Comfy-Org/comfyci-runpod-worker

Fork 0

mirror of https://github.com/Comfy-Org/comfyci-runpod-worker.git synced 2026-11-04 23:47:32 +00:00

[PR #10] [MERGED] Enable the four template workflows from #8 #10

New issue

Closed

opened 2026-10-31 22:06:13 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-31 22:06:13 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/Comfy-Org/comfyci-runpod-worker/pull/10
Author: @marawan206
Created: 10/18/2026
Status: ✅ Merged
Merged: 10/18/2026
Merged by: @marawan206

Base: main ← Head: enable-template-corpus

📝 Commits (1)

39f219b Enable the four template workflows from #8

📊 Changes

1 file changed (+4 additions, -4 deletions)

View changed files

📝 manifest/workflows.json (+4 -4)

📄 Description

What

Flips enabled to true for the four template workflows #8 added:

Workflow	Golden	Noise floor	Exec (4090)	VRAM peak
`qwen_image_2_1_t2i`	v0.42.0	identical (r1 == r2)	83 s	15.9 GB
`qwen_image_2_1_turbo_t2i`	v0.42.0	identical (r1 == r2)	52 s	15.9 GB
`mage_flow_t2i_int8`	v0.42.0	identical (r1 == r2)	54 s	13.2 GB
`pixeldit_t2i`	v0.42.0	identical (r1 == r2)	30 s	8.9 GB

Models

The six new files (28.5 GB) were synced from the #8 branch before the merge (run 39175189810, under three minutes, no failures); the post-merge sync (run 39176033770) found all 21 manifest files present and skipped them. The volume was grown from 150 to 200 GB the same day, so about 126 GB is in use.

Goldens

Generated twice each at v0.42.0 (master's tag today) by golden run https://github.com/Comfy-Org/comfyci-runpod-worker/actions/runs/39175491009 and blessed by https://github.com/Comfy-Org/comfyci-runpod-worker/actions/runs/39176386718. Every run reports validation.ok with no filled defaults and no stripped inputs, so the prompts the importer produced needed no server-side patching.

Each output was looked at before blessing and matches its template prompt: the greyscale editorial portrait with the lime backdrop (Qwen 2.1), the "QWEN IMAGE 2.1 TURBO" poster with the portrait grid (Qwen 2.1 Turbo), the overcoat figure under the brutalist concrete canopy in the snow (Mage-Flow), and the white dress in the sandstone canyon (PixelDiT).

Effect

With #9 this takes every new master commit from two workflows to nine. These four add about 3.7 minutes of execution plus model loads per run, roughly $0.10 per tested commit on top of #9's estimate. At the recent 4 to 9 tested commits a day the regression endpoint lands around $2 to $4/day in total.

Independent of #9; the two merge in either order.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/Comfy-Org/comfyci-runpod-worker/pull/10 **Author:** [@marawan206](https://github.com/marawan206) **Created:** 10/18/2026 **Status:** ✅ Merged **Merged:** 10/18/2026 **Merged by:** [@marawan206](https://github.com/marawan206) **Base:** `main` ← **Head:** `enable-template-corpus` --- ### 📝 Commits (1) - [`39f219b`](https://github.com/Comfy-Org/comfyci-runpod-worker/commit/39f219bc064c81239fb9bdd0a189d5a627249d2f) Enable the four template workflows from #8 ### 📊 Changes **1 file changed** (+4 additions, -4 deletions) <details> <summary>View changed files</summary> 📝 `manifest/workflows.json` (+4 -4) </details> ### 📄 Description ## What Flips `enabled` to `true` for the four template workflows #8 added: | Workflow | Golden | Noise floor | Exec (4090) | VRAM peak | |---|---|---|---|---| | `qwen_image_2_1_t2i` | v0.42.0 | identical (r1 == r2) | 83 s | 15.9 GB | | `qwen_image_2_1_turbo_t2i` | v0.42.0 | identical (r1 == r2) | 52 s | 15.9 GB | | `mage_flow_t2i_int8` | v0.42.0 | identical (r1 == r2) | 54 s | 13.2 GB | | `pixeldit_t2i` | v0.42.0 | identical (r1 == r2) | 30 s | 8.9 GB | ## Models The six new files (28.5 GB) were synced from the #8 branch before the merge (run 39175189810, under three minutes, no failures); the post-merge sync (run 39176033770) found all 21 manifest files present and skipped them. The volume was grown from 150 to 200 GB the same day, so about 126 GB is in use. ## Goldens Generated twice each at v0.42.0 (master's tag today) by golden run https://github.com/Comfy-Org/comfyci-runpod-worker/actions/runs/39175491009 and blessed by https://github.com/Comfy-Org/comfyci-runpod-worker/actions/runs/39176386718. Every run reports `validation.ok` with no filled defaults and no stripped inputs, so the prompts the importer produced needed no server-side patching. Each output was looked at before blessing and matches its template prompt: the greyscale editorial portrait with the lime backdrop (Qwen 2.1), the "QWEN IMAGE 2.1 TURBO" poster with the portrait grid (Qwen 2.1 Turbo), the overcoat figure under the brutalist concrete canopy in the snow (Mage-Flow), and the white dress in the sandstone canyon (PixelDiT). ## Effect With #9 this takes every new master commit from two workflows to nine. These four add about 3.7 minutes of execution plus model loads per run, roughly $0.10 per tested commit on top of #9's estimate. At the recent 4 to 9 tested commits a day the regression endpoint lands around $2 to $4/day in total. Independent of #9; the two merge in either order. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-31 22:06:13 +00:00

closed this issue
added the
pull-request
label

No labels

pull-request

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

Comfy-Org/comfyci-runpod-worker#10

No description provided.

Rows
Columns