WithinUsAI
/

IBM-Grok4-Ultra.Fast.Coder-1B

Text Generation

granitemoehybrid

Model card Files Files and versions

IBM-Grok4-UltraFast-Coder-1B

This model is a full fine-tuned derivative of ibm-granite/granite-4.0-1b.

Training notes: - Full model fine-tuning - No adapters - No LoRA - No QLoRA - Dual-GPU DDP training on Kaggle

Downloads last month: 1,341

Safetensors

Model size

2B params

Tensor type

BF16

·

Model tree for WithinUsAI/IBM-Grok4-Ultra.Fast.Coder-1B

Base model

ibm-granite/granite-4.0-1b-base

Finetuned

ibm-granite/granite-4.0-1b

Finetuned

(9)

this model

Quantizations

Datasets used to train WithinUsAI/IBM-Grok4-Ultra.Fast.Coder-1B

Collection including WithinUsAI/IBM-Grok4-Ultra.Fast.Coder-1B

“IBM Granite 4.0”

All models are “IBM Granite 4.0” at core fine-tuned, merged & trained by (WithIn Us AI) • 2 items • Updated 2 days ago • 1