Aquiles-ai/Chargaff-Tokenizer
Updated • 1
AI & ML interests
None defined yet.
Recent Activity
FredyRivera-dev
updated a
model 9 days ago
FredyRivera-dev
published a
model 9 days ago
FredyRivera-dev
updated 2
models 24 days ago
FredyRivera-dev
updated a
collection 24 days ago
FredyRivera-dev
posted an update 24 days ago
Post
75
I ported Evo 2 inference to plain PyTorch and Transformers so the official checkpoints load through AutoModelForCausalLM with no Vortex, Transformer Engine, or custom kernels required.
What is included:
- GitHub repo with the port, a Vortex to HF converter, and a gene completion repro script: https://github.com/Aquiles-ai/Evo2-transformers
- Aquiles-ai/Evo2-1B-Base: the 1B base checkpoint (8k context) in Transformers format: Aquiles-ai/Evo2-1B-Base
- Aquiles-ai/Evo2-7B: the 7B checkpoint (1M context) in Transformers format: Aquiles-ai/Evo2-7B
Both repos vendor the modeling files, so loading needs trust_remote_code=True. The tokenizer matches the original byte level behavior, including the vortex_tokenize and vortex_detokenize helpers.
One honest note: this port is less efficient than the original Vortex stack. It has no FlashAttention, no FP8 path, and no recurrent state decoding, so generation recomputes the prefix and long contexts are slow. For large scale or 1M context work, use the original implementation or NVIDIA NIM.
All credit for the model, data, and research goes to the Evo 2 team at Arc Institute and collaborators. Original project: https://github.com/ArcInstitute/evo2. Paper: https://www.nature.com/articles/s41586-026-10176-5
Feedback and corrections are welcome as GitHub issues on the port repo.
What is included:
- GitHub repo with the port, a Vortex to HF converter, and a gene completion repro script: https://github.com/Aquiles-ai/Evo2-transformers
- Aquiles-ai/Evo2-1B-Base: the 1B base checkpoint (8k context) in Transformers format: Aquiles-ai/Evo2-1B-Base
- Aquiles-ai/Evo2-7B: the 7B checkpoint (1M context) in Transformers format: Aquiles-ai/Evo2-7B
Both repos vendor the modeling files, so loading needs trust_remote_code=True. The tokenizer matches the original byte level behavior, including the vortex_tokenize and vortex_detokenize helpers.
One honest note: this port is less efficient than the original Vortex stack. It has no FlashAttention, no FP8 path, and no recurrent state decoding, so generation recomputes the prefix and long contexts are slow. For large scale or 1M context work, use the original implementation or NVIDIA NIM.
All credit for the model, data, and research goes to the Evo 2 team at Arc Institute and collaborators. Original project: https://github.com/ArcInstitute/evo2. Paper: https://www.nature.com/articles/s41586-026-10176-5
Feedback and corrections are welcome as GitHub issues on the port repo.
FredyRivera-dev
published 2
models 24 days ago
FredyRivera-dev
updated a
Space about 1 month ago
FredyRivera-dev
updated a
dataset about 1 month ago
FredyRivera-dev
posted an update about 2 months ago
Post
3447
I've written a technical blog post about how we create a multimodal model:
Kairos: a multimodal model built with LFM2.5-2.6B as the LLM, MoonViT-3D (the vision tower of Kimi-K2.6) as the vision encoder, and a custom projector.
The original plan was LLaVA's approach, two stages: first align the projector with the LLM frozen, and then train the projector + LLM together. The first stage worked in terms of loss (ablation with +3.7 nats in favor of the image), but in free generation the image shifted the logits without changing the argmax: the model received the image and ignored it. That's why we jumped directly to early fusion, with a reasoning dataset.
For that, we created Kairos-Multimodal-Reasoning: 116,357 examples with explicit reasoning traces, generated through distillation (60,041 from LLaVA-CC3M, 2,295 from WebSight, and 54,021 from Zebra-CoT), with GPT 5.6 Luna, Inkling, Qwen 3.6 27B, and Qwen 3.7 Plus as teachers.
The training, in two phases:
1. Projector through backbone with 80k image-text pairs (Kairos-Proj-80k).
2. Projector + LoRA (r=16) with 30k examples from the reasoning dataset (Kairos-Alig-30k).
Everything is open source:
- Full blog post with the process: https://aquiles-ai.vercel.app/blog/kairos-a-multimodal-model
- Implementation: https://github.com/Aquiles-ai/Kairos
To be honest: the checkpoints are not a competent model, they are experimental artifacts. But they validated the approach and precisely defined what the next iteration needs.
https://huggingface.co/collections/Aquiles-ai/kairos
Aquiles-ai/MoonViT-3D
LiquidAI/LFM2.5-2.6B
Kairos: a multimodal model built with LFM2.5-2.6B as the LLM, MoonViT-3D (the vision tower of Kimi-K2.6) as the vision encoder, and a custom projector.
The original plan was LLaVA's approach, two stages: first align the projector with the LLM frozen, and then train the projector + LLM together. The first stage worked in terms of loss (ablation with +3.7 nats in favor of the image), but in free generation the image shifted the logits without changing the argmax: the model received the image and ignored it. That's why we jumped directly to early fusion, with a reasoning dataset.
For that, we created Kairos-Multimodal-Reasoning: 116,357 examples with explicit reasoning traces, generated through distillation (60,041 from LLaVA-CC3M, 2,295 from WebSight, and 54,021 from Zebra-CoT), with GPT 5.6 Luna, Inkling, Qwen 3.6 27B, and Qwen 3.7 Plus as teachers.
The training, in two phases:
1. Projector through backbone with 80k image-text pairs (Kairos-Proj-80k).
2. Projector + LoRA (r=16) with 30k examples from the reasoning dataset (Kairos-Alig-30k).
Everything is open source:
- Full blog post with the process: https://aquiles-ai.vercel.app/blog/kairos-a-multimodal-model
- Implementation: https://github.com/Aquiles-ai/Kairos
To be honest: the checkpoints are not a competent model, they are experimental artifacts. But they validated the approach and precisely defined what the next iteration needs.
https://huggingface.co/collections/Aquiles-ai/kairos
Aquiles-ai/MoonViT-3D
LiquidAI/LFM2.5-2.6B
FredyRivera-dev
published a
model about 2 months ago
FredyRivera-dev
updated 3
models about 2 months ago
FredyRivera-dev
published a
model about 2 months ago
FredyRivera-dev
updated a
collection about 2 months ago