README.md

---
base_model:
- tiiuae/falcon-11B
library_name: transformers
tags:
- mergekit
- merge
- lazymergekit
license: apache-2.0
language:
- de
---
## Why prune?

Even though [Falcon-11B](https://huggingface.co/tiiuae/falcon-11B) is trained on 5T tokens, it is still undertrained, as can be seen by this graph:
![image/png](https://cdn-uploads.huggingface.co/production/uploads/660c0a02cf274b3ab77dd6b7/QeaL9bOrPskustzFpjMUP.png)
This is why the choice is made to prune 50% of the layers.
Note that \~1B of continued pre-training (\~1M rows of 1k tokens) is still required to restore the perplexity of this model in the desired language.
I'm planning on doing that for certain languages, depending on how much compute will be available.

# sliced

This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).

## Merge Details
### Merge Method

This model was merged using the passthrough merge method.

### Models Merged

The following models were included in the merge:
* [tiiuae/falcon-11B](https://huggingface.co/tiiuae/falcon-11B)

### Configuration

The following YAML configuration was used to produce this model:

```yaml
slices:
  - sources:
      - model: tiiuae/falcon-11B
        layer_range: [0, 24]
  - sources:
      - model: tiiuae/falcon-11B
        layer_range: [55, 59]
merge_method: passthrough
dtype: bfloat16
```

[PruneMe](https://github.com/arcee-ai/PruneMe) has been utilized using the wikimedia/wikipedia German (de) subset by investigating layer similarity with 2000 samples. The layer ranges for pruning were determined based on this analysis to maintain performance while reducing model size.

![Layer Similarity Plot](https://cdn-uploads.huggingface.co/production/uploads/660c0a02cf274b3ab77dd6b7/k9VKXgqUuUr0EjGZf7Ick.png)

```python
from transformers import AutoTokenizer, AutoModelForCausalLM
import transformers
import torch

model = "ssmits/Falcon2-5.5B-German"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    torch_dtype=torch.bfloat16,
)
sequences = pipeline(
   "Can you explain the concepts of Quantum Computing?",
    max_length=200,
    do_sample=True,
    top_k=10,
    num_return_sequences=1,
    eos_token_id=tokenizer.eos_token_id,
)
for seq in sequences:
    print(f"Result: {seq['generated_text']}")

```

💥 **Falcon LLMs require PyTorch 2.0 for use with `transformers`!**

For fast inference with Falcon, check-out [Text Generation Inference](https://github.com/huggingface/text-generation-inference)! Read more in this [blogpost]((https://huggingface.co/blog/falcon). 

## Direct Use
Research on large language models; as a foundation for further specialization and finetuning for specific usecases (e.g., summarization, text generation, chatbot, etc.)

## Out-of-Scope Use
Production use without adequate assessment of risks and mitigation; any use cases which may be considered irresponsible or harmful.

## Bias, Risks, and Limitations
Falcon2-5.5B is trained mostly on English, but also German, Spanish, French, Italian, Portuguese, Polish, Dutch, Romanian, Czech, Swedish. It will not generalize appropriately to other languages. Furthermore, as it is trained on a large-scale corpora representative of the web, it will carry the stereotypes and biases commonly encountered online.

## Recommendations
We recommend users of Falcon2-5.5B to consider finetuning it for the specific set of tasks of interest, and for guardrails and appropriate precautions to be taken for any production use.