XReyRobert/Qwopus3.6-27B-v2-GPTQ-Pro-v1

Update README.md

Document upstream vLLM reproducibility

Remove unpublished custom image target from card

Clarify custom vLLM image scope

Document vLLM deployment image for reproducibility

Align support buttons in model card

Add support button to model card

Restore X follow button

Delete misplaced 35B runtime image asset

Remove misplaced 35B runtime image from 27B card

Add Qwopus3.6-35B-A3B runtime image to model card

Add Qwopus3.6-35B-A3B RTX 3090 image

Clarify Terminal-Bench protocol settings

Update Qwen Terminal-Bench comparison table

Fix Qwen Terminal-Bench 2.0 comparison

Add Qwen Terminal-Bench comparison

Add Terminal-Bench 2.0 evaluation report

Fix X follow button handle

Add X follow button to model card

docs: align runtime notes with metrics card layout

docs: add vLLM 0.21 stable runtime note

docs: note reproducible vLLM runtime for 131K serving

Update MMLU-Pro one-shot benchmark

Refine validation scope wording

Clarify MMLU-Pro evaluation protocol

Use Jackrong-style benchmark HTML section

Update model card with 24 May MMLU-Pro results

Remove unused SVG model card visual

Clean model card for public release

Add personal project notice

Use simple benchmark visual in model card

Add simple RTX 3090 model card visual

Update RTX 3090 Ampere benchmark visual

Embed RTX 3090 benchmark visual

Add RTX 3090 benchmark visual

Rename public repo references to short v1 name

Remove internal build lineage from model card

Prepare public v1 model card

Update v1 model card

Clarify ns256 artifact is non-MTP as published

Format RTX 3090 metrics as table

Update RTX 3090 Hermes metrics

Update model card intro and reasoning notes

Update model card with RTX 3090 validation plan

Upload GPTQ-Pro FOEM 4-bit g128 ns256 v2

initial commit