README.md

metadata

license: apache-2.0
language:
  - en
base_model:
  - prithivMLmods/GCIRS-Reasoning-1.5B-R1
pipeline_tag: text-generation
library_name: transformers
tags:
  - text-generation-inference
  - code
  - math
  - RL
  - science

GCIRS-Reasoning-1.5B-R1-GGUF

GCIRS-Reasoning-1.5B-R1 is a research-grade reasoning model fine-tuned from Qwen2.5-1.5B-Instruct, focused on non-fictional reasoning, factual consistency, and scientific depth. Trained with reinforcement learning using the Big Reasoning Traces dataset from DeepSeek, this model is tailored for complex analytical tasks and scientific rigor in high-stakes or research environments.

Model Files

File Name	Format	Size	Precision	Use Case
GCIRS-Reasoning-1.5B-R1.F32.gguf	GGUF	7.11 GB	F32	Highest precision, research use
GCIRS-Reasoning-1.5B-R1.BF16.gguf	GGUF	3.56 GB	BF16	High precision, balanced performance
GCIRS-Reasoning-1.5B-R1.F16.gguf	GGUF	3.56 GB	F16	High precision, memory efficient
GCIRS-Reasoning-1.5B-R1.Q8_0.gguf	GGUF	1.89 GB	Q8_0	Excellent quality, moderate compression
GCIRS-Reasoning-1.5B-R1.Q6_K.gguf	GGUF	1.46 GB	Q6_K	Very good quality, good compression
GCIRS-Reasoning-1.5B-R1.Q5_K_M.gguf	GGUF	1.29 GB	Q5_K_M	Balanced quality/size (recommended)
GCIRS-Reasoning-1.5B-R1.Q5_K_S.gguf	GGUF	1.26 GB	Q5_K_S	Good quality, smaller size
GCIRS-Reasoning-1.5B-R1.Q4_K_M.gguf	GGUF	1.12 GB	Q4_K_M	Good balance for most users
GCIRS-Reasoning-1.5B-R1.Q4_K_S.gguf	GGUF	1.07 GB	Q4_K_S	Decent quality, compact size
GCIRS-Reasoning-1.5B-R1.Q3_K_L.gguf	GGUF	980 MB	Q3_K_L	Lower quality, very compact
GCIRS-Reasoning-1.5B-R1.Q3_K_M.gguf	GGUF	924 MB	Q3_K_M	Fast inference, limited quality
GCIRS-Reasoning-1.5B-R1.Q3_K_S.gguf	GGUF	861 MB	Q3_K_S	Fastest inference, basic quality
GCIRS-Reasoning-1.5B-R1.Q2_K.gguf	GGUF	753 MB	Q2_K	Minimal size, experimental use

Quick Selection Guide

• For Research/Development: Use F32 or BF16 for maximum accuracy
• For Production (Recommended): Use Q5_K_M or Q6_K for best quality/performance balance
• For General Use: Use Q4_K_M or Q4_K_S for good performance
• For Resource-Constrained Environments: Use Q3_K_M or Q3_K_L
• For Edge Devices: Use Q2_K for minimal footprint

Quants Usage

(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)

Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):