README.md

---
license: apache-2.0
pipeline_tag: text-generation
tags:
- model_hub_mixin
- pytorch_model_hub_mixin
- RxNN
- SparseQueryAttention
- SQA
language:
- en
datasets:
- wikimedia/wikipedia
library_name: RxNN
---

# SQAT-m: Sparse Query Attention Transformer mini
Research model for [Sparse Query Attention](https://github.com/RxAI-dev/RxNN/blob/main/docs/research/sparse_query_attention.md)
experiments - extension to Grouped Query Attention, that's also reducing the number of used query heads, instead of further
reducing key/value heads count (up to Multi Query Attention). That approach results in huge computational complexity reduction
and much faster training, while the performance stays on GQA level (almost unnoticeable decrease, when compared to GQA, and
noticeable better than MQA).

### Architecture details:
- trainable params: ~10.7M
- dim: 256
- layers: 8
- self-attention: Sparse Query Attention
  - heads: 16 (for dimension split)
  - query groups: 8
  - key/value groups: 4
- SwiGLU feed forward with 768 dim
- RoPE
- RMS Norm
- vocab: 10k (english only)
- message length: 1024
- Library: RxNN

### Training details:
This model was only trained for research purposes, on a small number of training steps. As it's the most promising from
tested attention architectures, it will be developed further soon.
- dataset: 50% from english subset of [wikimedia/wikipedia](https://huggingface.co/datasets/wikimedia/wikipedia) (45% train / 5% validation)
- single epoch
- 1.5B processed tokens
- learning rate: 5e-4, cosine annealing scheduler with 25% warmup steps

### Results
Validation mean loss/accuracy:
- MHA: 1.1976 / ~77.35%
- GQA: 1.2177 / ~77.12%
- MQA: 1.2497 / ~76.64%
- **SQA: 1.2272 / ~76.97%**

Training time / time per batch:
- MHA: ~269 min / 0.7173s
- GQA: ~258 min / 0.6877s
- MQA: ~261 min / 0.6947s
- **SQA: ~241 min / 0.6417s**

### Computational complexity comparison
- MHA: `O(N*d * N*d)`
- GQA `O(N*d * N*(d/heads*groups))`
- MQA `O(N*d * N*(d/heads))`
- SQA `O(N*(d/heads*query_groups) * N*(d/heads*groups))`

SQA has reduced two factors instead of one. That means it will better scale for longer sequences and training time gains
will be even greater.

Furthermore, even _the extreme version_ of **SQA** with only 4/16 used query heads (and also 4/16 key/value heads), seems to perform a little
better than a reference MQA model, with even shorter training times. It suggests that **SQA** could be a gamechanger for efficient
long context handling. More info in [ReactiveAI/xSQAT-m](https://huggingface.co/ReactiveAI/xSQAT-m)

### Model size difference
SQA has reduced dimensions of query heads linear projection and output projection, which results in a little smaller model size:
- MHA: 12M Params
- GQA: 11.2M Params
- MQA: 11M Params
- **SQA: 10.7M Params**

### Usage
Model requires [RxNN framework](https://github.com/RxAI-dev/RxNN) for training/inference. It's integrated with HuggingFace Hub and libraries.

#### Inference:
- Install RxNN, PyTorch and dependencies: `pip install rxnn torch transformers tokenizers`
```python
import torch
from rxnn.experimental.models import ExperimentalAttentionTransformer
from rxnn.transformers.sampler import Sampler, SampleDecoder
from rxnn.training.tokenizer import load_tokenizer_from_hf_hub

model = ExperimentalAttentionTransformer.from_pretrained('ReactiveAI/SQAT-m')
tokenizer = load_tokenizer_from_hf_hub('ReactiveAI/SQAT-m')
sampler = Sampler(model, torch.device('cuda' if torch.cuda.is_available() else 'cpu'), end_token_id=3)
sample = SampleDecoder(sampler, tokenizer)

# 0.1 and 0.9 are default values for temperature and top_p
generated = sample('Example model input for text generation...', temperature=0.1, top_p=0.9, max_seq_len=1024)
sample('Example model input for text generation - print streamed response...', temperature=0.1, top_p=0.9, max_seq_len=1024, print_stream=True)
```

#### Train:
- Install RxNN, PyTorch and dependencies: `pip install rxnn torch transformers tokenizers tensorboard` (`tensorboard` is optional)
```python
import torch
from rxnn.experimental.models import ExperimentalAttentionTransformer
from rxnn.training.tokenizer import load_tokenizer_from_hf_hub
from rxnn.training.dataset import AutoregressiveLMDataset
from rxnn.training.bml import AutoregressiveTrainer
from rxnn.training.callbacks import PrintLossCallback, PrintAccuracyCallback, TokenCounterCallback, ModelSaveCallback
from rxnn.training.scheduler import get_transformer_lr_scheduler

model = ExperimentalAttentionTransformer.from_pretrained('ReactiveAI/SQAT-m')
tokenizer = load_tokenizer_from_hf_hub('ReactiveAI/SQAT-m')

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

batch_size = 128 # Require ~40GB GPU Memory (trained on L40S)
epochs = 1
gradient_acc_steps = 1
seq_len = 1024
vocab_size = 10_000

peak_lr = 5e-4 * gradient_acc_steps

train_dataset = AutoregressiveLMDataset.from_hf_hub('hf-dataset-id', 'subset', tokenizer=tokenizer, max_seq_len=seq_len) # split is 'train' by default
valid_dataset = AutoregressiveLMDataset.from_hf_hub('hf-dataset-id', split='validation', tokenizer=tokenizer, max_seq_len=seq_len)

dataset_len = len(train_dataset)

steps_per_epoch = int(dataset_len / batch_size - 1)
total_steps = int((epochs * steps_per_epoch) / gradient_acc_steps)
warmup_steps = int(0.25 * steps_per_epoch)


logs_dir = './tensorboard_logs' # require tensorboard `pip install tensorboard`

print_cb = PrintLossCallback(batches_per_epoch=steps_per_epoch)
count_cb = TokenCounterCallback()
acc_cb = PrintAccuracyCallback()
save_cb = ModelSaveCallback('./path/to/save', push_to_hub=True,
                            hub_model_id='your-model-id', private_repo=True,
                            push_checkpoint_weights=True, final_commit_message='Final commit message', hf_token=YOUR_HF_TOKEN)

trainer = AutoregressiveTrainer(model, device, dataset=train_dataset, validation_dataset=valid_dataset,
                         vocab_size=vocab_size, callbacks=[print_cb, acc_cb, count_cb, save_cb], use_amp=True,
                         dtype=torch.bfloat16, log_dir=logs_dir, gradient_accumulation_steps=gradient_acc_steps)

optimizer = torch.optim.AdamW(model.parameters(), lr=peak_lr, weight_decay=0.01)
scheduler = get_transformer_lr_scheduler(
    optimizer,
    warmup_steps=warmup_steps,
    num_training_steps=total_steps
)

trainer(epochs=epochs, batch_size=batch_size, optimizer=optimizer, scheduler=scheduler)
```

## Summary
According to experiment results, SparseQueryAttention seems to be the most cost-effective variant of GroupedQueryAttention,
leading to noticeable training time reduction and is a promising research direction. Currently, for our **Reactive Tranformer**
architectures that were initially designed with GQA for self-attention and MQA for memory-attention, we consider using SQA
instead for all attention layer types. More info will be released soon.