Commit Graph

4664 Commits

Author SHA1 Message Date
comfyanonymous
e78d230496
Only enable cuda malloc on cuda torch. (#9031) 2025-07-23 19:37:43 -04:00
comfyanonymous
d3504e1778
Enable pytorch attention by default for gfx1201 on torch 2.8 (#9029) 2025-07-23 19:21:29 -04:00
comfyanonymous
a86a58c308
Fix xpu function not implemented p2. (#9027) 2025-07-23 18:18:20 -04:00
comfyanonymous
39dda1d40d
Fix xpu function not implemented. (#9026) 2025-07-23 18:10:59 -04:00
comfyanonymous
5ad33787de
Add default device argument. (#9023) 2025-07-23 14:20:49 -04:00
Simon Lui
255f139863
Add xpu version for async offload and some other things. (#9004) 2025-07-22 15:20:09 -04:00
comfyanonymous
5ac9ec214b
Try to fix line endings workflow. (#9001) 2025-07-22 04:07:51 -04:00
comfyanonymous
0aa1c58b04
This is not needed. (#8991) 2025-07-21 16:48:25 -04:00
comfyanonymous
5249e45a1c
Add hidream e1.1 example to readme. (#8990) 2025-07-21 15:23:41 -04:00
comfyanonymous
54a45b9967
Replace torchaudio.load with pyav. (#8989) 2025-07-21 14:19:14 -04:00
comfyanonymous
9a470e073e ComfyUI version 0.3.45 2025-07-21 14:05:43 -04:00
bymyself
1eb00c0a51 Upgrade to OpenAPI 3.1.0 with prefixItems and streamlined validation
- Upgraded OpenAPI spec from 3.0.3 to 3.1.0
- Replaced QueueItem oneOf pattern with prefixItems for precise tuple validation
- Simplified GitHub Actions workflow to use only Python tests (removed redundant swagger-editor validation)
- All validation now handled by openapi-spec-validator which supports OpenAPI 3.1
- QueueItem now enforces exact tuple structure: [position, prompt_id, prompt, extra_data, outputs_to_execute]
2025-07-20 19:49:41 -07:00
bymyself
7a691c980f Fix OpenAPI validation by resolving duplicate content blocks
- Fixed all duplicate YAML content entries in both main and test files
- Reverted from OpenAPI 3.1 to 3.0.3 for GitHub Actions compatibility
- QueueItem now uses minItems/maxItems instead of prefixItems
- All endpoints now have proper response schema organization
2025-07-20 16:55:47 -07:00
bymyself
6a70191868 Fix remaining duplicate content blocks in internal endpoints
- Fixed duplicate content in /api/system_stats GET
- Fixed duplicate content in /api/features GET
- Fixed duplicate content in /internal/logs GET
- Fixed duplicate content in /internal/logs/raw GET
- Fixed duplicate content in /internal/folder_paths GET
- Fixed duplicate content in /internal/files/{directory_type} GET

OpenAPI spec now passes all validation checks!
2025-07-20 14:44:27 -07:00
ComfyUI Wiki
7d627f764c
Update template to 0.1.39 (#8981) 2025-07-20 15:58:35 -04:00
comfyanonymous
a0c0785635
Document what the fast_fp16_accumulation is in the portable. (#8973) 2025-07-20 01:24:09 -04:00
chaObserv
100c2478ea
Add SamplingPercentToSigma node (#8963)
It's helpful to adjust start_percent or end_percent based on the corresponding sigma.
2025-07-19 23:09:11 -04:00
bymyself
f2cd7d3e13 Fix duplicate content blocks in model and system endpoints
- Fixed duplicate content in /api/models GET
- Fixed duplicate content in /api/models/{folder} GET
- Fixed duplicate content in /api/embeddings GET
- Fixed duplicate content in /api/extensions GET
- Consistently moved data schemas to 200 responses
2025-07-19 19:04:29 -07:00
bymyself
90e701b411 Fix more duplicate content blocks (upload and view endpoints)
- Fixed duplicate content in /api/object_info/{node_class} GET
- Fixed duplicate content in /api/upload/image POST
- Fixed duplicate content in /api/upload/mask POST
- Fixed duplicate content in /api/view GET
- Moved response schemas to appropriate status codes
2025-07-19 18:15:59 -07:00
bymyself
90d6b88739 Fix more duplicate content blocks in OpenAPI spec
- Fixed duplicate content in /api/history/{prompt_id} GET response
- Fixed duplicate content in /api/object_info GET response
- Moved actual response schemas to 200 responses where they belong
2025-07-19 15:14:17 -07:00
bymyself
f0ffa84e2f Fix first two duplicate content blocks in OpenAPI spec
- Fixed duplicate content in /api/queue GET response
- Fixed duplicate content in /api/history GET response
- Working systematically through all YAML duplicate key errors
2025-07-19 14:32:58 -07:00
ComfyUI Wiki
1da5639e86
Update template to 0.1.37 (#8967) 2025-07-19 06:08:00 -04:00
comfyanonymous
1b96fae1d4
Add nested style of dual cfg to DualCFGGuider node. (#8965) 2025-07-19 04:55:23 -04:00
bymyself
c88afc94c8 Fix duplicate content in /api/prompt GET response
- Fixed duplicate content entries in OpenAPI spec that were causing validation failures
- Maintained OpenAPI 3.1.0 and prefixItems structure
- Ensures proper YAML structure for GitHub Actions validation
2025-07-18 14:11:58 -07:00
bymyself
78fcc57e33 Update to OpenAPI 3.1.0 and use prefixItems for QueueItem
- Changed OpenAPI version from 3.0.3 to 3.1.0 as suggested in PR review
- Updated QueueItem schema to use prefixItems instead of oneOf for proper tuple representation
- Added items: false to enforce the 5-element tuple structure
- Follows OpenAPI 3.1 best practices for representing fixed-length arrays (tuples)
2025-07-18 13:46:36 -07:00
comfyanonymous
7f492522b6
Forgot this (#8957) 2025-07-18 05:43:02 -04:00
comfyanonymous
650838fd6f
Experimental CFGNorm node. (#8942)
This is from the new hidream e1 1 model code. Figured it might be useful as a generic cfg trick.
2025-07-17 04:11:07 -04:00
comfyanonymous
491fafbd64
Silence clip tokenizer warning. (#8934) 2025-07-16 14:42:07 -04:00
Harel Cain
9bc2798f72
LTXV VAE decoder: switch default padding mode (#8930) 2025-07-16 13:54:38 -04:00
comfyanonymous
50afba747c
Add attempt to work around the safetensors mmap issue. (#8928) 2025-07-16 03:42:17 -04:00
Brandon Wallace
6b8062f414
Fix MaskComposite error when destination has 2 dimensions (#8915)
Fix code that is using the original `destination` input instead of the reshaped value.
2025-07-15 21:08:27 -04:00
bymyself
13a970b3a6 Create minimal valid OpenAPI specification in root directory
- Place openapi.yaml in root directory where GitHub Actions workflow expects it
- Remove duplicate content and response keys that caused validation failures
- Use proper server URL format for validator compatibility
- Include essential endpoints and schema components for basic validation

This should resolve the 'Unable to render this definition' error in the OpenAPI validation workflow.
2025-07-15 12:43:47 -07:00
bymyself
c7f1f656a5 Fix OpenAPI specification validation errors
- Remove duplicate 'content' and '500' response entries that were causing validation failures
- Ensure proper YAML structure with unique response codes and single content definitions
- All OpenAPI spec validation tests now pass

Fixes GitHub Actions test failure in openapi-check workflow
2025-07-15 09:07:18 -07:00
comfyanonymous
b1ae4126c3
Add action to detect windows line endings. (#8917) 2025-07-15 02:27:18 -04:00
bymyself
16893d8b00 [openapi] Replace view metadata generic object with detailed schema
Replace generic 'type: object' for /api/view_metadata endpoint with
comprehensive ModelMetadata schema that includes:

- Standard modelspec fields (date, architecture, title, description)
- Implementation details (format, license, author)
- Additional metadata with proper typing
- Base64 thumbnail support

This provides much better API documentation for model metadata
responses than the previous generic object definition.
2025-07-14 19:24:51 -07:00
bymyself
3e0a4d5988 [openapi] Add comprehensive 500 Internal Server Error responses
Add 500 error responses to endpoints that previously only had 200:
- /api/prompt GET
- /api/queue GET
- /api/interrupt POST
- /api/free POST
- /api/features GET
- /api/history GET/POST
- /api/object_info endpoints
- /api/embeddings, /api/extensions, /api/system_stats
- /internal/logs and /internal/folder_paths endpoints

This provides complete error coverage for system failures and
improves API reliability documentation.
2025-07-14 19:24:51 -07:00
bymyself
5c9b27e6e7 [openapi] Improve queue item structure with proper schema
Replace generic additionalProperties: true for queue items with
detailed QueueItem schema that properly describes the array structure:

- Position number (integer)
- Prompt ID (UUID string)
- Workflow graph (object)
- Extra metadata (object)
- Output node IDs (array)

This provides much better API documentation and validation for
queue responses than the previous generic object definition.
2025-07-14 19:24:51 -07:00
bymyself
b5d6064974 [openapi] Add missing /api/features endpoint and improve schema organization
- Add /api/features endpoint with ServerFeatures response schema
- Replace inline objects in /internal/logs/raw with RawLogsResponse schema
- Add LogEntry and TerminalSize component schemas for better reusability
- Replace inline LogsSubscribeRequest object with proper schema component
- Add required field constraints to LogsSubscribeRequest

This addresses the missing features endpoint and improves schema
consistency by moving complex inline objects to reusable components.
2025-07-14 19:24:51 -07:00
bymyself
1d2b704160 [openapi] Improve upload endpoint request schemas
Enhance multipart/form-data schemas for image upload endpoints:

- Add required field constraints for image uploads
- Specify enum values for overwrite parameter ("true"/"false")
- Add default values for type ("input") and subfolder ("")
- Mark image and original_ref as required fields for mask uploads
- Improve field descriptions and validation

These changes make the API specification more precise and match
the actual server implementation requirements.
2025-07-14 19:24:51 -07:00
bymyself
7baed85b1d [openapi] Replace bare object schemas with proper component references
Replace inline object definitions for queue, history, and resource management
operations with reusable schema components:

- Add QueueManageRequest schema for /api/queue POST requests
- Add FreeResourcesRequest schema for /api/free POST requests
- Add HistoryManageRequest schema for /api/history POST requests
- Add UploadResponse schema for upload endpoints

This improves API consistency and reusability by moving common
object definitions to the components section.
2025-07-14 19:24:51 -07:00
Yoland Yan
9dabda19f0
Update nodes_gemini.py (#8912) 2025-07-14 20:59:35 -04:00
Yoland Yan
543c24108c
Fix wrong reference bug (#8910) 2025-07-14 20:45:55 -04:00
FeepingCreature
260a5ca5d9
Allow the prompt request to specify the prompt ID. (#8189)
This makes it easier to write asynchronous clients that submit requests, because they can store the task immediately.
Duplicate prompt IDs are rejected by the job queue.
2025-07-14 14:48:31 -04:00
ComfyUI Wiki
861c3bbb3d
Upate template to 0.1.36 (#8904) 2025-07-14 13:27:57 -04:00
comfyanonymous
9ca581c941
Remove windows line endings. (#8902) 2025-07-14 13:10:20 -04:00
comfyanonymous
4831e9c2c4
Refactor previous pr. (#8893) 2025-07-13 04:59:17 -04:00
Christian Byrne
480375f349
Remove auth tokens from history storage (#8889)
Remove auth_token_comfy_org and api_key_comfy_org from extra_data before
storing prompt history to prevent sensitive authentication tokens from
being persisted in the history endpoint response.
2025-07-13 04:46:27 -04:00
comfyanonymous
b40143984c
Add model detection error hint for lora. (#8880) 2025-07-12 03:49:26 -04:00
chaObserv
b43916a134
Fix fresca's input and output (#8871) 2025-07-11 12:52:58 -04:00
JettHu
7bc7dd2aa2
Execute async node earlier (#8865) 2025-07-11 12:51:06 -04:00