AudioLDM 2
Free.ai (self-hosted)
·
music
·
~5000 Ii-token track
·
3.9
I- 8 Abasebenzisi
AudioLDM 2 yi an Imodeli yokudala umculo eyenziwe ngu Haohe Liu. Enamandla kakhulu kwi Music generation from text. I-self-hosted kwi-Free.ai GPUs — iqhuba ngokukhululekileyo kwi-token pool yakho yosuku (5,000 tokens Ingoma nganye). Ikhutshwe phantsi kwe Apache 2.0 - ukusetyenziswa korhwebo kuvunyelwe kwi Free.ai.
Sebenzisa nge API
I-REST API ehambelana ne-OpenAI. Yenza iqhosha kwaye unxulumane nale modeli kwimizuzu.
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"}'Iimodeli ezifanayo
Imibuzo ebuzwa rhoqo
AudioLDM 2 ivelisa iingoma zesithethi sangaphakathi (kunye kwezinye iinjini, iingoma) ukusuka kumbhalo okhawulezayo. Chaza umphefumlo, uhlobo, isithethi, okanye i BPM kwaye AudioLDM 2 ibuyisela umkhondo wesandi ogqibekileyo.
10 ukuya kwimizuzu eyi-60 nganye yohlobo oluthile luxhomekeke kwinjini. Kwiinxalenye ezide, dibanisa uhlobo oluninzi lugqitha /music/stitch/ okanye yenza kwicandelo lemizuzu eli-30 nokufa okuthe tyaba.
Ezinye iinjini zidibanisa iingoma ezivela kwiingoma ozinikayo (cofa iingoma kwifomu); ezinye zii-instrumental kuphela. Khangela ifomu kule phepha - indawo yengoma ithetha ukuba iingoma zixhaswe.
I-pop, i-rock, i-EDM, i-jazz, i-ambient, i-classical, i-lo-fi, i-hip-hop instrumentation zonke zibonisa ngokufanelekileyo. Ii-genres ze-niche (i-gamelan, i-Carnatic, ukudanisa kwengcuka) zibuthathaka - chaza ii-instrumentations ngokucacileyo ukuze ufumane iziphumo ezilungileyo.
AudioLDM 2 iqhuba kwi-GPU yethu ekhethekileyo echaphazela i-pool yakho ekhululekileyo yosuku kuqala; $5 → 200,000 i-token ehlawulweyo emva koko. malunga ne-~5,000 i-token nganye.
WAV ngokumiselweyo (engenakulahleka, ilungile ekuxhumeni okungaphezulu). MP3 ifumaneka kwimveliso yokukhethiweyo yefayile ezincinci.
U-No. AudioLDM 2 ubuyisela umculo odityanisiweyo. Ukuwuhlula ngezijungqu, qhuba i-output nge /music/stem-splitter/, ekhupha iingoma, i-drums, i-bass nezinye.
Ewe - /iqela/ lixhasa uluhlu lweempendulo. Ingoma nganye iwela kwi /iakhawunti/?tab=imbali ngempendulo njengegama lefayili. Ulawulo oluninzi, sebenzisa i-API.
Ewe - UTHENGA umyalezo wakho ku /v1/music/generate/ ngemodeli "AudioLDM 2". /api/ inesiqendu esipheleleyo esiquka amagama, ixesha, kunye nexesha elimiselweyo.
AudioLDM 2 ayiphindanga iinkcazo ezikhuselwe yi-copyright - iziphumezo zidityanisiwe zitsha. Free.ai inikezela ukusetyenziswa korhwebo; ugcina ilungelo lemihlaba yakho eyenziweyo. Ukunika igama umculi kwi prompt icela "kwindlela ye" - isandi sangempela sivela ekuqaleni.
15 seconds to 2 minutes per track depending on duration + engine. Premium models finish faster. Use the queue button to close the tab - result lands in /account/?tab=history when done.
Ewe - i-failure ibonisa ubuyiselo oluzenzekelayo. Kwiingoma eziqhubayo kodwa ezingenasandi esibi, thumela i-imeyili contact@free.ai ngekhonkco lokudibanisa kwaye siza kujonga iimeko nganye ngenye.