File size: 4,586 Bytes
3ee8a95 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 | {
"reference_materials": {
"OpenGVLab--InternVL3-1B__config.json": {
"repo": "OpenGVLab/InternVL3-1B",
"revision": "4415a3b810e636d11dfa86b0e9ba40bb00535aa8",
"file": "config.json",
"url": "https://huggingface.co/OpenGVLab/InternVL3-1B/resolve/4415a3b810e636d11dfa86b0e9ba40bb00535aa8/config.json",
"sha256": "8a252be385dbef3abf1c58d028ea5116849a33dc7a1601c350c3ee28f202524c"
},
"OpenGVLab--InternVL3-1B__modeling_intern_vit.py": {
"repo": "OpenGVLab/InternVL3-1B",
"revision": "4415a3b810e636d11dfa86b0e9ba40bb00535aa8",
"file": "modeling_intern_vit.py",
"url": "https://huggingface.co/OpenGVLab/InternVL3-1B/resolve/4415a3b810e636d11dfa86b0e9ba40bb00535aa8/modeling_intern_vit.py",
"sha256": "d3e0f68f29e86be1f3e6e101dc45303ec24e5115e0034de101665972576eb348"
},
"OpenGVLab--InternVL3-1B__configuration_intern_vit.py": {
"repo": "OpenGVLab/InternVL3-1B",
"revision": "4415a3b810e636d11dfa86b0e9ba40bb00535aa8",
"file": "configuration_intern_vit.py",
"url": "https://huggingface.co/OpenGVLab/InternVL3-1B/resolve/4415a3b810e636d11dfa86b0e9ba40bb00535aa8/configuration_intern_vit.py",
"sha256": "22dd687a5f142985407675fc6b1dc688c2ded93e54192543ecf28ef495a6efa8"
},
"OpenGVLab--InternVL3-1B-hf__config.json": {
"repo": "OpenGVLab/InternVL3-1B-hf",
"revision": "014c0583a0d4bedf29fbe2dbff4f865eb998e171",
"file": "config.json",
"url": "https://huggingface.co/OpenGVLab/InternVL3-1B-hf/resolve/014c0583a0d4bedf29fbe2dbff4f865eb998e171/config.json",
"sha256": "7dd610d98f11ea7d37f5086a990eca7df22f0b5d7a36c5dbf41d060b97bdf901"
},
"Efficient-Large-Model--Sana_600M_512px_diffusers__transformer__config.json": {
"repo": "Efficient-Large-Model/Sana_600M_512px_diffusers",
"revision": "83d7a190bfd1fd070570a793d2dab5c7a3231b9d",
"file": "transformer/config.json",
"url": "https://huggingface.co/Efficient-Large-Model/Sana_600M_512px_diffusers/resolve/83d7a190bfd1fd070570a793d2dab5c7a3231b9d/transformer/config.json",
"sha256": "44dfc5f16b396035371d4b5a56ef3773a9be9e0c679337b0336b2b81b78be3b8"
},
"Efficient-Large-Model--Sana_600M_512px_diffusers__vae__config.json": {
"repo": "Efficient-Large-Model/Sana_600M_512px_diffusers",
"revision": "83d7a190bfd1fd070570a793d2dab5c7a3231b9d",
"file": "vae/config.json",
"url": "https://huggingface.co/Efficient-Large-Model/Sana_600M_512px_diffusers/resolve/83d7a190bfd1fd070570a793d2dab5c7a3231b9d/vae/config.json",
"sha256": "4e5669aa03caa77615b73b21c9347c384cb11124a9590683e32a510df0176eb4"
},
"Efficient-Large-Model--Sana_600M_512px_diffusers__scheduler__scheduler_config.json": {
"repo": "Efficient-Large-Model/Sana_600M_512px_diffusers",
"revision": "83d7a190bfd1fd070570a793d2dab5c7a3231b9d",
"file": "scheduler/scheduler_config.json",
"url": "https://huggingface.co/Efficient-Large-Model/Sana_600M_512px_diffusers/resolve/83d7a190bfd1fd070570a793d2dab5c7a3231b9d/scheduler/scheduler_config.json",
"sha256": "70e8603120cc387316aa3bc7f08cb2872f6b6eb6eff6f25df1b88b6b76389e81"
},
"mit-han-lab--dc-ae-f32c32-sana-1.1-diffusers__config.json": {
"repo": "mit-han-lab/dc-ae-f32c32-sana-1.1-diffusers",
"revision": "6f7b3f3b289a439a11ef4fb1034989fd4b9a4766",
"file": "config.json",
"url": "https://huggingface.co/mit-han-lab/dc-ae-f32c32-sana-1.1-diffusers/resolve/6f7b3f3b289a439a11ef4fb1034989fd4b9a4766/config.json",
"sha256": "d448e075595d708140d53102b3e21362845ffdfcf5c62638256fe37559c55c34"
}
},
"external_decoder": {
"url": "https://raw.githubusercontent.com/nnnth/UniLIP/937ee73991d68e1a53c1ac156e1a62e6b251e6aa/unilip/model/vae_modules.py",
"sha256": "5666de0ad561036d5206583881a81367502ee7f8628ae67dae9fb022b521838a",
"included": false,
"reason": "The official code repository has no root license; do not redistribute this file. Supply a local official checkout explicitly to the loader."
},
"local_loader": "Independent packaging glue; uses the original trained tensor names and licensed constructors.",
"changes": [
"InternVision configuration disables optional FlashAttention for CPU; native attention equations retained.",
"DC-AE from_pretrained constructor is replaced with config-only meta construction inside a scoped patch; strict loading subsequently supplies every trained decoder tensor.",
"Qwen connector keeps the original final six layers and removes its unused token embedding."
]
}
|