🧠
Model

Qwen3 4b Base Think Grpo Kl0.001 Lr1e 5 Bf16 540steps

by L1nkee l1nkee/qwen3-4b-base-think-grpo-kl0.001-lr1e-5-bf16-540steps
Free2AITools Nexus Index
39.0
S: Semantic 50

Query-time baseline · scored live at search

A: Authority 0
P: Popularity 0
R: Recency 99
Q: Quality 65
Tech Context
4.02 Params
32.768K Ctx
Vital Performance —

Technical Constraints

Experimental / High Latency
Low FNI signal 39 FNI Score
4.02B Params
32k Context
0 Downloads
8G GPU ~6GB Est. VRAM
Dense QWEN3FORCAUSALLM Architecture
Commercial APACHE License
Model Information Summary
Entity Passport
Registry ID l1nkee/qwen3-4b-base-think-grpo-kl0.001-lr1e-5-bf16-540steps
License Apache-2.0
Provider huggingface
πŸ’Ύ

Compute Threshold

~5.5GB VRAM

Interactive
Estimate fit
β–Ό

* Static estimation for 4-Bit Quantization.

πŸ“œ

Cite this model

Academic & Research Attribution

BibTeX
@misc{l1nkee_qwen3_4b_base_think_grpo_kl0_001_lr1e_5_bf16_540steps,
  author = {L1nkee},
  title = {Qwen3 4b Base Think Grpo Kl0.001 Lr1e 5 Bf16 540steps Model},
  year = {2026},
  howpublished = {\url{https://huggingface.co/L1nkee/Qwen3-4B-Base-Think-GRPO-KL0.001-LR1e-5-bf16-540steps}},
  note = {Accessed via Free2AITools.}
}
APA Style
L1nkee. (2026). Qwen3 4b Base Think Grpo Kl0.001 Lr1e 5 Bf16 540steps [Model]. Free2AITools. https://huggingface.co/L1nkee/Qwen3-4B-Base-Think-GRPO-KL0.001-LR1e-5-bf16-540steps

πŸ”¬Technical Deep Dive

Full Specifications [+]

Quick Commands

πŸ¦™ Ollama Run
ollama run qwen3-4b-base-think-grpo-kl0.001-lr1e-5-bf16-540steps
πŸ€— HF Download
huggingface-cli download l1nkee/qwen3-4b-base-think-grpo-kl0.001-lr1e-5-bf16-540steps
πŸ“¦ Install Lib
pip install -U transformers

βš–οΈ Free2AITools Nexus Index V2.0

Semantic (S) 50

Query-time baseline · scored live at search

Authority (A) 0
Popularity (P) 0
Recency (R) 99
Quality (Q) 65

πŸ’¬ Index Insight

FNI V2.0 for Qwen3 4b Base Think Grpo Kl0.001 Lr1e 5 Bf16 540steps: Authority (A:0), Popularity (P:0), Recency (R:99), Quality (Q:65). Semantic (S) is a query-time baseline scored live at search.

Free2AITools Nexus Index

Data Sources / Provenance

Open data Updated: Live data
---

πŸš€ What's Next?

Technical Deep Dive

⚠️ Incomplete Data

Some information about this model is not available. Use with Caution - Verify details from the original source before relying on this data.

View Original Source β†’

πŸ“ Limitations & Considerations

  • β€’ Benchmark scores may vary based on evaluation methodology and hardware configuration.
  • β€’ VRAM requirements are estimates; actual usage depends on quantization and batch size.
  • β€’ FNI scores are relative rankings and may change as new models are added.
πŸ”„ Updated daily

Source summary: Based on Hugging Face metadata. Not a recommendation.

πŸ“Š FNI Methodology πŸ“š Knowledge Baseℹ️ Verify with original source

πŸ›‘οΈ Model Transparency Report

Technical metadata sourced from upstream repositories.

Open Metadata

πŸ†” Identity & Source

id
hf-model--l1nkee--qwen3-4b-base-think-grpo-kl0.001-lr1e-5-bf16-540steps
slug
l1nkee--qwen3-4b-base-think-grpo-kl0.001-lr1e-5-bf16-540steps
source
huggingface
author
L1nkee
license
Apache-2.0
tags
transformers, safetensors, qwen3, text-generation, verl, reinforcement-learning, grpo, math, dataset:bytedtsinghua-sia/dapo-math-17k, base_model:qwen/qwen3-4b-base, base_model:finetune:qwen/qwen3-4b-base, license:apache-2.0, text-generation-inference, endpoints_compatible, region:us

βš™οΈ Technical Specs

architecture
Qwen3ForCausalLM
params billions
4.02
context length
32,768
pipeline tag
reinforcement-learning
vram gb
5.5
vram is estimated
true
vram formula
VRAM β‰ˆ (params * 0.75) + 2GB (KV) + 0.5GB (OS)

πŸ“Š Engagement & Metrics

downloads
0

Data indexed from public sources. Updated daily.