Downloads · 30 days
6
16% of all-time downloads
namhokaist/appgen-qwen3-vl-8b-sft-ngc-v1-frozen-policyclean-ground2x-tool2x-policy-v1-lr5e7-1ep
appgen-qwen3-vl-8b-sft-ngc-v1-frozen-policyclean-ground2x-tool2x-policy-v1-lr5e7-1ep is a image-text-to-text model from namhokaist. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as apache-2.0.
This is sweep arm ground2xtool2x from a four-arm controlled SFT study for Android GUI use. The language model was fully fine-tuned while the complete Qwen3-VL visual tower, merger, and deep-stack mergers remained froz…
Downloads · 30 days
6
16% of all-time downloads
All-time downloads
38
Public
Parameters
8.8B
17.5 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors17.5 GB · 100%
From the Hugging Face model README
This is sweep arm ground2x_tool2x from a four-arm controlled SFT study for
Android GUI use. The language model was fully fine-tuned while the complete
Qwen3-VL visual tower, merger, and deep-stack mergers remained frozen. The
uploaded weights were verified byte-for-byte against the pinned base for every
model.visual* tensor, while at least one language-model tensor was verified
to have changed.
Qwen/Qwen3-VL-8B-Instruct at commit 0c351dd01ed87e9c1b53cbc748cba10e6187ff3bluca0621/appgen-sft-ngc-v1, config ngc-f-qwen3-normalized, revision 769ea99dbc4ff190048ae0db37eb6310dba595e0policy_clean_grounding2xec34c34e50ee78ce426baaa58dcd71427cf4309688f2b6a79285fcfa319fd15dpolicy_v1; SHA-256 f47248dbafaff3e04750b75952c2ccab2dfe325705e9b850ec45452099e4f5e95e-7, cosine schedule, 5% warmuphermes loss-scale preset was used (the sweep's tool2x arm).Only the final checkpoint (checkpoint-125) is published in
this repository. The exact training system prompt is included as
appgen_system_prompt.txt; run_manifest.json records hashes and verification
evidence. Intermediate 25-step checkpoints remain local for controlled eval.
This arm starts from the 3,988-row grounding-2× variant. It removes six
publisher-identified mislabeled terminate(success) exposures: three original
rows and their three completion-replay rows. Selection uses the publisher
task_id plus publisher idx, not mutable row positions. It also removes the
undeclared time: 1 argument from exactly two long_press tool calls; only
those tool-call JSON objects were reserialized. The resulting dataset has
3,982 rows, 3,165 unique semantic examples, and no remaining undeclared
time fields.
The images are synthetic AppGen HTML-to-PNG screenshots from 50 generated training environments. They are not real-user or real-device captures. The pinned dataset declares no AndroidWorld, held-out, or public grounding benchmark images. No AndroidWorld or real-world benchmark score is claimed in this card; evaluate all four arms under the same decoding and benchmark setup.
This model is for Android visual-agent research, not safety-critical autonomous deployment.