After a lot positive feedback on Progenitor V1.1, I got some advice regarding a couple of settings which I could finetune for hopefully better results. Mainly changing the tokenizer and letting the merge compute at full float32 before scaling down to bfloat16. With these two changes in place, and the rest identical to the Progenitor V1.1, I present Progenitor V2.1!
merge
This is a merge of pre-trained language models created using
mergekit
.
Progenitor-V2.1-LLaMa-70B huggingface.co is an AI model on huggingface.co that provides Progenitor-V2.1-LLaMa-70B's model effect (), which can be used instantly with this TareksGraveyard Progenitor-V2.1-LLaMa-70B model. huggingface.co supports a free trial of the Progenitor-V2.1-LLaMa-70B model, and also provides paid use of the Progenitor-V2.1-LLaMa-70B. Support call Progenitor-V2.1-LLaMa-70B model through api, including Node.js, Python, http.
Progenitor-V2.1-LLaMa-70B huggingface.co is an online trial and call api platform, which integrates Progenitor-V2.1-LLaMa-70B's modeling effects, including api services, and provides a free online trial of Progenitor-V2.1-LLaMa-70B, you can try Progenitor-V2.1-LLaMa-70B online for free by clicking the link below.
TareksGraveyard Progenitor-V2.1-LLaMa-70B online free url in huggingface.co:
Progenitor-V2.1-LLaMa-70B is an open source model from GitHub that offers a free installation service, and any user can find Progenitor-V2.1-LLaMa-70B on GitHub to install. At the same time, huggingface.co provides the effect of Progenitor-V2.1-LLaMa-70B install, users can directly use Progenitor-V2.1-LLaMa-70B installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
Progenitor-V2.1-LLaMa-70B install url in huggingface.co: