Commit Graph

Select branches

Hide Pull Requests

dev/4.2.0

dev/4.2.1

dev/4.3.0

dev/4.3.0-deprecated

dev/docs

dev/photography-expert-agent

dev/z-image

issue/4.1.1

main

nightly

release/4.0.0

release/4.1.0

release/4.2.0

4.1.0

latest

nightly

v4.0.0

v4.1.0

v4.2.0

v4.2.1

v4.3.0

Mono Color

• ba3b9edd84 feat: add clear_cache toggle and anti-medical jargon rules across all modes dev/4.2.1 slahiriandClaude Opus 4.5 2025-12-31 14:31:35 +05:30
• 6a83d59b7f feat: add NSFW modifier dropdown with 100 presets slahiriandClaude Opus 4.5 2025-12-31 12:41:31 +05:30
• a482473cd5 feat: add smartphone cameras to camera body dropdown slahiriandClaude Opus 4.5 2025-12-31 12:35:34 +05:30
• 8e2108b0f7 feat: add camera technical specs dropdowns to PromptModifier slahiriandClaude Opus 4.5 2025-12-31 12:31:27 +05:30
• 873a059c20 fix: differentiate token limits for local vs API LLMs slahiriandClaude Opus 4.5 2025-12-31 12:21:47 +05:30
• 3c03ac5a0b feat: optimize token sizes and add 6-Section Generate mode slahiriandClaude Opus 4.5 2025-12-31 12:14:40 +05:30
• 7d2db2d824 feat: add Extreme mode with 6-pass LLM generation slahiriandClaude Opus 4.5 2025-12-31 12:05:31 +05:30
• 24281139ba refactor: update decision engine templates with camera thinking slahiriandClaude Opus 4.5 2025-12-31 11:57:24 +05:30
• 94bf1addcf refactor: optimize prompts for Z-Image/Flux best practices slahiriandClaude Opus 4.5 2025-12-31 11:51:27 +05:30
• 2b259253e6 fix: include character presets in LLM system prompts slahiriandClaude Opus 4.5 2025-12-31 11:27:12 +05:30
• 59a63df005 feat: add comprehensive preset dropdowns to PromptModifier slahiriandClaude Opus 4.5 2025-12-31 11:19:11 +05:30
• e56308fb8b chore: move documentation to dev/docs branch slahiriandClaude Opus 4.5 2025-12-31 10:47:22 +05:30
• 5bfd23187a feat: add caching to PromptGenerator based on seed and inputs dev/docs slahiriandClaude Opus 4.5 2025-12-31 10:39:04 +05:30
• 88bb7ad1e2 feat: add caching to PromptModifier based on seed and inputs slahiriandClaude Opus 4.5 2025-12-31 10:35:10 +05:30
• dcfd7fd6c9 feat: add Expand mode to PromptModifier for generating detailed prompts slahiriandClaude Opus 4.5 2025-12-31 10:27:12 +05:30
• aac6274fbe fix: remove hard word limits, keep only repetition detection slahiriandClaude Opus 4.5 2025-12-31 10:14:46 +05:30
• bd3ceaa1e1 feat: add interrupt handling, increase token limits, add analysis report slahiriandClaude Opus 4.5 2025-12-31 10:12:21 +05:30
• 435c2bb24d fix: filter Qwen3 <think> blocks and add seed/dynamic tokens slahiriandClaude Opus 4.5 2025-12-31 09:41:36 +05:30
• e2b14e7c13 feat: add seed field and dynamic token calculation slahiriandClaude Opus 4.5 2025-12-31 09:37:40 +05:30
• 28d74114a6 fix: prevent runaway local LLM generation with repetition penalty and truncation slahiriandClaude Opus 4.5 2025-12-31 09:28:35 +05:30
• 03fb2b8f4e feat: add release_vram toggle to PromptModifier (default: enabled) slahiriandClaude Opus 4.5 2025-12-31 09:20:17 +05:30
• 295553e326 feat: add photography templates, effects, and photographer styles to PromptModifier slahiriandClaude Opus 4.5 2025-12-31 09:18:38 +05:30
• 426dd358d5 fix: use LLM node temperature in PromptModifier, prompt as input only slahiriandClaude Opus 4.5 2025-12-31 09:13:07 +05:30
• 5d914f7e01 feat: add SID_LLM_Local_Text node for text-only local models slahiriandClaude Opus 4.5 2025-12-31 08:52:18 +05:30
• 0dfd6fd294 feat: add SID_PromptModifier node for semantic prompt modification slahiriandClaude Opus 4.5 2025-12-31 08:44:59 +05:30
• 4361a7b403 feat: add Instagram caption generation and yellow node colors slahiriandClaude Opus 4.5 2025-12-31 08:28:43 +05:30
• 2fb79e2a15 fix: align detailed mode system prompt with standard/quick modes slahiriandClaude Opus 4.5 2025-12-31 01:53:46 +05:30
• a78ebb3b00 feat: add fashion pose detection with body/face angle analysis slahiriandClaude Opus 4.5 2025-12-31 01:24:08 +05:30
• 274b36a620 fix: improve pose detection with proper keypoint extraction slahiriandClaude Opus 4.5 2025-12-31 01:15:14 +05:30
• 474b02a021 feat: Z-Image/Flux prompts, female anatomy injection, detailed mode for all LLMs slahiriandClaude Opus 4.5 2025-12-31 01:10:10 +05:30
• 031f316b4b fix: remove hardcoded limits and add max_injected_tags parameter slahiriandClaude Opus 4.5 2025-12-31 00:37:10 +05:30
• 0012ad531b feat: v4.2.1 - Modular prompt generator with enhanced LLM modes slahiriandClaude Opus 4.5 2025-12-31 00:23:50 +05:30
• c1d26ad03e fix: enforce higher tag coverage in LLM prompts dev/4.3.0 slahiriandClaude Opus 4.5 2025-12-28 17:35:07 +05:30
• 11cccc5c19 feat: Stage 3 LLM enhancement with word composition analysis slahiriandClaude Opus 4.5 2025-12-28 17:30:54 +05:30
• 7c9bae939d feat: hybrid prompt composition with conflict resolution slahiriandClaude Opus 4.5 2025-12-28 16:59:14 +05:30
• 290020740d feat: enhanced category system with body/pose detection slahiriandClaude Opus 4.5 2025-12-28 11:42:51 +05:30
• 7e081049a4 fix: Places365 model download from MIT CSAIL server slahiriandClaude Opus 4.5 2025-12-28 03:57:28 +05:30
• 786926982f feat: add scene/landmark/sky taggers, improve wildlife detection, add progress bar slahiriandClaude Opus 4.5 2025-12-28 03:48:10 +05:30
• 9f6f30805b feat: Z-Image prompt pipeline redesign (WIP) slahiriandClaude Opus 4.5 2025-12-27 23:30:13 +05:30
• 5ef42c8c8e fix: enforce exact max_tokens_per_category limit slahiriandClaude Opus 4.5 2025-12-27 20:56:14 +05:30
• ec05fb561f fix: detect and clean garbage LLM output slahiriandClaude Opus 4.5 2025-12-27 20:52:01 +05:30
• 4d107597ea fix: max_tokens_per_category now properly overrides category limits slahiriandClaude Opus 4.5 2025-12-27 20:48:21 +05:30
• e0b2f000b6 fix: improve classification accuracy and conflict resolution slahiriandClaude Opus 4.5 2025-12-27 20:40:46 +05:30
• 7a56c1c220 feat: add RAG reindex script slahiriandClaude Opus 4.5 2025-12-27 20:31:51 +05:30
• 913a5c6129 feat: add visual RAG for prompt vocabulary enhancement slahiriandClaude Opus 4.5 2025-12-27 20:29:45 +05:30
• 0cac3aefe2 fix: set temperature slider minimum to 0.1 slahiriandClaude Opus 4.5 2025-12-27 19:53:51 +05:30
• 6d57304db0 feat: add pose extraction and expand ethnicity keywords slahiriandClaude Opus 4.5 2025-12-27 19:49:50 +05:30
• bb6a31bab6 fix: improve prompt accuracy with conflict resolution and token balancing slahiriandClaude Opus 4.5 2025-12-27 19:42:56 +05:30
• 2740f5dbe6 feat: add Florence pass selection and standardize node options slahiriandClaude Opus 4.5 2025-12-27 19:24:27 +05:30
• d11328ff90 feat: add STRUCTURED prompt style with category-segregated output slahiriandClaude Opus 4.5 2025-12-27 19:05:47 +05:30
• e7c28cf672 feat(nodes): integrate new compose pipeline into SID_PromptCompose slahiriandClaude Opus 4.5 2025-12-27 12:02:18 +05:30
• 84ce4de0d2 feat(compose): implement unified pipeline with NLP and LLM modes slahiriandClaude Opus 4.5 2025-12-27 11:49:21 +05:30
• 700607305b feat(compose): implement Phase 5 - Prompt Validation slahiriandClaude Opus 4.5 2025-12-27 11:41:40 +05:30
• fb05f55c9e feat(compose): implement Phase 4 - Prompt Assembly slahiriandClaude Opus 4.5 2025-12-27 11:26:40 +05:30
• da21ad1806 feat(compose): add hybrid classifier with embeddings and LLM layers slahiriandClaude Opus 4.5 2025-12-27 11:07:16 +05:30
• 596c860db2 docs: mark Phase 3 as complete slahiriandClaude Opus 4.5 2025-12-27 10:28:28 +05:30
• 574b976728 feat(compose): implement Phase 3 - Classification slahiriandClaude Opus 4.5 2025-12-27 10:27:32 +05:30
• 2721d58cd3 docs: mark Phase 2 as complete slahiriandClaude Opus 4.5 2025-12-27 10:07:00 +05:30
• bdb58fdeec feat(compose): implement Phase 2 - Tokenization slahiriandClaude Opus 4.5 2025-12-27 10:05:58 +05:30
• 06ff172b1d docs: mark Phase 1 as complete slahiriandClaude Opus 4.5 2025-12-27 09:52:28 +05:30
• c7859cf9f4 feat(compose): implement Phase 1 - Core Infrastructure slahiriandClaude Opus 4.5 2025-12-27 09:51:09 +05:30
• 7f4989f145 docs: update CLAUDE.md as implementation checkpoint slahiriandClaude Opus 4.5 2025-12-27 02:35:43 +05:30
• 7f1186d412 feat: add prompt composition pipeline and tokenization spec slahiriandClaude Opus 4.5 2025-12-27 02:33:17 +05:30
• fe041330b5 docs: add CLAUDE.md for development context slahiri 2025-12-22 02:26:51 +05:30
• f4f0a5c937 feat: expanded ethnic/traditional fashion vocabulary slahiri 2025-12-22 02:21:58 +05:30
• cb327c29a8 fix: saliency works with standard opencv-python slahiri 2025-12-22 01:41:55 +05:30
• e3a89228d0 fix: correct WD14 model list in configurator slahiri 2025-12-22 01:37:20 +05:30
• 7b7a37f459 refactor: simplify dependency management slahiriandClaude Opus 4.5 2025-12-22 01:32:43 +05:30
• 0fecf76a66 refactor: remove RAM++ and simplify JSON output slahiriandClaude Opus 4.5 2025-12-22 01:30:09 +05:30
• 1a46a9c998 fix: register Florence2Processor from model local files v4.3.0 slahiriandClaude Opus 4.5 2025-12-22 01:11:11 +05:30
• d8d2eeb726 fix: simplify Florence-2 loading - remove version checking slahiriandClaude Opus 4.5 2025-12-22 00:48:04 +05:30
• 31fbc370ad fix: remove torch/torchvision from auto-install slahiriandClaude Opus 4.5 2025-12-21 19:28:21 +05:30
• 58fe3076f8 feat: add multi-tagger pipeline and fix Florence-2 loading slahiriandClaude Opus 4.5 2025-12-21 19:26:14 +05:30
• 238ef20a59 feat: add multi-model captioning with tiered architecture slahiriandClaude Opus 4.5 2025-12-21 02:35:38 +05:30
• 706bf2ac7f docs: consolidate SID_Caption and SID_CaptionAdvanced into single node slahiriandClaude Opus 4.5 2025-12-20 20:36:12 +05:30
• 2a35976780 feat: implement SID_Caption node with Florence-2 PromptGen slahiriandClaude Opus 4.5 2025-12-20 20:22:30 +05:30
• 08bdb58fb6 docs: add v5.0 architecture and MVP implementation plan slahiriandClaude Opus 4.5 2025-12-20 19:11:27 +05:30
• f9bbb1e081 docs: add comprehensive VLM captioning documentation slahiri 2025-12-20 15:53:56 +05:30
• c92dc7f532 chore: remove assets, sample_workflows, and github workflows slahiri 2025-12-20 15:31:37 +05:30
• bb8e629091 chore: clean slate for re-architecture slahiri 2025-12-20 15:29:14 +05:30
• e3031e95bf feat: add prompt override, raw output, template registry, and caching fixes dev/4.3.0-deprecated slahiriandClaude 2025-12-20 15:14:30 +05:30
• 9a198af12a refactor: keep only Qwen3/Mistral/Llama text models slahiriandClaude Opus 4.5 2025-12-19 13:52:26 +05:30
• f148c004ba refactor: remove non-Qwen vision models, keep only QwenVL slahiriandClaude Opus 4.5 2025-12-19 13:46:47 +05:30
• f5827c3a86 chore: remove dead code after mode consolidation slahiriandClaude Opus 4.5 2025-12-19 13:06:10 +05:30
• c45de428cd fix: update default analysis_mode to 'detailed' for consistency slahiriandClaude Opus 4.5 2025-12-19 12:58:31 +05:30
• 09ac4070de refactor: consolidate analysis modes to Standard/Detailed slahiriandClaude Opus 4.5 2025-12-19 12:44:59 +05:30
• 5b7159bd54 feat: add LLM response caching for ANY provider slahiri 2025-12-19 11:33:21 +05:30
• b9f8827ba5 fix: remove leftover realism_tags reference + add rate limiting slahiri 2025-12-19 11:29:16 +05:30
• c2540d7f64 fix: add retry with exponential backoff for Anthropic 529 overload errors slahiri 2025-12-19 11:26:54 +05:30
• 994ced7a1e improve: add ANALYZE IMAGE FIRST section to replace CV hints slahiri 2025-12-19 11:18:40 +05:30
• 72605333c9 fix: remove CV hints from prompts - LLM analyzes image directly slahiri 2025-12-19 11:16:40 +05:30
• fc56efd22e fix: critical framing accuracy - stop hallucinating body parts beyond frame slahiri 2025-12-19 11:08:48 +05:30
• d7f3d18b9d improve: add anti-distortion guidance for clean generation slahiri 2025-12-19 11:00:52 +05:30
• fd73385e12 improve: template learnings from 9 unchecked debug sessions slahiri 2025-12-19 10:57:56 +05:30
• 0496987e2b feat: session-based model loading for keep_model_loaded checkbox slahiri 2025-12-19 10:42:20 +05:30
• 241d7918df improve: add meta-commentary prohibition to prompt templates slahiri 2025-12-19 10:26:19 +05:30
• c4322bd555 improve: templates from non-excluded debug session (score 7.75) slahiriandClaude Opus 4.5 2025-12-19 10:06:58 +05:30
• 8a38d6a681 feat: add exclude checkbox to debug results viewer slahiriandClaude Opus 4.5 2025-12-19 10:02:02 +05:30
• 8e2b043cd1 feat: add "exclude from learning" checkbox to generation results slahiriandClaude Opus 4.5 2025-12-19 09:56:31 +05:30
• 1bb1da8231 fix: add GAZE DIRECTION guidance for accurate pose description slahiriandClaude Opus 4.5 2025-12-19 09:51:17 +05:30