Llama-3.2-1B-Instruct-bnb-4bit-lima - Merged Model

Full-precision (16-bit) merged model with LoRA adapters integrated.

Model Details

Base Model: unsloth/Llama-3.2-1B-Instruct-bnb-4bit
Format: merged_16bit
Dataset: GAIR/lima
Size: ~8-16GB
Usage: transformers

Related Models

LoRA Adapters: fs90/Llama-3.2-1B-Instruct-bnb-4bit-lima-lora - Smaller LoRA-only adapters
GGUF Quantized: fs90/Llama-3.2-1B-Instruct-bnb-4bit-lima-GGUF - GGUF format for llama.cpp/Ollama

Prompt Format

This model uses the Llama 3.2 chat template.

Python Usage

Use the tokenizer's apply_chat_template() method:

messages = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Your question here"}
]
inputs = tokenizer.apply_chat_template(messages, tokenize=True, return_tensors="pt")

Training Details

LoRA Rank: 64
Training Steps: 480
Training Loss: 1.1123
Max Seq Length: 2048
Training Scope: 1,278 samples (3.0 epoch(s), full dataset)

For complete training configuration, see the LoRA adapters repository/directory.

Benchmark Results

Evaluated: 2025-11-24 03:10 Comparison: Fine-tuned vs Base model

HuggingFace Transformers (16-bit merged model)

IFEval (Instruction Following)

Model	Strict Prompt	Strict Inst	Loose Prompt	Loose Inst
Base	0.4399	0.5731	0.4787	0.6067
Fine-tuned	0.3050	0.4376	0.3327	0.4700
Δ	↘ -0.1349	↘ -0.1355	↘ -0.1460	↘ -0.1367

Summary

Benchmark	What It Tests	Base	Fine-tuned	Improvement
IFEval	Tests ability to follow specific instructions	43.99%	30.50%	↘ -13.49% (-30.7%)
GSM8K	Tests math reasoning and chain-of-thought	-	-	-
HellaSwag	Tests real-world knowledge and common sense	-	-	-
MMLU	Tests broad knowledge retention (detects catastrophic forgetting)	-	-	-
TruthfulQA	Tests tendency to generate truthful answers	-	-	-

Usage

With Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained(
    "./outputs/Llama-3.2-1B-Instruct-bnb-4bit-lima/merged_16bit",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("./outputs/Llama-3.2-1B-Instruct-bnb-4bit-lima/merged_16bit")

messages = [{"role": "user", "content": "Your question here"}]
inputs = tokenizer.apply_chat_template(messages, tokenize=True, return_tensors="pt").to("cuda")
outputs = model.generate(inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0]))