ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
unsloth/studio/backend/tests
History
Daniel Han 6412efd7d9
 Studio: auto-detect completion masking markers, stop silent full-sequence training (#7054) 
* Auto-detect completion masking markers with template table fallback

Studio's train_on_completions previously relied only on the hardcoded
MODEL_TO_TEMPLATE_MAPPER / TEMPLATE_TO_RESPONSES_MAPPER tables and
silently disabled masking when a model was not in the table, so unmapped
models (LFM2-8B-A1B, DeepSeek, and others) trained on full sequences
without telling the user. Several mapped templates (glm, mistral, llama,
starling, zephyr, qwen3-thinking) also carried markers that mask every
assistant token, which made every row drop in the post-masking filter.

Both training callsites (CUDA trainer.py and MLX worker.py) now share
utils.datasets.completion_masking.apply_completion_masking:

- Try unsloth_zoo chat template auto-detection first; it raises loudly
  when the template cannot be parsed and never masks the EOS token.
- gpt-oss models keep their manual markers so non-final assistant
  <|end|> tokens stay trained, matching current behavior.
- If auto-detection raises, fall back to the template table exactly as
  before.
- If the table also misses, emit an explicit user-visible warning that
  completion masking could not be applied and full-sequence training
  will occur, instead of a quiet log line.

The >30 percent dropped-rows safety net in trainer.py now guards the
auto path as well. Table consumers for inference and chat templates are
unchanged. Validated against one representative tokenizer for every
template in TEMPLATE_TO_RESPONSES_MAPPER plus the unmapped models:
no template regresses; unit tests cover the four decision paths.

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

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

* Restrict masking fallback to marker detection failures

The auto branch wrapped the whole train_on_responses_only call, so a real
failure while applying the masking (dataset map, tokenization) was treated
as a detection miss and training silently proceeded on full sequences.
Detect markers separately via get_chat_template_parts (test seam via
detect_fn), then apply them with errors propagating, matching the manual
path. Tokenizers with preset unsloth marker attrs skip detection and call
bare so zoo reuses the stored parts.

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

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

* Fail the run when applying completion masking raises

The helper already falls back internally on detection failures and returns
applied=False on a double miss, so an exception reaching the callsites is a
real failure applying the masking. Remove the callsite catches that
downgraded it to full-sequence training; the run now fails visibly instead.

Also use the explicit re-export alias form in utils/datasets/__init__.py for
the two new names, satisfying the import-hoist source lint.

* Import completion masking from its submodule

The import-hoist source lint counts only real name loads, so package-level
re-exports of the two new names cannot satisfy it. Import
apply_completion_masking from utils.datasets.completion_masking directly at
both callsites and leave utils/datasets/__init__.py untouched.

* Completion masking: gpt-oss renames and MLX raw/alpaca parity

Renamed or private gpt-oss checkpoints are name-detected as gpt-oss but miss
the exact-name table; default them to the gpt-oss template markers instead of
falling through to full-sequence training.

Gate the MLX masking call on not raw_text_mode and format_type != alpaca,
mirroring the CUDA path: raw/CPT text has no chat turns to mask and
Alpaca-rendered text lacks the tokenizer's chat markers.

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

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

* Define raw_text_mode outside the MLX feature-detect block

With an older zoo lacking the append_eos config field, the masking
gate referenced raw_text_mode before assignment. Hoist the assignment
above the feature detection so both consumers see it.

* Gate MLX masking on the formatter's resolved format

format_type auto can resolve to alpaca or raw text; the masking skip
checked only the requested value, so auto-detected Alpaca data got
chat-template markers applied to rendered prompt text. Track the
final_format returned by format_and_template_dataset and gate on it,
matching the CUDA path.

* Unwrap the mlx-lm TokenizerWrapper before marker checks

The wrapper delegates plain reads to the wrapped HF tokenizer but hides
underscore attrs, so preset unsloth markers were invisible and detection
relied on the loader's call patch. Unwrap to the real tokenizer first,
as the zoo MLX resolver does.

* Tighten masking comments

* gpt-oss: auto-detect markers first like every other template

The quantized and BF16 gpt-oss checkpoints ship a chat template without
the channel final header, so the pinned manual markers match nothing
there and masking trained zero tokens. Auto-detection derives markers
from whichever template the checkpoint ships and keeps the final
terminator trained; the manual gpt-oss markers remain the detection
failure fallback, including for renamed checkpoints.

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

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

* Tighten comments

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2026-07-11 05:13:45 -07:00
..
__init__.py Final cleanup 2026-03-12 18:28:04 +00:00
conftest.py Studio: clean-room compact RAG (knowledge bases, hybrid search, fast indexing) (#5910) 2026-06-09 21:17:04 -07:00
test_amd_apu_unified_memory.py Studio: graceful recovery ladder when llama-server hard-crashes at startup (#6291) 2026-06-18 09:07:25 -07:00
test_anthropic_cache_ttl.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_anthropic_citations.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_anthropic_citations_edge.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_anthropic_code_execution.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_anthropic_compaction.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_anthropic_fast_mode_and_refusal.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_anthropic_fast_mode_edge.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_anthropic_messages.py Studio /v1/messages: accept thinking and unknown content blocks (#7017) 2026-07-09 12:20:02 +02:00
test_anthropic_thinking_translation.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_anthropic_tool_versions.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_anthropic_web_fetch.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_api_key_expiry.py Studio: add --secure Cloudflare-only mode and revamp API usage examples (#6300) 2026-06-15 04:18:15 -07:00
test_api_monitor.py Studio: trim serving-log noise and surface llama-server engine stats (#6377) 2026-06-17 05:37:57 -07:00
test_api_perf_serialization.py fix(studio/llama_cpp): disable trust_env on the loopback health probe (#6750) (#6752) 2026-06-30 19:09:26 +02:00
test_apple_gpu_sensors.py Studio: show Apple GPU temperature and power in the GPU monitor (macOS) (#6187) 2026-06-12 01:50:45 -07:00
test_audio_token_detection.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_bootstrap_timeout.py Studio: auto-shut-down an exposed first-run instance if the admin password is never changed (#6651) 2026-06-26 01:27:27 -07:00
test_browse_folders_route.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_bypass_permissions.py Use UTF-8 for Python code-execution subprocess I/O (#6489 class) (#6548) 2026-06-22 09:06:03 -07:00
test_cache_case_resolution.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_cached_gguf_routes.py studio: allow updating HF models through UI (#5388) 2026-07-01 01:54:57 +03:00
test_capability_detection.py Harden model fetching (#6391) 2026-06-18 05:39:52 -07:00
test_chat_eos_template_refresh.py Studio: stop chat generation on the assistant-turn-end token (fixes Qwen3.5 loop) (#6804) 2026-07-06 10:07:56 -07:00
test_chat_history_routes.py Studio chat: tool-call nudging on by default (API stays opt-in) (#6883) 2026-07-06 19:41:19 -07:00
test_chat_history_storage.py Sort chat recents by last activity (#6844) 2026-07-07 17:54:32 +01:00
test_chat_load_during_training.py Studio: free chat model VRAM at training start only when the GPU is tight (#6243) 2026-06-18 09:04:01 -07:00
test_chat_only_reason.py Studio macOS: faster startup, MLX self-heal, drop obsolete prebuilt pins (#6494) 2026-06-22 02:20:08 -07:00
test_chat_template_tool_arguments.py Studio: coerce tool_call arguments to dict before chat templating (fixes MLX tool follow-up error) (#6807) 2026-07-06 10:12:22 -07:00
test_chat_turn_end_eos.py Studio: stop chat generation on the assistant-turn-end token (fixes Qwen3.5 loop) (#6804) 2026-07-06 10:07:56 -07:00
test_checkpoints_scan.py (feat) Add project names to studio training runs (#6512) 2026-06-29 16:06:36 +02:00
test_cleanup_cancelled_checkpoints.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_cloudflare_tunnel.py studio: explicit Cloudflare tunnel notice and public-exposure warning at startup (#6515) 2026-06-30 17:47:48 +02:00
test_coding_agents.py feat: detect installed coding agent CLIs in Studio settings (#6909) 2026-07-08 05:26:50 -07:00
test_completion_masking.py Studio: auto-detect completion masking markers, stop silent full-sequence training (#7054) 2026-07-11 05:13:45 -07:00
test_compute_buffer.py Studio: account for DeepSeek-V4 compute buffer in context auto-fit (#6940) 2026-07-07 07:20:31 -07:00
test_consent_gate.py Studio: persistent per-user trust_remote_code approval cache (#6551) 2026-06-22 05:12:49 -07:00
test_context_overflow_truncation.py Studio: report the real llama-server context window and add an opt-in overflow policy for OpenAI-compatible serving (#6164) 2026-06-11 07:49:55 -07:00
test_cpu_threads.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_data_recipe_github_progress.py Studio: add github_repo seed reader and GitHub Support Bot recipe (#5169) 2026-04-24 12:02:03 -07:00
test_data_recipe_pump_resilience.py Studio: harden background consumer loops and streaming paths against silent UI freezes (#6653) 2026-06-26 03:31:33 -07:00
test_data_recipe_seed.py Fix per-block ID collisions and add block cleanup for unstructured uploads (#6944) 2026-07-08 20:03:03 -07:00
test_datacenter_gpu_tuning.py Studio: tune llama.cpp env for data-center GPUs (#6098) 2026-06-12 02:39:01 -07:00
test_dataset_upload_limits.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_deepseek_v4_thinking_effort.py Add DeepSeek-V4-Flash-GGUF to Studio with none/high/max reasoning (#6908) 2026-07-07 06:13:43 -07:00
test_default_output_dir_name.py Studio: fix training output dir escaping outputs root for models on another drive (#6293) 2026-06-13 04:06:17 -07:00
test_desktop_auth.py Studio macOS: faster startup, MLX self-heal, drop obsolete prebuilt pins (#6494) 2026-06-22 02:20:08 -07:00
test_detect_mmproj_file.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_embedding_model_security_gate.py Run the malware gate on the RAG embedding model before it loads (#6887) 2026-07-07 04:30:21 -07:00
test_embedding_model_settings.py Studio: customizable RAG embedding model with HF search, settings tab reorganization (#6800) 2026-07-02 05:26:33 -07:00
test_exec_utf8.py Use UTF-8 for Python code-execution subprocess I/O (#6489 class) (#6548) 2026-06-22 09:06:03 -07:00
test_export_absolute_paths.py Studio: scale export GGUF size estimates from the real model size (#6418) 2026-06-18 05:44:17 -07:00
test_export_capability.py Studio: multi-select export formats, portable FP8/INT8, GGUF LoRA, and source parity (#6767) 2026-07-03 08:25:10 -07:00
test_export_imatrix_compressed.py Studio: multi-select export formats, portable FP8/INT8, GGUF LoRA, and source parity (#6767) 2026-07-03 08:25:10 -07:00
test_export_log_cursor.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_export_size_estimate.py Studio: scale export GGUF size estimates from the real model size (#6418) 2026-06-18 05:44:17 -07:00
test_external_provider_proxy_env.py Studio: ignore unsupported env proxy during Studio startup (#6102) 2026-06-11 05:13:27 -07:00
test_external_provider_usage_chunk.py Studio: Add custom provider option to Connections (#6112) 2026-06-12 13:09:35 +02:00
test_file_security.py Harden model fetching (#6391) 2026-06-18 05:39:52 -07:00
test_frontend_resolution.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_gemini_provider.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_gemma4_chat_template_override.py Studio: bundle Gemma 4 chat templates (E2B/E4B + larger) and auto-apply to unsloth/gemma-4-*-GGUF (#6245) 2026-06-12 05:49:39 -07:00
test_gemma_tool_parse_edge_cases.py Studio: parse Mistral [TOOL_CALLS] and rehearsal tool-call shapes (#5704) 2026-07-06 18:52:13 -07:00
test_gguf_completion_usage.py Studio: improve OpenAI- and Anthropic-compatible API spec compliance (#6010) 2026-06-09 17:13:25 +02:00
test_gguf_metadata.py Studio: add 'Load on selection' toggle to configure load options before loading (#6348) 2026-06-17 16:24:10 +01:00
test_gguf_reload_inheritance.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_gguf_route_cursor_reset.py Studio: extend llama.cpp first-token timeout (#5841) 2026-06-12 18:41:38 +02:00
test_gguf_routing.py Fix GGUF variant file selection (#6342) 2026-06-16 12:42:58 +02:00
test_gguf_tool_non_streaming.py Studio: honor stream=false on the GGUF agentic tool path (#6570) (#6618) 2026-06-24 15:37:08 +01:00
test_gguf_xet_fallback_integration.py Studio: Xet-primary model downloads with automatic HTTP fallback on stall (#6372) 2026-06-16 06:17:54 -07:00
test_gpu_selection.py feat: Implementation of the Portuguese (Brazil) language and VRAM/RAM monitor (#6509) 2026-07-02 18:06:39 +02:00
test_gpu_selection_sandbox.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_hf_xet_fallback.py Studio: don't pin transformers before the training worker activates the 5.x sidecar (#6968) 2026-07-08 05:33:16 -07:00
test_host_defaults.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_identity.py CLI: stop unsloth connect from leaking Studio credentials to unverified servers (#6479) 2026-06-21 21:28:38 -07:00
test_index_bootstrap_origin.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_index_bootstrap_origin_extra.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_inference_default_models_non_blocking.py Speed up Studio startup path (#6899) 2026-07-07 18:08:07 -07:00
test_inference_dispatcher_resilience.py Studio: harden OpenAI-compatible GGUF streaming (#6950) 2026-07-09 12:09:08 -03:00
test_inference_model_validation.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_inference_orchestrator_crash_message.py Studio: fix Gemma-4-12B-it not loading (#6054) 2026-06-10 08:39:07 -07:00
test_install_resolve_prebuilt.py Studio: detect Windows Intel GPUs via the registry before WMI (#7064) 2026-07-10 17:59:04 -03:00
test_kv_cache_estimation.py studio: deterministic VRAM auto-fit for GGUF (MTP reserve, compute buffer, total-based budget) (#6312) 2026-06-17 03:10:22 -07:00
test_lifespan_shutdown.py Studio: make lifespan shutdown resilient to a dead default executor (#6307) 2026-06-15 22:51:46 -07:00
test_linux_external_media_paths.py Fix Studio custom folders on Linux external drives (#6799) 2026-07-03 19:10:04 +01:00
test_llama_admission.py Studio: queue local GGUF OpenAI-compatible requests before llama-server (#7047) 2026-07-10 17:05:48 -03:00
test_llama_cpp_cache_aware_disk_check.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_llama_cpp_context_fit.py Studio: cap GGUF context to unified memory on Apple Silicon (#6622) 2026-06-24 04:39:33 -07:00
test_llama_cpp_effective_parallel_slots.py Studio: queue local GGUF OpenAI-compatible requests before llama-server (#7047) 2026-07-10 17:05:48 -03:00
test_llama_cpp_freshness.py Studio: source CPU llama.cpp prebuilts from unslothai/llama.cpp (#6311) 2026-07-08 05:34:59 -07:00
test_llama_cpp_load_progress.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_llama_cpp_load_progress_live.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_llama_cpp_load_progress_matrix.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_llama_cpp_max_context_threshold.py studio: deterministic VRAM auto-fit for GGUF (MTP reserve, compute buffer, total-based budget) (#6312) 2026-06-17 03:10:22 -07:00
test_llama_cpp_mmproj_fallback.py Studio: graceful recovery ladder when llama-server hard-crashes at startup (#6291) 2026-06-18 09:07:25 -07:00
test_llama_cpp_mtp_detection.py Studio: fix Gemma 4 separate-drafter MTP detection and fallback (#6459) 2026-06-20 04:43:37 -07:00
test_llama_cpp_no_context_shift.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_llama_cpp_props_readback.py fix(studio/llama_cpp): disable trust_env on the loopback health probe (#6750) (#6752) 2026-06-30 19:09:26 +02:00
test_llama_cpp_start_failure_classification.py report a complete load once llama-server is healthy (#6790) 2026-07-03 14:13:54 +01:00
test_llama_cpp_stream_cancel.py Studio: harden OpenAI-compatible GGUF streaming (#6950) 2026-07-09 12:09:08 -03:00
test_llama_cpp_tool_loop.py Studio: stream reasoning tokens in the tool-loop generator (fixes DeepSeek thinking not streaming with a pill on) (#6947) 2026-07-07 19:50:40 -03:00
test_llama_cpp_update.py Studio: add Vulkan llama.cpp support (#5819) 2026-07-09 03:39:48 -07:00
test_llama_cpp_vulkan_probe.py Studio: add Vulkan llama.cpp support (#5819) 2026-07-09 03:39:48 -07:00
test_llama_cpp_wait_for_health.py report a complete load once llama-server is healthy (#6790) 2026-07-03 14:13:54 +01:00
test_llama_cpp_wait_for_vram_settle.py Studio: cross-session backstop to reap a leftover llama-server on startup (#6431) 2026-06-18 05:52:18 -07:00
test_llama_cpp_windows_nvidia_path.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_llama_route.py Studio: fix llama.cpp update toast tag and reload hint (#6493) 2026-06-21 05:40:49 -07:00
test_llama_route_timeouts.py Studio: harden OpenAI-compatible GGUF streaming (#6950) 2026-07-09 12:09:08 -03:00
test_llama_server_args.py studio: deterministic VRAM auto-fit for GGUF (MTP reserve, compute buffer, total-based budget) (#6312) 2026-06-17 03:10:22 -07:00
test_llama_stats.py Studio: trim serving-log noise and surface llama-server engine stats (#6377) 2026-06-17 05:37:57 -07:00
test_llm_assist_startup_opt_in.py Studio: make Helper LLM startup pre-cache opt in (#6113) 2026-06-09 15:28:34 +02:00
test_load_progress_ready_fraction.py report a complete load once llama-server is healthy (#6790) 2026-07-03 14:13:54 +01:00
test_local_llama_cpp_link.py [Studio] Add --with-llama-cpp-dir installer flag to reuse a local llama.cpp (#6472) 2026-07-02 22:11:20 +01:00
test_local_model_format.py Studio: redesign Select model dropdown to match Hub design (#6364) 2026-06-22 04:33:04 -07:00
test_log_filter_no_truncation.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_logging_middleware.py Studio: trim serving-log noise and surface llama-server engine stats (#6377) 2026-06-17 05:37:57 -07:00
test_login_rate_limit.py Studio: harden background consumer loops and streaming paths against silent UI freezes (#6653) 2026-06-26 03:31:33 -07:00
test_mcp_config_import.py studio: show MCP "Import config" on the add-server form (#6030) 2026-06-11 16:17:22 +01:00
test_mcp_servers.py studio: tool calling for DeepSeek (R1/V3/V3.1), GLM 4.x, Kimi K2 on safetensors + MLX (#5624) 2026-07-06 15:40:46 -07:00
test_mcp_stdio_improvements.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_mcp_stdio_pr5863.py Studio: enable stdio MCP servers on a loopback bind (#6295) 2026-06-15 03:02:32 +01:00
test_message_content.py fix(studio): handle multimodal list content in inference text paths (#4383) (#6480) 2026-06-23 01:26:11 -07:00
test_middleware.py test: add regression guards for SecurityHeadersMiddleware pure-ASGI (#6424) 2026-06-17 22:53:43 -07:00
test_mlx_inference_backend.py feat(cli): support MLX distributed inference (#6845) 2026-07-08 03:25:39 -07:00
test_mlx_repair.py Studio: exclude mlx-lm 0.31.3 (broke gemma4/qwen3_5 QK-norm load on Apple Silicon) (#6803) 2026-07-06 19:40:06 -07:00
test_mlx_training_worker_config.py Add has_blackwell_gpu to the mlx worker test's wheel_utils stub (#6980) 2026-07-08 07:22:54 -07:00
test_mmproj_vram_accounting.py Studio: account for mmproj VRAM in GGUF fit budget (#5825) (#5849) 2026-06-12 15:04:08 +01:00
test_model_defaults_none_guard.py Studio macOS: faster startup, MLX self-heal, drop obsolete prebuilt pins (#6494) 2026-06-22 02:20:08 -07:00
test_model_ids.py studio: list the full local model catalog from /v1/models (#6519) 2026-06-26 20:42:06 -03:00
test_model_update_robustness.py Auto Xet to HTTP download fallback in from_pretrained; share Studio's fallback via unsloth_zoo (#6638) 2026-07-06 05:13:25 -07:00
test_models_get_model_config_case_resolution.py Harden model fetching (#6391) 2026-06-18 05:39:52 -07:00
test_mtp_drafter_companion.py Studio: resolve the repo-root MTP drafter after the MTP/ GGUF rename (#7031) 2026-07-09 06:46:00 -07:00
test_mtp_mla_target_ctx.py Studio: gate the MTP target-KV reserve to MTP spec mode, not just MLA (#6449) 2026-06-19 05:51:35 -07:00
test_mtp_vram_budget.py Studio: simplify the inference backend (#6490) 2026-06-21 20:01:09 -03:00
test_multimodal_document.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_namespace_shadow_guard_pr6269.py Studio: self-heal unsloth namespace shadows; clearer failed-load messages (#6532) 2026-06-21 22:43:31 -07:00
test_native_context_length.py Expose runtime context length for hub models (#6154) 2026-06-11 22:13:53 +03:00
test_native_template_trust_remote_code.py studio: tool calling for DeepSeek (R1/V3/V3.1), GLM 4.x, Kimi K2 on safetensors + MLX (#5624) 2026-07-06 15:40:46 -07:00
test_nudge_tool_calls_wiring.py Studio: keep the nudge wiring test collectable without the unsloth stack (#6924) 2026-07-06 21:57:56 -07:00
test_offline_gguf_cache_fallback.py Fix case-variant model matching and GGUF cache reuse in unsloth start (#6900) 2026-07-08 02:32:06 -07:00
test_offline_inference_parent.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_openai_auto_switch.py Studio: hint at Model auto-switch in the OpenAI "No model loaded" 400 (#7006) 2026-07-10 17:48:27 -03:00
test_openai_catalog.py Studio: opt-in OpenAI /v1 model auto-switch and idle keep-warm (#6392) 2026-07-01 06:42:23 -07:00
test_openai_citation_markers.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_openai_citation_markers_edge.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_openai_code_execution.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_openai_compaction.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_openai_container_crud.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_openai_image_generation.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_openai_models_path_leak.py studio: return a clean model id from the OpenAI API instead of the local .gguf path (#6518) 2026-06-26 16:07:53 -03:00
test_openai_responses_translation.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_openai_tool_passthrough.py Studio: queue local GGUF OpenAI-compatible requests before llama-server (#7047) 2026-07-10 17:05:48 -03:00
test_openai_tool_result_fallbacks.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_orchestrator_unload_cancel.py Studio: serialize the compare-mode dispatcher lifecycle to fix a start race (#6922) 2026-07-06 22:09:41 -07:00
test_passthrough_healing.py Studio: harden OpenAI-compatible GGUF streaming (#6950) 2026-07-09 12:09:08 -03:00
test_personalization_settings.py studio: persist personalization (profile, avatar, theme) server-side (#6516) 2026-06-22 04:09:48 -07:00
test_pr5624_regressions.py studio: tool calling for DeepSeek (R1/V3/V3.1), GLM 4.x, Kimi K2 on safetensors + MLX (#5624) 2026-07-06 15:40:46 -07:00
test_presence_penalty.py Studio: apply presence_penalty on the safetensors and MLX inference paths (#6923) 2026-07-06 22:24:47 -07:00
test_preview.py Studio: shareable per-checkpoint preview links (#6486) 2026-06-24 06:31:53 -07:00
test_preview_followups.py Studio: require signed capability tokens for /p preview links (#6666) 2026-06-25 21:40:48 -07:00
test_preview_routes.py Studio: require signed capability tokens for /p preview links (#6666) 2026-06-25 21:40:48 -07:00
test_preview_sharing_settings.py Studio: require signed capability tokens for /p preview links (#6666) 2026-06-25 21:40:48 -07:00
test_preview_token.py Studio: require signed capability tokens for /p preview links (#6666) 2026-06-25 21:40:48 -07:00
test_pricing.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_pricing_edge.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_process_lifetime.py Reap Studio child processes when the parent dies abnormally (#6425) 2026-06-18 05:51:22 -07:00
test_providers_api.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_pytorch_mirror.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_rag_captioning.py Studio RAG: disable trust_env on loopback llama-server httpx clients (#6775) 2026-07-01 03:58:08 -07:00
test_rag_chunking.py Studio: clean-room compact RAG (knowledge bases, hybrid search, fast indexing) (#5910) 2026-06-09 21:17:04 -07:00
test_rag_embed_llama_server.py Studio: customizable RAG embedding model with HF search, settings tab reorganization (#6800) 2026-07-02 05:26:33 -07:00
test_rag_embeddings.py Studio: clean-room compact RAG (knowledge bases, hybrid search, fast indexing) (#5910) 2026-06-09 21:17:04 -07:00
test_rag_ingestion.py Whole-document context for RAG chat attachments (#6693) 2026-06-30 15:55:23 +02:00
test_rag_job_events_queue_lifecycle.py Studio: harden background consumer loops and streaming paths against silent UI freezes (#6653) 2026-06-26 03:31:33 -07:00
test_rag_loopback_trust_env.py Studio RAG: disable trust_env on loopback llama-server httpx clients (#6775) 2026-07-01 03:58:08 -07:00
test_rag_ocr_fallback.py Whole-document context for RAG chat attachments (#6693) 2026-06-30 15:55:23 +02:00
test_rag_parsing.py Studio RAG: fix RTL/Indic PDF corruption and dropped DOCX tables (#6780) 2026-07-02 11:38:48 +01:00
test_rag_preview.py Whole-document context for RAG chat attachments (#6693) 2026-06-30 15:55:23 +02:00
test_rag_reconcile_orphaned.py Studio: harden background consumer loops and streaming paths against silent UI freezes (#6653) 2026-06-26 03:31:33 -07:00
test_rag_retrieval.py Studio: clean-room compact RAG (knowledge bases, hybrid search, fast indexing) (#5910) 2026-06-09 21:17:04 -07:00
test_rag_store.py Studio: clean-room compact RAG (knowledge bases, hybrid search, fast indexing) (#5910) 2026-06-09 21:17:04 -07:00
test_rag_whole_document.py Whole-document context for RAG chat attachments (#6693) 2026-06-30 15:55:23 +02:00
test_recommended_folders_has_model.py Studio: redesign Select model dropdown to match Hub design (#6364) 2026-06-22 04:33:04 -07:00
test_recommended_folders_permission.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_resolve_quant_gguf.py Studio: redesign Select model dropdown to match Hub design (#6364) 2026-06-22 04:33:04 -07:00
test_responses_api.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_responses_tool_passthrough.py Studio: parse Mistral [TOOL_CALLS] and rehearsal tool-call shapes (#5704) 2026-07-06 18:52:13 -07:00
test_rocm_oom_guard.py Windows/WSL installer: fix winget msstore cert failure, amd-smi DiskPart prompt, and enable AMD GPU (Strix Halo gfx1151) (#5940) 2026-06-10 04:24:49 -07:00
test_s3_dataset.py feat(studio): implement S3 dataset loading (completes #5951) (#6222) 2026-06-12 14:52:04 +02:00
test_safetensors_capability_advertise.py Add DeepSeek-V4-Flash-GGUF to Studio with none/high/max reasoning (#6908) 2026-07-07 06:13:43 -07:00
test_safetensors_reasoning_stream.py Studio: parse Mistral [TOOL_CALLS] and rehearsal tool-call shapes (#5704) 2026-07-06 18:52:13 -07:00
test_safetensors_tool_loop.py Studio chat: tool-call nudging on by default (API stays opt-in) (#6883) 2026-07-06 19:41:19 -07:00
test_safetensors_toolcall_wiring.py studio: deterministic backend tool-calling wiring test (#6836) 2026-07-06 10:08:39 -07:00
test_sandbox_tools.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_secure_tools_execute.py Keep server-side tools enabled under --secure (#6403) 2026-06-18 05:52:40 -07:00
test_secure_tunnel_gate.py studio: explicit Cloudflare tunnel notice and public-exposure warning at startup (#6515) 2026-06-30 17:47:48 +02:00
test_security_gate_consistency.py Run the malware gate on the RAG embedding model before it loads (#6887) 2026-07-07 04:30:21 -07:00
test_server_disk_logging.py Windows/WSL installer: fix winget msstore cert failure, amd-smi DiskPart prompt, and enable AMD GPU (Strix Halo gfx1151) (#5940) 2026-06-10 04:24:49 -07:00
test_setup_cache_env_hf_home.py Studio: honor custom HF_HOME for model download and load (#6510) 2026-06-22 08:21:59 -07:00
test_sf_client_tools_passthrough.py Studio: client-tool passthrough healing for safetensors and MLX (#6870) 2026-07-06 19:48:36 -07:00
test_slot_offload_fit.py Studio: reserve CUDA context and mmproj/MTP soft overhead in the GGUF fit budget (#6718) 2026-07-03 13:07:30 -03:00
test_sse_streaming_headers.py Studio: simplify the inference backend (#6490) 2026-06-21 20:01:09 -03:00
test_ssm_runtime.py Fix test isolation: restore sys.modules after the pre-import gate test (#6578) 2026-06-22 08:45:34 -07:00
test_startup_banner_loopback.py Fix Windows Studio UTF-8 startup handling (#6614) 2026-07-01 13:47:33 +01:00
test_startup_llama_probe_non_blocking.py Studio macOS: faster startup, MLX self-heal, drop obsolete prebuilt pins (#6494) 2026-06-22 02:20:08 -07:00
test_studio_api.py Studio: Add Tensor-Parallel llama.cpp support (#6040) 2026-06-12 04:00:52 -07:00
test_studio_train_validation.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_tensor_parallel.py Studio: reserve CUDA context and mmproj/MTP soft overhead in the GGUF fit budget (#6718) 2026-07-03 13:07:30 -03:00
test_think_prefill_reemit.py Studio: render thinking blocks for safetensors inference with prefilled <think> templates (#6816) 2026-07-08 08:14:03 -07:00
test_thinking_parameter.py feat: add Anthropic-compatible thinking parameter (#5856) 2026-06-15 10:35:33 +01:00
test_tool_approvals.py Studio: Add inline confirmation (Allow/Always allow/Deny) for tool calls (#5869) 2026-06-12 10:55:26 +02:00
test_tool_call_parser_strict.py Studio: parse Mistral [TOOL_CALLS] and rehearsal tool-call shapes (#5704) 2026-07-06 18:52:13 -07:00
test_tool_confirm_loop.py Studio: Bypass Permissions (skip confirmation, disable tool sandbox) (#5895) 2026-06-15 04:04:22 -07:00
test_tool_confirm_stream.py Studio: Add inline confirmation (Allow/Always allow/Deny) for tool calls (#5869) 2026-06-12 10:55:26 +02:00
test_tool_loop_controller.py Rename chat artifacts copy to canvas (#6298) 2026-06-15 13:38:39 +02:00
test_tool_message_empty_content.py Studio: report the real llama-server context window and add an opt-in overflow policy for OpenAI-compatible serving (#6164) 2026-06-11 07:49:55 -07:00
test_tool_policy_gates.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_tool_policy_state.py unsloth run: add --enable-tools/--disable-tools server-side tool policy (#5277) 2026-05-05 12:45:15 +04:00
test_tool_strip_guard.py Studio: Gemma tool-call streaming follow-ups + nested-XML escape fix (#6476) (#6611) 2026-07-06 10:39:37 -07:00
test_tool_xml_strip.py Studio: parse Mistral [TOOL_CALLS] and rehearsal tool-call shapes (#5704) 2026-07-06 18:52:13 -07:00
test_torchao_select.py Studio: fix flash-attn and torchao install on Blackwell (sm_100+) GPUs (Closes #6961) (#6970) 2026-07-08 06:38:10 -07:00
test_tp_vision_regression.py Studio: restore tensor parallelism for vision/mmproj GGUFs (#6659) 2026-06-27 01:52:18 -07:00
test_trained_model_scan.py Harden model fetching (#6391) 2026-06-18 05:39:52 -07:00
test_training_before_spawn.py Studio: free chat model VRAM at training start only when the GPU is tight (#6243) 2026-06-18 09:04:01 -07:00
test_training_history_update.py Studio: require signed capability tokens for /p preview links (#6666) 2026-06-25 21:40:48 -07:00
test_training_nan_loss_handling.py fix(studio): surface live step with null loss through the SSE progress stream (#6206) 2026-06-11 07:50:13 -07:00
test_training_preflight.py Add MLX backend support for CLI unsloth train (#6709) 2026-07-08 03:25:26 -07:00
test_training_progress_prep_timeout.py Studio: harden background consumer loops and streaming paths against silent UI freezes (#6653) 2026-06-26 03:31:33 -07:00
test_training_progress_stream_nan.py Studio: harden background consumer loops and streaming paths against silent UI freezes (#6653) 2026-06-26 03:31:33 -07:00
test_training_pump_resilience.py Studio: keep the training event pump alive so progress can't silently freeze (#6643) 2026-06-25 05:19:32 -07:00
test_training_raw_support.py MLX Training updates (#5656) 2026-06-14 04:58:50 -07:00
test_training_resume.py feat(studio): implement S3 dataset loading (completes #5951) (#6222) 2026-06-12 14:52:04 +02:00
test_training_runs.py (feat) Add project names to studio training runs (#6512) 2026-06-29 16:06:36 +02:00
test_training_streaming.py (feat) Add project names to studio training runs (#6512) 2026-06-29 16:06:36 +02:00
test_training_vram_coexistence.py Studio: free chat model VRAM at training start only when the GPU is tight (#6243) 2026-06-18 09:04:01 -07:00
test_training_worker_flash_attn.py Studio: fix flash-attn and torchao install on Blackwell (sm_100+) GPUs (Closes #6961) (#6970) 2026-07-08 06:38:10 -07:00
test_training_worker_import_discipline.py Studio: don't pin transformers before the training worker activates the 5.x sidecar (#6968) 2026-07-08 05:33:16 -07:00
test_training_xet_fallback.py Studio: Xet-primary model downloads with automatic HTTP fallback on stall (#6372) 2026-06-16 06:17:54 -07:00
test_transformers_version.py Fix offline checkpoint load/export: "tokenizer is weirdly not loaded" (#6554) 2026-06-25 23:16:53 -07:00
test_trc_approval_cache.py Studio: persistent per-user trust_remote_code approval cache (#6551) 2026-06-22 05:12:49 -07:00
test_utils.py Studio macOS: faster startup, MLX self-heal, drop obsolete prebuilt pins (#6494) 2026-06-22 02:20:08 -07:00
test_validate_gguf_runtime_message.py Studio: show an actionable message when the GGUF runtime is missing (#6327) 2026-06-18 06:00:00 -07:00
test_validate_model_error.py Harden model fetching (#6391) 2026-06-18 05:39:52 -07:00
test_vision_cache.py Fix offline checkpoint load/export: "tokenizer is weirdly not loaded" (#6554) 2026-06-25 23:16:53 -07:00
test_vram_estimation.py Studio: make code comments and docstrings more succinct (#6029) 2026-06-08 23:07:28 -07:00
test_windows_gpu_detection_mock.py studio: deterministic VRAM auto-fit for GGUF (MTP reserve, compute buffer, total-based budget) (#6312) 2026-06-17 03:10:22 -07:00
test_worker_activates_correct_transformers.py Studio: don't pin transformers before the training worker activates the 5.x sidecar (#6968) 2026-07-08 05:33:16 -07:00
test_yaml_trust_remote_code_removed.py Harden trust_remote_code consent: scan GGUF-only auto_map and drop pre-set TRC defaults (#6478) 2026-06-22 02:10:35 -07:00