this isthe 4bit model !!!
Updated to include LangChain code and documentation:
as the model previoulsy fcould not generate correct code: Transformer documentation was also re added (not lost) as python code fragments as well as markdown pages and html pages: This data was fit to 0.9 ( the bible was attempted again but the problem is still at 2,4.....perhaps i will have to have a specialized session for this data. ).
.. in training this model only 2million parameters were moved: ....its easy fit data as it can already program!! ( no need for deep embedding)
a lora of rank 1:1 <<< Usefull for data which will only need to be slightly fit or data that the model is already close to:
hence for bible training i will have to move more tensors!!! making a more drastic effect on the model !!!(not greatly desirable)<<<< Here we will have to create a specific merge candidate : and merge the data into the prime model instead!!
Uploaded model
- Developed by: LeroyDyer
- License: apache-2.0
- Finetuned from model : LeroyDyer/Mixtral_AI_LCARS_tg_1
This mistral model was trained 2x faster with Unsloth and Huggingface's TRL library.
- Downloads last month
- 25