HyPLM-Flash
HyPLM-Flash is a 650-million-parameter masked protein language model that returns per-residue or pooled 1280-dimensional embeddings and masked-language-model logits.
Results are recorded on linked configurations and versions of this model. Their settings and evaluations are kept separate below.
Results on linked records
This model has no results of its own in this release. 2 linked records have results. They are not assigned to the underlying model: each record keeps its own settings and evaluations, and its scores are on its own page.
- HyPLM-Flash, BF16 evaluation (HyPLM report) Configuration · configuration of this model · 40 evaluations · 40 results
- HyPLM-Flash, INT8 inference engine (HyPLM report) Configuration · configuration of this model · 1 evaluation · 1 result
Overview
An explanatory profile has not yet been reviewed for this record.
Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.
HyPLM-Flash has no ranked result on a use-case protocol in this release.
Recorded parameters: Not recorded. Licence: MIT (AI4Science/hyplm_flash/LICENSE); third-party weights, datasets and compatibility implementations keep their own licences.
Model Capabilities & Hardware Requirements
Deployment facts recorded for HyPLM-Flash. Fields the record does not carry are marked not recorded.
Evaluated Checkpoint: First release of the HyPLM family, published 2026-11-02 with FP32 weights (2.61 GB safetensors) as two GitHub Release shards; technical report dated 2026-10-12
Use this model
How it works, versions and access
Additional methodological details, versions and access terms have not yet been extracted into a structured profile for this model. Consult the linked sources and evaluation records below.
Evidence
Source checking verifies the cited claim or transcription. It does not establish independent reproduction.
Evidence table
Inspect claims, sources and review details
Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.
One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.
0 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|
No evidence rows match these filters. Choose another scope or clear the search.
Sources and history
Release 2026-11-21-c8bea25e5abb · Record review: source checked
4 source records and release history
- HyPLM Technical Report (Hyra Team, 2026-10-12) · Original source · Dated 2026-10-12 on page 1; retrieved at commit 0dcc533cc9f8f76c90ecbfe944df816b6672328a (13 pages)
- HyPLM-Flash package README (Tencent Hunyuan, commit 0dcc533c) · Original source · GitHub repository Tencent-Hunyuan/Hyra-results at commit 0dcc533cc9f8f76c90ecbfe944df816b6672328a
- HyPLM-Flash LICENSE (MIT, commit 0dcc533c) · Original source · GitHub repository Tencent-Hunyuan/Hyra-results at commit 0dcc533cc9f8f76c90ecbfe944df816b6672328a
- Hyra Results repository README (Tencent Hunyuan, commit 0dcc533c) · Original source · GitHub repository Tencent-Hunyuan/Hyra-results at commit 0dcc533cc9f8f76c90ecbfe944df816b6672328a
Technical metadata and record identifiers
Technical metadata for this record is tracked under stable ID hyplm-flash-20261106-model-hyplm-flash.
Source locator: Technical report Sections 2.1 and 2.2, Table 1; AI4Science/hyplm_flash/README.md 'Installation', 'Usage', 'Layout'
Related records
- subject: HyPLM report: accuracy in bf16
- subject: HyPLM report: baseline throughput range
- subject: HyPLM report: bf16 recommendation
- subject: HyPLM report: block design
- subject: HyPLM report: corpus sequences
- subject: HyPLM report: data ratio e1
- subject: HyPLM report: data ratio esm2
- subject: HyPLM report: data ratio esmc
- subject: HyPLM report: developer
- subject: HyPLM report: embedding dimension
- subject: HyPLM report: graph path bf16
- subject: HyPLM report: hardware
- subject: HyPLM report: heads
- subject: HyPLM report: inputs
- subject: HyPLM report: int8 default
- subject: HyPLM report: int8 preservation
- subject: HyPLM report: int8 threshold
- subject: HyPLM report: layers
- subject: HyPLM report: licence
- subject: HyPLM report: logits
- subject: HyPLM report: max length
- subject: HyPLM report: metagenomic share
- subject: HyPLM report: overlap exact
- subject: HyPLM report: overlap structures
- subject: HyPLM report: parameters
- subject: HyPLM report: pretraining stage
- subject: HyPLM report: release date
- subject: HyPLM report: second stage
- subject: HyPLM report: speed range
- subject: HyPLM report: structure chains
- subject: HyPLM report: swiglu width
- subject: HyPLM report: throughput bf16
- subject: HyPLM report: throughput readme int8
- subject: HyPLM report: training flops
- subject: HyPLM report: training tokens
- subject: HyPLM report: variant similarity
- subject: HyPLM report: vocabulary
- subject: HyPLM report: weights size install
- subject: HyPLM report: weights size layout
- subject: HyPLM report: width
- configuration of: HyPLM-Flash, BF16 evaluation (HyPLM report)
- configuration of: HyPLM-Flash, INT8 inference engine (HyPLM report)
How to cite
HyPLM-Flash. Rewire biological model benchmark database, release 2026-11-21-c8bea25e5abb. https://benchmarks.rewirebio.io/database/model/hyplm-flash-20261106-model-hyplm-flash/
Release 2026-11-21-c8bea25e5abb. Rankings and values change between releases, so keep the release id in the citation. Cite the original publication as well for any value you reuse: each result links its source.
The compilation and annotations made by Rewire in this release are offered under CC BY 4.0; values and text taken from cited sources stay under the terms of those sources and are not covered by this licence. Data licence