README.md

---
library_name: residuals
base_model: ibm-granite/granite-4.0-h-tiny-base
base_model_relation: adapter
instruct_model: ibm-granite/granite-4.0-h-tiny
pipeline_tag: text-generation
tags:
- residuals
- delta
- task-arithmetic
- finetune
---


# Instruction Residuals

This repository contains instruction residuals (delta weights) computed as the parameter-wise difference between `ibm-granite/granite-4.0-h-tiny` and `ibm-granite/granite-4.0-h-tiny-base`.

Apply these residuals to the base model to reconstruct the instruction-tuned weights without retraining.


## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
from residuals import Residuals

base = AutoModelForCausalLM.from_pretrained("ibm-granite/granite-4.0-h-tiny-base")
tok = AutoTokenizer.from_pretrained("ibm-granite/granite-4.0-h-tiny-base")

res = Residuals.from_pretrained("residuals/granite-4.0-h-tiny")
res.apply(base, base_tokenizer=tok)
```


## Provenance
- **Created at**: 2025-10-25T17:44:55.705169+00:00
- **DType**: float32
- **Parameters**: 587
- **Shapes hash**: 6dc91196b3c1b84ed135c5dcd5152fd5f2dcf49c5865e1cd097b1c26693cab11
- **Names hash**: c8c3f98304c698c0e6fad6202d193586a30c58e61d250b2972183fa48fe29a9b
- **Base model**: `ibm-granite/granite-4.0-h-tiny-base`
- **Instruction model**: `ibm-granite/granite-4.0-h-tiny`

## Files
- **model.safetensors**: Serialized residual tensors (safetensors format).
- (optional) **model.safetensors.index.json** + shard files `model-00001-of-000N.safetensors`, ... for multi-part weights.
- **config.json**: Residuals metadata and provenance.
- **tokenizer files**: Saved tokenizer for compatibility.

## About this format
These are additive residuals (task vectors). Applying them to the base model's parameters reconstructs the instruction-tuned model.

## Tools
Generated with the `residuals` Python package. Install via: `pip install residuals`.
- PyPI: https://pypi.org/project/residuals/
- Source: https://github.com/omarish/residuals