Verify a download

No account, no token, no throttle. Three commands from clean machine to verified model.

The three commands

1 · Download the torrent

Start from the .torrent file — it carries the webseed url-list (BEP-19), so your client can fall back to plain HTTPS when the swarm is thin:

aria2c https://aiseedbank.org/torrents/1-800-BAD-CODE_xlm-roberta_punctuation_fullstop_truecase.torrent

aria2c https://aiseedbank.org/torrents/AdamCodd_vit-base-nsfw-detector.torrent

aria2c https://aiseedbank.org/torrents/Alibaba-NLP_gte-Qwen2-1.5B-instruct.torrent

aria2c https://aiseedbank.org/torrents/Alibaba-NLP_gte-base-en-v1.5.torrent

aria2c https://aiseedbank.org/torrents/Alibaba-NLP_gte-large-en-v1.5.torrent

aria2c https://aiseedbank.org/torrents/Alibaba-NLP_gte-multilingual-base.torrent

aria2c https://aiseedbank.org/torrents/Alibaba-NLP_gte-multilingual-reranker-base.torrent

aria2c https://aiseedbank.org/torrents/Alibaba-NLP_gte-reranker-modernbert-base.torrent

aria2c https://aiseedbank.org/torrents/Arunavaonly_Bangla-twoclass-Sentiment-Analyzer.torrent

aria2c https://aiseedbank.org/torrents/AungMoonLord_bert-log-anomaly-detection.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-base-en-v1.5.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-base-zh-v1.5.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-large-en-v1.5.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-large-zh-v1.5.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-m3.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-multilingual-gemma2.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-reranker-base.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-reranker-large.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-reranker-v2-m3.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-small-en-v1.5.torrent

aria2c https://aiseedbank.org/torrents/BAAI_bge-small-zh-v1.5.torrent

aria2c https://aiseedbank.org/torrents/Bingsu_adetailer.torrent

aria2c https://aiseedbank.org/torrents/ByteDance-Seed_UI-TARS-1.5-7B.torrent

aria2c https://aiseedbank.org/torrents/CIDAS_clipseg-rd64-refined.torrent

aria2c https://aiseedbank.org/torrents/Davlan_xlm-roberta-base-ner-hrl.torrent

aria2c https://aiseedbank.org/torrents/E-MIMIC_inclusively-classification.torrent

aria2c https://aiseedbank.org/torrents/ElKulako_cryptobert.torrent

aria2c https://aiseedbank.org/torrents/FacebookAI_roberta-base.torrent

aria2c https://aiseedbank.org/torrents/FacebookAI_roberta-large.torrent

aria2c https://aiseedbank.org/torrents/FacebookAI_xlm-roberta-base.torrent

aria2c https://aiseedbank.org/torrents/FacebookAI_xlm-roberta-large.torrent

aria2c https://aiseedbank.org/torrents/Falconsai_nsfw_image_detection.torrent

aria2c https://aiseedbank.org/torrents/Freepik_nsfw_image_detector.torrent

aria2c https://aiseedbank.org/torrents/FremyCompany_BioLORD-2023.torrent

aria2c https://aiseedbank.org/torrents/Harveenchadha_vakyansh-wav2vec2-tamil-tam-250.torrent

aria2c https://aiseedbank.org/torrents/HuggingFaceTB_SmolVLM-256M-Instruct.torrent

aria2c https://aiseedbank.org/torrents/HuggingFaceTB_SmolVLM-500M-Instruct.torrent

aria2c https://aiseedbank.org/torrents/HuggingFaceTB_SmolVLM2-500M-Video-Instruct.torrent

aria2c https://aiseedbank.org/torrents/IDEA-Research_grounding-dino-base.torrent

aria2c https://aiseedbank.org/torrents/IDEA-Research_grounding-dino-tiny.torrent

aria2c https://aiseedbank.org/torrents/IlyaGusev_saiga_llama3_8b.torrent

aria2c https://aiseedbank.org/torrents/Intel_dpt-hybrid-midas.torrent

aria2c https://aiseedbank.org/torrents/KBLab_wav2vec2-large-voxrex-swedish.torrent

aria2c https://aiseedbank.org/torrents/Khalsuu_filipino-wav2vec2-l-xls-r-300m-official.torrent

aria2c https://aiseedbank.org/torrents/LGAI-EXAONE_EXAONE-3.5-7.8B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/LiquidAI_LFM2.5-1.2B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/MCG-NJU_videomae-base.torrent

aria2c https://aiseedbank.org/torrents/MahmoudAshraf_mms-300m-1130-forced-aligner.torrent

aria2c https://aiseedbank.org/torrents/Marqo_marqo-fashionSigLIP.torrent

aria2c https://aiseedbank.org/torrents/MongoDB_mdbr-leaf-ir.torrent

aria2c https://aiseedbank.org/torrents/MoritzLaurer_DeBERTa-v3-base-mnli-fever-anli.torrent

aria2c https://aiseedbank.org/torrents/MoritzLaurer_mDeBERTa-v3-base-mnli-xnli.torrent

aria2c https://aiseedbank.org/torrents/MoritzLaurer_mDeBERTa-v3-base-xnli-multilingual-nli-2mil7.torrent

aria2c https://aiseedbank.org/torrents/NX-AI_TiRex.torrent

aria2c https://aiseedbank.org/torrents/NbAiLab_nb-wav2vec2-1b-bokmaal-v2.torrent

aria2c https://aiseedbank.org/torrents/NbAiLab_nb-wav2vec2-1b-nynorsk.torrent

aria2c https://aiseedbank.org/torrents/NeuML_pubmedbert-base-embeddings.torrent

aria2c https://aiseedbank.org/torrents/NousResearch_Hermes-3-Llama-3.1-8B.torrent

aria2c https://aiseedbank.org/torrents/OpenGVLab_InternVL2-1B.torrent

aria2c https://aiseedbank.org/torrents/OpenGVLab_InternVL2-2B.torrent

aria2c https://aiseedbank.org/torrents/OpenGVLab_InternVL2_5-4B.torrent

aria2c https://aiseedbank.org/torrents/OpenGVLab_InternVL3-1B.torrent

aria2c https://aiseedbank.org/torrents/PaddlePaddle_PP-DocLayoutV3_safetensors.torrent

aria2c https://aiseedbank.org/torrents/PaddlePaddle_PP-LCNet_x1_0_doc_ori.torrent

aria2c https://aiseedbank.org/torrents/PaddlePaddle_PP-OCRv5_server_det.torrent

aria2c https://aiseedbank.org/torrents/PaddlePaddle_UVDoc.torrent

aria2c https://aiseedbank.org/torrents/PaddlePaddle_en_PP-OCRv5_mobile_rec.torrent

aria2c https://aiseedbank.org/torrents/PekingU_rtdetr_r101vd_coco_o365.torrent

aria2c https://aiseedbank.org/torrents/PekingU_rtdetr_v2_r18vd.torrent

aria2c https://aiseedbank.org/torrents/QCRI_Fanar-1-9B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen-Image.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2-0.5B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2-1.5B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2-7B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2-VL-2B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2-VL-7B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2.5-0.5B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2.5-1.5B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2.5-14B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2.5-3B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2.5-7B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2.5-Coder-1.5B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2.5-Coder-14B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2.5-Coder-7B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2.5-Math-1.5B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen2.5-VL-7B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-0.6B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-1.7B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-14B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-30B-A3B-Instruct-2507-FP8.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-4B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-4B-Thinking-2507.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-8B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-ASR-0.6B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-ASR-1.7B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-Coder-30B-A3B-Instruct-FP8.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-Embedding-0.6B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-Embedding-4B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-VL-4B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3-VL-8B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3.5-9B.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3.6-35B-A3B-FP8.torrent

aria2c https://aiseedbank.org/torrents/Qwen_Qwen3.8-27B.torrent

aria2c https://aiseedbank.org/torrents/Wan-AI_Wan2.2-TI2V-5B-Diffusers.torrent

aria2c https://aiseedbank.org/torrents/allenai_Olmo-3-7B-Instruct.torrent

aria2c https://aiseedbank.org/torrents/baidu_Unlimited-OCR.torrent

aria2c https://aiseedbank.org/torrents/black-forest-labs_FLUX.2-klein-4B.torrent

aria2c https://aiseedbank.org/torrents/deepseek-ai_DeepSeek-OCR.torrent

The magnet link on each model page is the secondarypath — it opens the swarm directly, but magnet links do not carry webseed url-lists.

2 · Fetch the signed manifest set

curl -O https://aiseedbank.org/torrents.json
curl -O https://aiseedbank.org/torrents.json.minisig
curl -O https://aiseedbank.org/minisign.pub

3 · Verify the signature

minisign -V -m torrents.json -p minisign.pub

The manifest lists every published model with its infohash, per-file hashes, and provenance. If minisign says the signature is good, the bytes you compare against are the bytes the maintainers signed — a mirror that tampered with anything fails here.

What "verified" means here

Every payload in this archive was verified against upstream Hugging Face at fetch time, with the exact upstreamrevision pinned in the signed manifest. That is the claim — nothing more. Seedbank does not audit model behavior, and the upstream of record for each model is Hugging Face; this archive preserves and re-distributes those bytes.

Client guide

• Transmission and aria2 are recommended — both implement BEP-19 webseeds solidly, so HTTPS fallback just works.
• qBittorrent works for seeding, but its webseed handling has historically lagged behind: if the swarm shows no peers and the download stalls, switch to aria2c for the initial fetch.
• Leave the client running afterward — a finished torrent seeds automatically. Share ratio 2.0+ is the goal.

Which file do I need? (formats, briefly)

safetensors vs GGUF: the archive mirrors the upstream.safetensors files verbatim — those are the canonical weights of record. GGUF files are third-party quantized repacks for local inference tools (llama.cpp family); unless a GGUF is listed in the signed manifest, it did not come from this archive.

Picking a quant: if you run a GGUF locally, a Q4 quant needs roughly 0.6 GB of RAM per billion parameters plus overhead; Q8 needs about 1 GB per billion. When in doubt, start one size smaller than you think fits.

Client RAM: the 0.6B model in the current catalog runs comfortably in 1–2 GB at Q4; 7B-class models want ~5 GB at Q4.

Ready? Browse the catalog →