docs/SUBMISSION_GUIDE.md

Submission Guide

Required Package

• Hugging Face Space URL: https://huggingface.co/spaces/build-small-hackathon/ObjectverseDiary
• GitHub Repository URL: https://github.com/qqyule/Objectverse-Diary
• Demo Video Script: docs/DEMO_VIDEO_SCRIPT.md
• Social Media Post Draft: docs/SOCIAL_POST.md
• Fine-tuned Model URL: https://huggingface.co/qqyule/objectverse-diary-qwen15b-lora
• Dataset URL: https://huggingface.co/datasets/qqyule/objectverse-diary-sft-curated
• Trace Dataset: local public mock JSONL export at data/traces/samples/objectverse_public_mock_traces.jsonl
• Field Notes Draft: docs/FIELD_NOTES.md
• Short project description: available in README

Local Evidence Ready

• Initial mock MVP report: docs/INITIAL_STAGE_REPORT.md
• Runtime boundary: docs/RUNTIME.md
• Dataset plan and preview workflow: docs/DATASET.md
• External setup checklist: docs/EXTERNAL_SETUP.md
• Space VLM validation report: docs/SPACE_VLM_REPORT.md currently passes for public mug, keyboard, and shoe images on ZeroGPU with OBJECTVERSE_VISION_BACKEND=minicpm-v.
• Space VLM diagnostics: hidden /vision_runtime_probe API confirms Torch/Transformers, CUDA, and MiniCPM-V model load status.
• Live Space runtime: ZeroGPU zero-a10g with OBJECTVERSE_VISION_BACKEND=minicpm-v, VISION_MODEL_ID=openbmb/MiniCPM-V-2_6, and OBJECTVERSE_TEXT_BACKEND=mock.
• Space VLM trace evidence: data/traces/space-vlm/
• Public mock traces: data/traces/samples/
• Stable demo baseline: Gradio example buttons replay committed sample traces first, then fall back to the live generation pipeline if a cached trace is missing.
• Optional llama.cpp runtime wiring: src/models/llama_cpp_runner.py
• Published LoRA v2 Q4_K_M GGUF: https://huggingface.co/qqyule/objectverse-diary-qwen15b-lora/blob/main/objectverse-diary-qwen15b-lora-v2-q4_k_m.gguf

Completed Locally

• Mock MVP flow, archive-style UI, share card, trace logging, sample traces, dataset preview, and initial acceptance tooling.
• Stable local demo baseline with six replayable example outputs, shared cached/live UI formatting, chat wake state, share card, and trace panel output.
• MiniCPM-V 2.6 backend wiring with fallback markers.
• Optional llama.cpp text runtime wiring through TEXT_MODEL_PATH.
• Hosted Space VLM validation script, report, JSON summary, and trace evidence export.
• Hosted Space VLM probe support, latest failure-note update support, and passing MiniCPM-V ZeroGPU validation after adding an HF_TOKEN Space secret for gated model access.
• Local GGUF smoke-test helper passed with models/objectverse-diary-qwen15b-lora-v2-q4_k_m.gguf; trace text runtime was llama-cpp text generation and no text-fallback-to-mock was present.
• Synthetic curated v2 SFT dataset published to Hugging Face Datasets: 200 rows, 40 objects, 5 personality modes.
• Modal Qwen 1.5B LoRA v2 run completed and adapter published to Hugging Face Models.
• LoRA v2 adapter merged into Qwen/Qwen2.5-1.5B-Instruct, converted with pinned llama.cpp, quantized to Q4_K_M, and uploaded to the same model repo.
• Field Notes draft, demo video script, and social post draft for the stable submission package.

Not Completed Yet

• Hosted Space text runtime validation with the published GGUF. The local runtime passed, but the public Space intentionally remains on mock text for the live MiniCPM-V vision release.
• Real text-model traces from the hosted Space.
• Field Notes publication URL, recorded demo video URL, social post URL, and final public submission.

Final Checks

• Space is under the official organization.
• Space MiniCPM-V validation passes for mug, keyboard, and shoe.
• Space is configured for live MiniCPM-V vision on ZeroGPU with mock text.
• Space MiniCPM-V non-secret diagnostic probe is implemented locally.
• Demo video script targets under 2 minutes.
• README includes stable-baseline parameter budget and links to the model card.
• No commercial cloud AI APIs are used.
• Mock-safe local demo baseline is reproducible from committed sample traces.
• Fine-tuned model is linked.
• Dataset is linked.
• Traces are linked.
• Field Notes are linked.
• UI remains English-first and Chinese-second.
• Submission is complete before June 15, 2026.