---
base_model: unsloth/gemma-4-E4B-it
tags:
- text-generation-inference
- transformers
- unsloth
- gemma4
- gguf
license: apache-2.0
language:
- en
datasets:
- bitext/Bitext-customer-support-llm-chatbot-training-dataset
---

# Uploaded finetuned  model

- **Developed by:** Prompt48
- **License:** apache-2.0
- **Finetuned from model :** unsloth/gemma-4-E4B-it

This gemma4 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.

[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)

## Training Dataset

This model was fine-tuned on the [Bitext Customer Support LLM Chatbot Training Dataset](https://huggingface.co/datasets/bitext/Bitext-customer-support-llm-chatbot-training-dataset).

- **Creator:** Bitext Innovations
- **Dataset License:** CDLA-Sharing-1.0
- **Language:** English (en-US)
- **Size:** 26,872 question/answer pairs (~1,000 per intent), 3.57 million total tokens
- **Format:** Each sample contains: `flags`, `instruction`, `category`, `intent`, `response`

The dataset covers **10 categories** and **27 intents** for customer support chatbot applications:
ACCOUNT (create, delete, edit, switch), CANCELLATION_FEE, DELIVERY (options), FEEDBACK (complaint, review), INVOICE (check, get), NEWSLETTER (subscription), ORDER (cancel, change, place), PAYMENT (check methods, payment issue), REFUND (check policy, track), SHIPPING_ADDRESS (change, set up).

It includes 12 linguistic variation tags capturing lexical, syntactic, register, and stylistic diversity, and 30 entity/slot placeholder types. Applicable across 20+ industry verticals (retail, banking, healthcare, travel, etc.).

## GGUF Quantizations

| Filename | Quantization | Size | BPW | Description |
|----------|-------------|------|-----|-------------|
| `gemma-4-e4b-it-ft-IQ4_NL.gguf` | IQ4_NL | 4.9 GB | 5.53 | Good quality, recommended for most use cases |
| `gemma-4-e4b-it-ft-IQ3_M.gguf` | IQ3_M | 4.4 GB | 5.00 | Smaller size, slightly reduced quality |
| `gemma-4-e4b-it-ft-Q4_K_M.gguf` | Q4_K_M | 5.0 GB | 5.66 | K-quant medium, good balance of quality and size |

### Original Safetensors

The original finetuned safetensors model files are also included in this repository.

### Source

Quantized from [Prompt48/gemma-4-E4B-it-FT](https://huggingface.co/Prompt48/gemma-4-E4B-it-FT)