Synchronize MiniMax README with current baseline

This commit is contained in:
Morpheus
2026-09-28 12:56:18 +00:00
parent 1f76b723fe
commit 7408ddeb97
+24 -17
View File
@@ -5,16 +5,21 @@ baseline is the validated Turbo LoRA graph in `api-workflow.json`; the older
web-editor graph remains in `workflow.json` for reference and is not the
current baseline.
## Fixed comparison inputs
## Current baseline inputs
- Reference: `reference.png`, supplied as Picture 1
- Prompt: `prompt.txt`
- Seed: fixed within each comparison batch
- Resolution: 864×480 (480p landscape)
- Initial duration: 3 seconds
- Resolution: 960×544 landscape
- Duration: 10 seconds
- CFG: **1.0**
- Model, text encoder, VAE, batch size, and denoise: fixed
- Audio: none for the initial R2V matrix
- Audio: none
The earlier Phase A–O rows include historical 864×480/3-second and
864×480/10-second comparisons. New tests should use the current 960×544,
10-second baseline unless a row explicitly specifies another resolution or
duration.
Only the variables in the matrix may change. Before each row is run, confirm
that the named node or sampler exists on the live server. Unsupported rows
@@ -32,7 +37,7 @@ substitute another configuration.
## Current best
**MM-R2V-58** remains the 3-second production baseline: Turbo LoRA enabled, 8
**MM-R2V-58** remains the historical 3-second baseline: Turbo LoRA enabled, 8
steps, Res Multistep sampler, simple scheduler, CFG 1, 864×480, and 368.746s
runtime. **MM-R2V-61** is the quality reference for the original 864×480
10-second bright park-and-robins prompt. **MM-R2V-68** is now the production
@@ -41,9 +46,9 @@ Attention (4 chunks), and Chunk FeedForward (2 chunks), with 659.169s runtime.
## Test matrix
All rows are planned; only the scheduled MM-R2V-01 and MM-R2V-02 control/
memory tests should be run first. Later phases are gated on valid 3-second
results and should not be queued automatically with Phase 1.
The matrix records historical results, blocked plans, and the current
960×544 baseline. New tests must use the fixed current-baseline inputs unless
their row explicitly overrides resolution or duration.
| ID | Phase | Efficiency / node combination | Steps | Sampler | Scheduler | Duration (s) | CFG | Tested? | Valid? | Current Best | Runtime (s) |
|---|---|---|---:|---|---|---:|---:|---|---|---|---:|
@@ -107,7 +112,7 @@ results and should not be queued automatically with Phase 1.
| MM-R2V-58 | I Turbo sampler/scheduler | Turbo LoRA only | 8 | res_multistep | simple | 3 | 1.0 | Yes | Yes | Yes | 368.746 |
| MM-R2V-59 | J Turbo step comparison | Turbo LoRA only | 12 | res_multistep | beta | 3 | 1.0 | Yes | Yes | No | 551.740 |
| MM-R2V-60 | J Turbo step comparison | Turbo LoRA only | 12 | res_multistep | simple | 3 | 1.0 | Yes | Yes | No | 511.753 |
| MM-R2V-61 | K duration/prompt | Current Turbo baseline with bright park prompt | 8 | res_multistep | simple | 10 | 1.0 | Yes | Yes | Yes | 790.632 |
| MM-R2V-61 | K duration/prompt | Current Turbo baseline with bright park prompt | 8 | res_multistep | simple | 10 | 1.0 | Yes | Yes | No | 790.632 |
| MM-R2V-62 | L SLA attention | MM-R2V-61 + H3 SLA Attention, sparsity 0.85, block 32 | 8 | res_multistep | simple | 10 | 1.0 | Yes | Yes | No | 525.745 |
| MM-R2V-63 | L SLA attention | MM-R2V-61 + H3 SLA Attention, sparsity 0.90, block 32 | 8 | res_multistep | simple | 10 | 1.0 | Yes | Yes | No | 480.653 |
| MM-R2V-64 | M combined attention | SLA 0.85 / block 32 + Sol-Attention | 8 | res_multistep | simple | 10 | 1.0 | Yes | Yes | No | 597.210 |
@@ -145,11 +150,12 @@ Turbo/base graph. MM-R2V-51 through MM-R2V-58 completed with the validated
Turbo graph. User review promoted MM-R2V-58 for its environment quality and
lower runtime; MM-R2V-59/60 did not justify the additional 12-step runtime.
MM-R2V-61 was user-reviewed as clean with an acceptable reference match.
MM-R2V-62 was judged acceptable but lower quality than MM-R2V-61; its faster
runtime makes it the current 10-second speed baseline. MM-R2V-63 was judged
MM-R2V-62 was judged acceptable but lower quality than MM-R2V-61, and was
retained as the earlier 864×480 speed comparison. MM-R2V-63 was judged
acceptable but not as good as MM-R2V-62, so it was not promoted.
MM-R2V-68 was user-reviewed as having very good quality and is promoted as
the next baseline at 960×544.
the current baseline at 960×544. The current API graph and variables now
match this result.
MM-R2V-64 was user-reviewed as having no visible quality difference from the
SLA-only baseline, with the added Sol-Attention cost, so it was not promoted.
MM-R2V-65 was user-reviewed as having no clear quality difference from the
@@ -161,8 +167,9 @@ did not change the result quality.
## Execution order
1. Run MM-R2V-01 and MM-R2V-02 as the initial paired efficiency test.
2. Run Phase A one-node and safe-combination tests on 3 seconds.
3. Optimize steps, sampler, scheduler, and quality at 3 seconds.
4. Only then test 6/10/20-second clips.
5. Test continuation/concatenation, then latent upscale and RTX Super Resolution.
1. Use MM-R2V-68 as the current 960×544/10-second baseline.
2. Compare new attention, cache, and sampler changes against MM-R2V-68.
3. Keep prompt, reference, seed, CFG, duration, and resolution fixed unless a
test row explicitly changes them.
4. After the baseline is stable, test continuation/concatenation, then latent
upscale and RTX Super Resolution.