EvaByte Model Card

EvaByte is a 6.5B byte-level language model built upon an improved architecture with multibyte prediction and EVA -- an efficient attention mechanism designed for scalability and performance. Trained on 1.5T bytes spanning natural language text, math, and code, EvaByte demonstrates the viability of efficient byte-level processing at scale -- rivaling top open-source tokenizer-based LMs using 5x less training data, excelling in coding tasks, and decoding up to 2x faster.

Model Resources

Repository: https://github.com/openevabyte/evabyte
Blog: https://hkunlp.github.io/blog/2025/evabyte and https://sambanova.ai/blog/evabyte-efficient-byte-level-language-models-at-scale
Paper: Coming soon

Model Details

EvaByte is trained using the performant SambaNova SN30 RDU system with a batch size of 8M bytes and 32K context length. The training process consists of 3 phases: after pre-training on 1.2T bytes (yielding EvaByte-Phase1), two independent annealing runs (100B and 200B bytes respectively) are conducted with learning rate linearly decayed from 1e-4 to 0. The resulting checkpoints are merged via model soup (EvaByte), which then undergoes supervised fine-tuning (EvaByte-SFT).

Stage	Model
Base (before annealing)	EvaByte-Phase1 <-- you are here
Base	EvaByte
SFT	EvaByte-SFT

Usage

Note: Make sure to set trust_remote_code=True when loading the model (or tokenizer), as our implementation includes custom code.

The code snippet below demonstrates EvaByte-6.5B for completion:

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("evabyte/EvaByte-Phase1", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("evabyte/EvaByte-Phase1", torch_dtype=torch.bfloat16, trust_remote_code=True).eval().to("cuda")

prompt = "The quick brown fox jumps "

# Tokenize input
# Option 1: standard HF tokenizer interface
input_ids = tokenizer(prompt, return_tensors="pt").input_ids.to("cuda")

# Option 2: Direct UTF-8 byte encoding with offset
# Note: Each byte is offset by 64 with <bos> prepended.
input_ids = torch.tensor([[1] + [b + 64 for b in prompt.encode("utf-8")]]).to("cuda")

# byte-by-byte generation (default)
generation_output = model.generate(
    input_ids=input_ids, 
    max_new_tokens=32
)
# alternatively, use faster multibyte generation
generation_output = model.multi_byte_generate(
    input_ids=input_ids, 
    max_new_tokens=32
)

# Decode and print the output
response = tokenizer.decode(
    generation_output[0][input_ids.shape[1]:], 
    skip_special_tokens=False,
    clean_up_tokenization_spaces=False
)
print(response)
# Sample output:
# over the lazy dog.\n\nThe quick

⚙️ Generation Modes

EvaByte supports two generation interfaces:

model.generate(): The default generation method compatible with Huggingface transformers library. This approach generates one byte at a time and might be slow.
model.multi_byte_generate(): A faster alternative that generates multiple bytes per step and usually yields the same result as model.generate() under greedy decoding, with the implementation adapted from Medusa. model.multi_byte_generate() supports a subset of arguments in model.generate():
- input_ids: the input byte ids.
- temperature: the temperature for sampling.
- max_length: the maximum length of the generated sequence.
- max_new_tokens: the maximum number of new bytes to generate.
- stopping_criteria: the stopping criteria for generation.
- top_p: the top-p parameter for sampling.
- do_sample: greedy decoding or sampling.

Notes and Limitations:

device_map="auto" is not supported for >2 GPUs.
Only batch size of 1 (with attention_mask=None) is supported for decoding.
torch_dtype=torch.bfloat16 is required.
The multibyte generation model.multi_byte_generate() might return extra bytes after the end-of-sequence sentinel, due to the nature of the multibyte decoding. Manual truncation or cleaning may be needed.

Bias, Risks, and Limitations

As a pretrained base model, EvaByte-Phase1 has not been fine-tuned for chat or instruction following, so users should not expect reliable performance in conversational or instruction-based tasks. Like other base models, it does not incorporate any moderation mechanisms, making it possible to generate potentially harmful or inappropriate content.

Evaluation

For detailed evaluation results, check out our blog post at SambaNova or HKUNLP.

Citation

@misc{evabyte,
    title = {EvaByte: Efficient Byte-level Language Models at Scale},
    url = {https://hkunlp.github.io/blog/2025/evabyte},
    author = {Lin Zheng and Xueliang Zhao and Guangtao Wang and Chen Wu and David Dong and Angela Wang and Mingran Wang and Yun Du and Haige Bo and Amol Sharma and Bo Li and Kejie Zhang and Changran Hu and Urmish Thakker and Lingpeng Kong},
    year = {2025}
}