Undi95/Nete-13B

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kPublished:Oct 25, 2023License:cc-by-nc-4.0Architecture:Transformer0.0K Open Weights Warm

Undi95/Nete-13B is a 13 billion parameter language model developed by Undi95, built upon the Xwin-MLewd recipe. This model is a powered-up version of Xwin-MLewd-13B, incorporating elements from various other 13B models and LoRAs. It is designed for general text generation tasks, leveraging a blend of fine-tuned components to enhance its capabilities.

Loading preview...

Undi95/Nete-13B: A Blended 13B Language Model

Nete-13B is a 13 billion parameter language model developed by Undi95, representing an enhanced iteration of the Xwin-MLewd-13B recipe. This model integrates components from several other 13B models and LoRAs, aiming to achieve improved performance through a diverse blend of fine-tuned elements.

Key Components and Influences

Nete-13B's development involved merging aspects from various models and LoRAs, including:

  • Undi95/Mlewd-v2.4-13B
  • Xwin-LM/Xwin-LM-13B-V0.2
  • cgato/Thespis-13b-v0.4
  • Undi95/PsyMedRP-v1-13B
  • Undi95/Storytelling-v2.1-13B-lora
  • lemonilia/LimaRP-Llama2-13B-v3-EXPERIMENT

This blending approach suggests an effort to combine strengths from different specialized models, potentially leading to a more versatile and robust language generation capability. The model utilizes the Alpaca prompt template for instruction-following tasks.

Intended Use Cases

Given its foundation and the models it incorporates, Nete-13B is suitable for a range of general-purpose text generation and instruction-following applications, particularly where a blend of creative and structured response capabilities is beneficial.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p