AltAPIs

Aescleah/gemma-4-E4B-it-litert-lm-rope-fix

AI model by Aescleah

Open original source ↗

Older snapshot: check the original source for current information.

From the publisher

Model card & documentation

Source preview

Read the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.

A modified copy of gemma-4-E4B-it.litertlm from litert-community/gemma-4-E4B-it-litert-lm at revision 2eee7ac325f20eb8c9ac1d0e972f7c84663062da (sha256 0b2a8980ce155fd97673d8e820b4d29d9c7d99b8fa6806f425d969b145bd52e0). Weights are untouched. Only the rotary-position (RoPE) math in the graph is rewritten. The stock graph casts the INT32 position to a float tensor before computing rotary angles. With float16 activations (the GPU/Metal path) that tensor cannot represent positions above 2048 exactly, so neighbouring tokens receive the same rotary phase. The visible symptom is corrupted exact-copy sequences: digits, dates and short markers (1848 → 18848). See LiteRT-LM #3913 and #2814. The patch (patchrope.py --split, block-512 variant, from the research/3913 branch of PowerBeef/LiteRT-LM) rewrites the angle as 2048·h + l: l stays exact in float16 and the…

Aescleah/gemma-4-E4B-it-litert-lm-rope-fix on Hugging Face A short preview, not the full document.

Read the full model card ↗ · Preview checked 2026-10-29T00:44:40.750Z

Inside the original model card — Document outline
  1. Gemma 4 E4B-it (LiteRT-LM) — rotary-position fix for float16 activations
  2. What changed
  3. Measured (iPhone, GPU backend, MTP on, French numeric-fidelity corpus)
  4. Licence

Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.

Documentation belongs to its respective authors. Reported project/model license: apache-2.0. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.

Model overview

Aescleah/gemma-4-E4B-it-litert-lm-rope-fix is a task-unspecified model repository published by Aescleah on Hugging Face. The source reports the litert-lm library.

This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.

Model facts

Task
Not reported
Library
litert-lm
Recent downloads (30 days)
0
Cumulative likes
0
Hugging Face trending score
0
Reported safetensors parameters
Not reported
Architecture
Not reported
License
apache-2.0
Access
Not gated by Hugging Face
Created
2026-10-16T06:37:27.000Z
Last modified
2026-10-16T06:38:02.000Z

Compatibility and lineage

Language coverage was not reported.

Reported base models: litert-community/gemma-4-E4B-it-litert-lm

Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.

Use and evaluate

Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.

No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.

Model card and usage instructions

Model files and configuration

Community discussion

Source and freshness

Source: Hugging Face Hub. Metadata observed 2026-10-16T06:42:25.526Z. Daily imports are snapshots, not real-time monitoring.

Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.

Open original source

Related models