RichardErkhov/FairMind_-_Llama-3-8B-4bit-UltraChat-Ita-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Llama-3-8B-4bit-UltraChat-Ita - GGUF

• Model creator: https://huggingface.co/FairMind/
• Original model: https://huggingface.co/FairMind/Llama-3-8B-4bit-UltraChat-Ita/

Name	Quant method	Size
Llama-3-8B-4bit-UltraChat-Ita.Q2_K.gguf	Q2_K	2.96GB
Llama-3-8B-4bit-UltraChat-Ita.IQ3_XS.gguf	IQ3_XS	3.28GB
Llama-3-8B-4bit-UltraChat-Ita.IQ3_S.gguf	IQ3_S	3.43GB
Llama-3-8B-4bit-UltraChat-Ita.Q3_K_S.gguf	Q3_K_S	3.41GB
Llama-3-8B-4bit-UltraChat-Ita.IQ3_M.gguf	IQ3_M	3.52GB
Llama-3-8B-4bit-UltraChat-Ita.Q3_K.gguf	Q3_K	3.74GB
Llama-3-8B-4bit-UltraChat-Ita.Q3_K_M.gguf	Q3_K_M	3.74GB
Llama-3-8B-4bit-UltraChat-Ita.Q3_K_L.gguf	Q3_K_L	4.03GB
Llama-3-8B-4bit-UltraChat-Ita.IQ4_XS.gguf	IQ4_XS	4.18GB
Llama-3-8B-4bit-UltraChat-Ita.Q4_0.gguf	Q4_0	4.34GB
Llama-3-8B-4bit-UltraChat-Ita.IQ4_NL.gguf	IQ4_NL	4.38GB
Llama-3-8B-4bit-UltraChat-Ita.Q4_K_S.gguf	Q4_K_S	4.37GB
Llama-3-8B-4bit-UltraChat-Ita.Q4_K.gguf	Q4_K	4.58GB
Llama-3-8B-4bit-UltraChat-Ita.Q4_K_M.gguf	Q4_K_M	4.58GB
Llama-3-8B-4bit-UltraChat-Ita.Q4_1.gguf	Q4_1	4.78GB
Llama-3-8B-4bit-UltraChat-Ita.Q5_0.gguf	Q5_0	5.21GB
Llama-3-8B-4bit-UltraChat-Ita.Q5_K_S.gguf	Q5_K_S	5.21GB
Llama-3-8B-4bit-UltraChat-Ita.Q5_K.gguf	Q5_K	5.34GB
Llama-3-8B-4bit-UltraChat-Ita.Q5_K_M.gguf	Q5_K_M	5.34GB
Llama-3-8B-4bit-UltraChat-Ita.Q5_1.gguf	Q5_1	5.65GB
Llama-3-8B-4bit-UltraChat-Ita.Q6_K.gguf	Q6_K	6.14GB
Llama-3-8B-4bit-UltraChat-Ita.Q8_0.gguf	Q8_0	7.95GB

Original model description:

language:

• it license: apache-2.0 tags:
• text-generation-inference
• transformers
• unsloth
• llama
• trl
• sft base_model: unsloth/llama-3-8b-bnb-4bit datasets:
• mii-community/ultrafeedback-translated-ita

Evaluation

---

Uploaded model

• Developed by: walid-iguider
• License: apache-2.0
• Finetuned from model : unsloth/llama-3-8b-bnb-4bit

This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.

For a detailed comparison of model performance, check out the Leaderboard for Italian Language Models.

Here's a breakdown of the performance metrics:

Metric	hellaswag_it acc_norm	arc_it acc_norm	m_mmlu_it 5-shot acc	Average
Accuracy Normalized	0.6064	0.4611	0.5328	0.5334