AbstractPhil
·
AI & ML interests
datasets, research papers, experimentation, vision, classification, text encoders, tokenization, llms, diffusion, distillation, and more.
Recent Activity
repliedto their post about 2 hours ago We proudly present the preview version Qwen 3.5 4b - Lucia V1.1 apache-2.0 prototype.
https://huggingface.co/AbstractPhil/qwen3.5-4b-think-symbolic-image
The space is set up with ZeroGPU capacity for testing. You can use Look With Help to combine a blend of prompts, or simply Free Look to test your images with just the model herself.
https://huggingface.co/spaces/AbstractPhil/think-symbolic-images-testbed
The first of multiple prototypes. Trained using >10,000 rows of highly complex, high quality, curated apache-friendly data, sourced as described in the recent posts and the readme.
https://huggingface.co/datasets/AbstractPhil/thought-symbolic-image-recaptioned
Respect the licenses.
This model is not as strong as the original for museum data. In the original's case, Claude and I decided to focus on the museum information due to it being a weak-point of Qwen, sort of improve the weakness. For the new version of Lucia we couldn't use the Claude data, so the specific museum data is on the to-do list of data to curate for deep thought, not just shallow or medium thought. Really give the Lucia line some introspection capacity in the real academic sense. So the upcoming dataset improvements will improve the Lucia line substantially over time.
The current model and training process preserves most of the Qwen 3.5 4b structure via very low token drift from the original model, while presenting an extractable utilizable format of multi-introspective chain of thought, formatted into a series of image captioning structures.
The upcoming 4b train ready by tomorrow; is the non-commercial train with nearly 4x more rows. This non-commercial variant includes a large amount of deepfashion rows, docci rows, and coco rows. A humongous experimental model train, meant to be used for non-commercial and research purposes only.
We would like to have the 9b apache variant ready by Monday and the non-commercial variant by hopefully next Friday, so the specialized forms can be trained. View all activity Organizations
view article Alhena: Building a Rotary That Withstands Selection
AbstractPowered
• • 1
published an article about 1 month ago view article One Address, Three Systems: The Formulas of AlephLM, AlephLLM and AMOE, and the Measurements That Set Them
published an article about 1 month ago view article Scaling Beatrix: A 376-Million-Parameter Byte Model, the Arms That Followed Her Trunk, and the First Sliders in Pictures
view article Twinning Beatrix: A Full-Splat Byte Model, Its Softmax Control, and What Reaches an Image Generator
view article Raising Beatrix: A Byte-Level Model's Measured Childhood
view article Agreement, Anchors, Addresses: A Week of Geometric Training
view article Geometric Memory FT4 — Distill Against a Consensus, Ship a Rotation
AbstractPhil
• • 1
view article The Loss Manifest: A Field History of Objective Functions, and What a Machine Can Actually Be Asked to Compute
view article Aleph Differentiation, Parts 3 & 3-D: Two Laws, Five Days, One Framework
view article The Aleph Moves Into a Pretrained Trunk: Relays, Registers, and the Two-Regime Dispatch Law
view article The Aleph Under Autoregressive Pressure: Bottleneck Priors, Sign Codes, and the Consumption Law
view article Subject Bucketing: Teaching a Diffusion Model New Prompt Languages Without Forgetting
AbstractPhil
• • 1
view article geolip-aleph-void: The First Relational Geometric Vocabulary Patchwork
view article Reading the Voids: Topological Contribution Signals in Frozen Geometric Codebooks
view article Fused Batched Thin SVD, Part II: Extending the Jacobi Pipeline to N=6 with Configurable Convergence
view article H2 Omega Confirmed, Paradigm Shift: Attempting to Disprove Omega As A Whole
AbstractPhil
• • 1
view article The Polygonal Omega: Trained Sphere-Solvers Are Projective Codebooks
AbstractPhil
• • 1
view article Three Geometric Bands in a Sphere-Normalized Patch Autoencoder