AudioLDM 2
Free.ai (self-hosted)
·
music
·
~5000 Ama-token ngalinye track
·
4.0
kusuka 4 abasebenzisi balolu hlu
AudioLDM 2 yi an imodeli yokukhiqiza umculo eyenziwe ngu Haohe Liu. Onamandla kakhulu ku Music generation from text. I-self-hosted ku-Free.ai GPUs - isebenza ngokukhululekile ngaphezu kwe-token pool yakho yansuku zonke (5,000 tokens isiqongo ngasinye). Kukhishwa ngaphansi kwe-Apache 2.0 — ukusetshenziswa kokuthengiswa kwempahla kuvunyelwe ku-Free.ai.
Sebenzisa nge-API
I-REST API ehambisana ne-OpenAI. Dala isithonjana bese ubiza le modeli emaminithini.
curl -X POST https://api.free.ai/v1/music/generate/ \
-H "Authorization: Bearer sk-free-..." \
-H "Content-Type: application/json" \
-d '{"model":"audioldm2","prompt":"your prompt here"}'Amamodeli afana
Imibuzo ebuzwa kaningi
AudioLDM 2 ikhiqiza umculo wesizinda (kunye nakwezinye izinjini, izingoma) kusuka kumbhalo osheshayo. Yazisa imizwa, uhlobo, isizinda, noma i-BPM futhi AudioLDM 2 ibuyisela umculo oqediwe.
10 kuya ku-60 imizuzwana ngayinye yohlobo ngokuya ngenjini. Ukwenza amaphuzu ade, hlanganisa izizukulwane eziningi nge /music/stitch/ noma khiqiza iziqephu zesikondi esingu-30 bese ushintsha.
Ezinye izinjini zihlanganisa izingoma ezivela kumazwi akho (faka amazwi efomu); ezinye ziyizinhlamvu kuphela. Khangela ifomu kule khasi — indawo yemazwi ikhombisa ukuthi izingoma zixhaswe.
I-pop, i-rock, i-EDM, i-jazz, i-ambient, i-classic, i-lo-fi, i-hip-hop instrumentals zonke zibonisa ngokunengqondo. Izinhlobo ze-niche (i-gamelan, i-Carnatic, ukudansa kwe-throat) zincane — zichaza amathuluzi ngokucacile ukuze zithole imiphumela engcono.
AudioLDM 2 isebenza kuma-GPU ethu ahlukile eqenjini lakho elimahhala ngosuku lokuqala; $5 → 200,000 ama-token akhokhelwe ngemuva kwalokho. Ama-token angama-5,000 ngayinye.
WAV ngokuzenzakalela (akukho phutha, ilungile ukuxhuma okuningi). MP3 itholakala ku-output picker yefayela elincane.
Hayi — AudioLDM 2 ibuyisela umculo oxhunywe. Ukuhlukanisa zibe izihlahla, sebenzisa i-output nge /music/separate/ esebenzisa i-Demucs ukukhishwa kwama-vocals, ama-drums, ama-bass, nezinye.
Yebo — /batch/ ivuma uhlu lwezicelo. Ingoma ngayinye iwela ku /account/?tab=history nesicelo njengegama lefayela. Ukulawula okuningi, sebenzisa i-API.
Yebo — POST isinqumo sakho ku /v1/music/generate/ ngemodeli="AudioLDM 2". /api/ inesihloko esigcwele esifaka phakathi amagama, isikhathi, kanye ne-tempo parameters.
AudioLDM 2 ayikhiqizi ama-copyrighted compositions — ama-outputs akhiqizwa ngokusha. Free.ai inikeza ukusetshenziswa kokuthengiswayo; ugcina ilungelo lemisindo yakho ekhiqizwe. Ukubiza umculi kwi-prompt icela "ngesitayela se" — umsindo okhona uvela ekuqaleni.
15 imizuzwana kuya kumizuzu emi-2 ngayinye ye-track ngokuya ngesikhathi + injini. Imodeli ye-premium iphele ngokushesha. Sebenzisa inkinobho yokulinganisa ukuvala i-tab — imiphumela iwela ku /account/?tab=history uma kwenziwe.
Yebo — iphutha libuyisela emuva ngokuzenzakalela. Amabhayisikobho ahambayo kodwa azwakala anzima, thumela i-imeyili ku contact@free.ai nge-link yokuhlukanisa futhi sizobheka i-case-by-case.