab oracle tripwire able to fire, and stop blaming start.py for a hung agent CLI (#7838) 2026-08-04 07:42:42 -07:00
python Windows: stop compiling C# for colour on hosts that already render it (#8767) 2026-08-14 04:44:37 -07:00
qlora Formatting: ruff line-length 100, kwarg-spacing passes, drop blank after short local imports (#6079) 2026-06-08 04:24:13 -07:00
saving Use Kaggle's large overlay for saves, and refuse a GGUF export that cannot fit (#8439) 2026-08-13 18:56:43 -07:00
security Revert "Desktop: ship a complete Linux AppImage (#8695)" (#8823) 2026-08-14 06:55:54 -07:00
sh Let the uv signal handler exit without going through setup_fail (#8744) 2026-08-13 17:16:18 -07:00
studio Bump install.sh / install.ps1 pin to unsloth>=2026.8.18 (#8836) 2026-08-14 08:06:18 -07:00
studio_setup_ps1 Studio: route every Windows installer line through the UTF-8 stdout sink (#8148) 2026-08-08 06:47:03 -07:00
utils Carry the sliding window into SDPA, and treat a zero window as no window (#8253) 2026-08-09 05:33:38 -07:00
version_compat Hand the LoRA to vLLM on TRL 1.x GRPO rollouts (#8701) 2026-08-14 06:46:48 -07:00
vllm_compat Reduce and tighten comments and docstrings across the test suite (#6429) 2026-06-18 01:07:09 -07:00
__init__.py Qwen 3, Bug Fixes (#2445) 2025-04-30 22:38:39 -07:00
_grpo_dispatch_source.py GRPO: dispatch on width at the remaining lm_head matmul call sites (#8204) 2026-08-11 10:44:06 -07:00
_rl_source.py Stop the GRPO hidden-states wrapper paying for logits it discards (#8576) 2026-08-13 05:15:07 -07:00
_zoo_aggressive_cuda_spoof.py Stop the CUDA test spoofs reporting an exhausted card (#8123) 2026-08-08 04:45:25 -07:00
_zoo_rocm_spoof.py AMD: CI coverage for recent fixes, plus three wrong gfx ids (#7431) 2026-07-25 18:58:02 -05:00
conftest.py Stop the CUDA test spoofs reporting an exhausted card (#8123) 2026-08-08 04:45:25 -07:00
run_all.sh Anchor the host-defaults assertions and take the file off the skip list (#7774) 2026-08-03 02:42:26 -07:00
test_allow_cpu_import_driverless.py Ask whether a device is present before asking what it can do (#8653) 2026-08-13 04:40:07 -07:00
test_attention_implementation.py fix(gpt-oss): prefer flex attention over sdpa (#5701) 2026-05-22 08:38:38 -07:00
test_attn_impl_honor_explicit.py Honor an explicit sdpa or flex_attention request when flash is disabled (#6847) 2026-07-06 05:45:45 -07:00
test_bad_mappings_redirect.py Keep the newer-mapper probe from replacing the installed FP8 mappers (#7478) 2026-07-27 04:21:27 -07:00
test_broken_tf_does_not_break_import.py Do not let a broken TensorFlow install break the Unsloth import (#8124) 2026-08-09 05:15:40 -07:00
test_broken_torchvision_probe.py Name a torchvision whose compiled ops do not match torch (#8128) 2026-08-09 04:30:12 -07:00
test_callback_signature_drift.py Reduce and tighten comments and docstrings across the test suite (#6429) 2026-06-18 01:07:09 -07:00
test_cli_export_unpacking.py Reduce and tighten comments and docstrings across the test suite (#6429) 2026-06-18 01:07:09 -07:00
test_collection_hygiene.py tests: make pytest tests collect cleanly and drop the deleted QVQ registry mirror (#8206) 2026-08-08 20:03:55 -07:00
test_compressed_export_gpu_release.py studio: shard export checkpoint loads across all visible GPUs (#7215) 2026-07-26 04:16:36 -07:00
test_cuda_spoof_reports_free_memory.py Stop the CUDA test spoofs reporting an exhausted card (#8123) 2026-08-08 04:45:25 -07:00
test_dataclass_default_backfill.py Keep up with transformers 5.x, TRL 0.22.x and peft (#7866) 2026-08-07 06:33:05 -07:00
test_deliberate_crashes_suppress_cores.py Studio: close seven detection gaps in the deliberate-crash guard (#8788) 2026-08-14 04:12:45 -07:00
test_device_helpers.py Report the real cause when a save or GGUF export fails (#7861) 2026-08-06 00:12:32 -07:00
test_enforce_kwargs_spacing.py Reduce and tighten comments and docstrings across the test suite (#6429) 2026-06-18 01:07:09 -07:00
test_fa2_fast_generate_bypass.py Bypass fast_generate for flash_attention_2 models (StaticCache + FA2 produces gibberish) (#7429) 2026-07-26 23:07:33 -07:00
test_fast_gemv_dispatch.py Fix fast inference crash on compressed-tensors FP8 models (#7025) 2026-07-09 04:10:59 -07:00
test_fast_generate_slow_guard.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_finetune_last_n_layers.py Reduce and tighten comments and docstrings across the test suite (#6429) 2026-06-18 01:07:09 -07:00
test_flex_attention_needs_ampere.py Do not choose flex attention on a card that cannot run its kernel (#8137) 2026-08-08 05:31:23 -07:00
test_float32_generate_autocast.py Do not enter CUDA autocast for a float32 model during generate (#7900) 2026-08-05 04:35:21 -07:00
test_float32_no_fp16_autocast.py Keep an explicitly requested float32 model in float32 without bfloat16 (#7867) 2026-08-09 05:12:02 -07:00
test_fp8_device_context.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_fp8_restore_dropped_scale.py Restore dropped FP8 weight_scale_inv tensors on load (#6978) 2026-07-09 06:44:44 -07:00
test_fp8_tiny_e8m0.py fix: enable XPU support and update hardcoded CUDA selections for tests (#7401) 2026-07-28 15:38:57 -07:00
test_fused_ce_not_return_dict_logits.py fix: use EMPTY_LOGITS on the fused-CE not-return_dict path (#2068) (#6482) 2026-06-23 01:26:55 -07:00
test_gemma4_chat_template.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_gemma_2b_mapper_key.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_generate_kwarg_gate.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_generation_failure_is_visible.py Restore warnings_issued for trl trainers on transformers 5.1+ (#7854) 2026-08-05 04:34:23 -07:00
test_get_model_name.py Keep the newer-mapper probe from replacing the installed FP8 mappers (#7478) 2026-07-27 04:21:27 -07:00
test_gguf_basename_platform_matrix.py Fix GGUF export writing outside the chosen save directory on Windows (#7897) (#7937) 2026-08-05 08:24:48 -07:00
test_gguf_disk_headroom.py Free the intermediate 16-bit merge when the GGUF quants will not fit (#8500) 2026-08-13 04:42:31 -07:00
test_gguf_model_basename.py Fix GGUF export writing outside the chosen save directory on Windows (#7897) (#7937) 2026-08-05 08:24:48 -07:00
test_gguf_windows_export_routing.py Fix GGUF export writing outside the chosen save directory on Windows (#7897) (#7937) 2026-08-05 08:24:48 -07:00
test_gguf_windows_native.py Fix GGUF export writing outside the chosen save directory on Windows (#7897) (#7937) 2026-08-05 08:24:48 -07:00
test_gradient_checkpointing_restore.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_grouped_gemm_optional_gather_indices.py fix grouped_gemm crash on the documented gather_indices=None default (#8629) 2026-08-14 05:27:25 -07:00
test_grpo_accumulated_loss_hidden_states_signal.py GRPO: dispatch on width at the remaining lm_head matmul call sites (#8204) 2026-08-11 10:44:06 -07:00
test_grpo_autocast_disabled.py Keep an explicitly requested float32 model in float32 without bfloat16 (#7867) 2026-08-09 05:12:02 -07:00
test_grpo_autocast_per_trainer.py Keep an explicitly requested float32 model in float32 without bfloat16 (#7867) 2026-08-09 05:12:02 -07:00
test_grpo_hidden_states_logits_cost.py Stop the GRPO hidden-states wrapper paying for logits it discards (#8576) 2026-08-13 05:15:07 -07:00
test_grpo_hidden_states_per_call_degradation.py GRPO: dispatch on width at the remaining lm_head matmul call sites (#8204) 2026-08-11 10:44:06 -07:00
test_grpo_hidden_states_signal.py GRPO: dispatch on width at the remaining lm_head matmul call sites (#8204) 2026-08-11 10:44:06 -07:00
test_grpo_hidden_states_wrap_target.py GRPO: wrap the model that owns the head when falling back to hidden states (#8231) 2026-08-09 04:48:46 -07:00
test_grpo_packed_raw_logits_nograd.py GRPO: dispatch on width at the remaining lm_head matmul call sites (#8204) 2026-08-11 10:44:06 -07:00
test_grpo_padded_raw_logits.py GRPO: dispatch on width at the remaining lm_head matmul call sites (#8204) 2026-08-11 10:44:06 -07:00
test_grpo_width_dispatch_sites.py GRPO: dispatch on width at the remaining lm_head matmul call sites (#8204) 2026-08-11 10:44:06 -07:00
test_ignored_tokenizer_casing.py Match IGNORED_TOKENIZER_NAMES case-insensitively (#6620) 2026-06-23 19:40:50 -03:00
test_import_fixes_drift.py Fix Apple Silicon M4+ CPU frequency reported as MHz instead of GHz (#8571) 2026-08-12 17:02:52 -03:00
test_installer_interactive_prompts.py Make install.ps1 work with the user's PowerShell profile loaded (#8161) 2026-08-09 04:27:28 -07:00
test_installer_profile_hardening.py Windows: stop depending on the generated unsloth.exe console script (#8592) 2026-08-13 07:54:51 -07:00
test_installer_shortcut_icons.py feat(install): use pre-built Tauri icon for macOS .app bundle (#7628) 2026-07-31 12:05:34 +02:00
test_installer_skip_autostart.py fix(install): fail non-tauri installer errors (#7123) 2026-07-15 04:37:03 -07:00
test_installer_system32_guard.py Windows: stop the oversize guard tests from emptying os.environ (#8696) 2026-08-13 07:00:50 -07:00
test_installer_unsloth_version.py Installer: report the installed Unsloth version (#7265) 2026-07-21 03:48:49 -07:00
test_kaggle_gguf_error_message.py Report the real cause when a save or GGUF export fails (#7861) 2026-08-06 00:12:32 -07:00
test_lint_no_parallel_clamp.py Studio: stop MTP forcing llama-server to a single parallel slot (#8172) 2026-08-08 21:09:58 -07:00
test_loader_glob_skip.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_map_eos_token.py Honor an explicit map_eos_token=False in get_chat_template (#7475) 2026-07-30 10:09:38 -03:00
test_mapper_no_duplicate_keys.py Add test to guard against duplicate keys in __INT_TO_FLOAT_MAPPER (#7419) 2026-07-29 02:37:10 -07:00
test_missing_optional_dep_skips.py Keep an explicitly requested float32 model in float32 without bfloat16 (#7867) 2026-08-09 05:12:02 -07:00
test_missing_torchvision_vlm.py Fix misleading 'only for image models' error for Qwen3-VL when torchvision is missing (#6525) 2026-06-23 01:28:09 -07:00
test_model_registry.py tests: build the hub-error fixtures without calling __init__ (huggingface_hub 1.x) (#8207) 2026-08-08 21:09:00 -07:00
test_moe_lora_targets.py Scope MoE expert LoRA detection to actual MLP projection targets (#6849) 2026-07-06 05:45:06 -07:00
test_multi_image_grpo_chunking.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_new_mapper_fetched_fp8.py Cover the FP8 row-scaling path in the newer-mapper probe (#7516) 2026-07-28 16:13:00 -07:00
test_new_mapper_no_global_leak.py Keep the newer-mapper probe from replacing the installed FP8 mappers (#7478) 2026-07-27 04:21:27 -07:00
test_nvfp4_quant_load.py Reduce and tighten comments and docstrings across the test suite (#6429) 2026-06-18 01:07:09 -07:00
test_offline_loading_helpers.py Report the network error when the offline retry finds nothing cached (#8121) 2026-08-09 05:01:02 -07:00
test_offload_embedding_hooks.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_offload_tied_autodisable.py Turn off embedding offload when the model is dispatched across devices (#8272) 2026-08-09 09:26:01 -07:00
test_offload_tied_guard.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_offloaded_parameter_hint.py Report the real cause when a save or GGUF export fails (#7861) 2026-08-06 00:12:32 -07:00
test_peft_stale_torchao.py Settle deferred compile-mode switches between training steps (#7955) 2026-08-06 02:45:50 -07:00
test_peft_symbol_backfill.py Keep up with transformers 5.x, TRL 0.22.x and peft (#7866) 2026-08-07 06:33:05 -07:00
test_peft_tensor_parallel_compat.py fix: keep LoRA reloads working with PEFT 0.19 (#6748) 2026-06-30 20:26:57 +01:00
test_peft_weight_converter_compat.py Formatting: ruff line-length 100, kwarg-spacing passes, drop blank after short local imports (#6079) 2026-06-08 04:24:13 -07:00
test_prefetch_snapshot_scope.py Fetch a sentence-transformers model's subfolder weights instead of pruning them (#7855) 2026-08-05 04:34:48 -07:00
test_pretrain_compile_reset.py Add regression tests for the stray-forward compile-cache reset (#6569) 2026-06-22 07:22:47 -07:00
test_profile_startup_gate.py Measure where Studio's startup time actually goes (#7553) 2026-07-28 22:24:34 -07:00
test_psutil_apple_cpu_freq.py Fix Apple Silicon M4+ CPU frequency reported as MHz instead of GHz (#8571) 2026-08-12 17:02:52 -03:00
test_public_api_surface.py Reduce and tighten comments and docstrings across the test suite (#6429) 2026-06-18 01:07:09 -07:00
test_python39_compatibility.py Do not scan vendored packages that refuse to import on our floor (#8252) 2026-08-09 04:06:19 -07:00
test_pythonpath_empty_components.py Settle deferred compile-mode switches between training steps (#7955) 2026-08-06 02:45:50 -07:00
test_raw_text.py tests: make pytest tests collect cleanly and drop the deleted QVQ registry mirror (#8206) 2026-08-08 20:03:55 -07:00
test_raw_text_json_loading.py Parse a .json dataset file as one JSON document instead of line-by-line (#7422) 2026-07-28 21:40:43 -03:00
test_resolve_model_class.py Formatting: ruff line-length 100, kwarg-spacing passes, drop blank after short local imports (#6079) 2026-06-08 04:24:13 -07:00
test_rl_config_compat.py Read rl.py with an explicit utf-8 encoding in the config compat test (#8333) 2026-08-10 05:47:09 -07:00
test_runtime_text_encoding.py Pin utf-8 on shipping-code text I/O instead of the operator locale (#7486) 2026-07-27 02:14:20 -07:00
test_save_entrypoints_reach_converter.py Report the real cause when a save or GGUF export fails (#7861) 2026-08-06 00:12:32 -07:00
test_save_lora_without_vllm.py Attach save_lora without a vLLM engine (#8127) 2026-08-08 06:14:12 -07:00
test_settle_eager_fallbacks_between_steps.py Settle deferred compile-mode switches between training steps (#7955) 2026-08-06 02:45:50 -07:00
test_sft_vision_dataset_gate.py Keep up with transformers 5.x, TRL 0.22.x and peft (#7866) 2026-08-07 06:33:05 -07:00
test_source_read_encoding.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_st_save_merged_signature.py Report the real cause when a save or GGUF export fails (#7861) 2026-08-06 00:12:32 -07:00
test_st_subfolder_weights_are_fetched.py Fetch a sentence-transformers model's subfolder weights instead of pruning them (#7855) 2026-08-05 04:34:48 -07:00
test_studio_install_workspace_guard.py Installer: suppress macOS uv developer tools dialog (#8479) 2026-08-11 12:19:30 -07:00
test_studio_root_resilience.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_studio_shutdown_thread_wait.py Replace standalone Studio wording with Unsloth (#7221) 2026-07-19 00:47:04 -07:00
test_synthetic_chunk_data.py Fix SyntheticDataKit.chunk_data emitting chunks over max_tokens (#7073) 2026-07-12 05:06:11 -07:00
test_synthetic_vllm_startup_failure.py Raise when SyntheticDataKit's vLLM server fails to start (#8297) 2026-08-10 00:55:00 -07:00
test_tool_mask_zoo_compat.py tests: read checked-in files as UTF-8 instead of the platform default (#7438) 2026-07-26 23:31:56 -07:00
test_torchao_aten_grouped_mm.py Settle deferred compile-mode switches between training steps (#7955) 2026-08-06 02:45:50 -07:00
test_torchao_nf4tensor_move.py Settle deferred compile-mode switches between training steps (#7955) 2026-08-06 02:45:50 -07:00
test_torchao_subprocess_fix.py Settle deferred compile-mode switches between training steps (#7955) 2026-08-06 02:45:50 -07:00
test_torchao_torch_symbol_skew.py Settle deferred compile-mode switches between training steps (#7955) 2026-08-06 02:45:50 -07:00
test_torchaudio_cuda_mismatch.py Do not let a CUDA-mismatched torchaudio take the whole import with it (#8496) 2026-08-13 05:12:36 -07:00
test_transformers5_bare_annotation_live.py Keep up with transformers 5.x, TRL 0.22.x and peft (#7866) 2026-08-07 06:33:05 -07:00
test_transformers_dependency_floor.py Tell users to upgrade the dependency, not transformers, when a --no-deps install leaves a floor unmet (#8212) 2026-08-09 01:45:07 -07:00
test_uma_safetensors_load.py Faster safetensors weight loading on unified-memory (integrated) GPUs (#5988) 2026-07-22 05:20:59 -07:00
test_uninitialized_position_ids.py Load DeepSeek-OCR and other VLMs that register AutoModel in auto_map (#6421) 2026-06-18 07:03:43 -07:00
test_version_single_source.py Report unsloth's own version on the MLX path (#8395) 2026-08-11 01:10:03 -07:00
test_video_path_validation.py Fix formatting_func race in UnslothVisionDataCollator without a lock (#7698) 2026-08-02 22:44:09 -07:00
test_vllm_broken_detection.py Fix fast_inference crash on ABI-broken vLLM: probe compiled extensions, not just import vllm (#6621) 2026-06-26 22:43:36 -07:00
test_vllm_cuda_mismatch_wheel_url.py Stop the vLLM CUDA mismatch hint printing a wheel URL that 404s (#8200) 2026-08-10 06:42:39 -07:00
test_warnings_issued_guard.py Apply the moved trainer kwargs to the config instead of dropping them (#8374) 2026-08-11 04:35:10 -07:00
test_windows_amd_gpu_scan_fallback.py Windows: stop depending on the generated unsloth.exe console script (#8592) 2026-08-13 07:54:51 -07:00
test_windows_rocm_bnb_version.py Reduce and tighten comments and docstrings across the test suite (#6429) 2026-06-18 01:07:09 -07:00