18 Commits

Author	SHA1	Message	Date
James Betker	ad0f3fdd58	update to v2 models (clvp pending)	2022-04-18 17:32:54 -06:00
James Betker	8e94abd341	Support CVVP & fix for major bug in API	2022-04-18 14:47:44 -06:00
James Betker	fc8d52a998	update do_tts	2022-04-18 10:22:36 -06:00
James Betker	713281e376	update api constants	2022-04-18 09:22:15 -06:00
James Betker	c52cc78632	update	2022-04-15 08:26:11 -06:00
James Betker	979ff6e65e	implement clip-guided generation (and never use it...)	2022-04-14 21:50:57 -06:00
James Betker	776e5634fd	Remove intelligibility refinement ... It's not longer a concern. :)	2022-04-13 17:04:19 -06:00
James Betker	56f8385b99	Update sweep & eval_multiple with new voices	2022-04-13 17:03:36 -06:00
James Betker	3214ca0dfe	support latents into the diffusion decoder	2022-04-12 20:53:09 -06:00
James Betker	e2ee843098	Updates	2022-04-12 16:40:42 -06:00
James Betker	17af2df44f	support presets for generation	2022-04-10 23:19:15 -06:00
James Betker	b07fb37a78	Clip diffusion inputs	2022-04-10 19:29:32 -06:00
James Betker	b1ba8416ff	Updates	2022-04-10 14:41:13 -06:00
James Betker	33e4bc7907	integrate new autoregressive model and fix new diffusion bug	2022-04-04 16:51:35 -06:00
James Betker	9043dde3f9	Integrate new diffusion network	2022-04-01 14:15:17 -06:00
James Betker	287debd1d3	port do_tts to use the API	2022-04-01 11:55:07 -06:00
James Betker	f625a9e443	Update API to have more expressive interface for controlling various generation knobs ... - Also adds typical decoder support; unfortunately this does not work well with the current model.	2022-03-29 13:59:39 -06:00
James Betker	b78ae92890	Upgrade CLIP model and add eval_multiple	2022-03-28 19:33:31 -06:00