granite-4.0-h-1b-DISTILL-glm-4.7

This model is a fine-tuned version of ibm-granite/granite-4.0-h-1b trained on conversational data.

Model Details

• Base Model: ibm-granite/granite-4.0-h-1b
• Fine-tuning Dataset: TeichAI/glm-4.7-2000x
• Training Loss: 0.6364
• Context Length: 1048576 tokens

Quantized Versions (GGUF)

🔗 GGUF versions available here: granite-4.0-h-1b-DISTILL-glm-4.7-GGUF

Format	Size	Use Case
Q2_K	Smallest	Low memory, reduced quality
Q4_K_M	Recommended	Best balance
Q5_K_M	Good	Higher quality
Q8_0	Large	Near lossless
F16	Largest	Original precision

Usage

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("glogwa68/granite-4.0-h-1b-DISTILL-glm-4.7")
tokenizer = AutoTokenizer.from_pretrained("glogwa68/granite-4.0-h-1b-DISTILL-glm-4.7")

messages = [{"role": "user", "content": "Hello, how are you?"}]
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True)
outputs = model.generate(inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Ollama (GGUF)

ollama run hf.co/glogwa68/granite-4.0-h-1b-DISTILL-glm-4.7-GGUF:Q4_K_M

llama.cpp

llama-cli --hf-repo glogwa68/granite-4.0-h-1b-DISTILL-glm-4.7-GGUF --hf-file granite-4.0-h-1b-distill-glm-4.7-q4_k_m.gguf -p "Hello"

Training Details

• Epochs: 3
• Learning Rate: 2e-5
• Batch Size: 1 (with gradient accumulation)
• Precision: FP16
• Hardware: Multi-GPU with DeepSpeed ZeRO-3

License

Apache 2.0