Llama 3.3 70B Instruct (Q4_K_M) — 47.4 GBon Scaleway L40S-1-48G
Overview
Llama 3.3 70B Instruct is a 70B parameter dense language model by Meta, with code, multilingual, tool-calls capabilities. It supports a context window of up to 131,072 tokens.
Llama 3.3 70B Instruct is a 70-billion-parameter dense transformer model from Meta, optimized for instruction following, code generation, and multilingual conversation. It delivers performance competitive with larger models in the Llama family while remaining practical for single-node GPU deployments. The model supports tool calling and eight languages including English, French, Spanish, and German. With a 128K context window and grouped-query attention, it quantizes efficiently down to Q4 levels for self-hosted inference on consumer hardware.
At Q4_K_M quantization (medium quality tier), the model weighs 39.6 GB. This fits within the 48 GB of VRAM on Scaleway L40S-1-48G, enabling full GPU inference.
The NVIDIA L40S is a datacenter GPU with 48 GB of GDDR6 VRAM and 864 GB/s memory bandwidth. It delivers 362 FP16 TFLOPS with Ada Lovelace architecture. A versatile GPU for AI inference, training, and graphics workloads. Handles quantized models up to 40B parameters comfortably.
Hardware Requirements
| Model size | 39.6 GB |
| VRAM available | 48 GB |
| VRAM used | 47.4 GB |
| System RAM | 96 GB |
| Min RAM required | 0 GB |
| GPU layers | 80 / 80 |
| Context size | 21,480 |
| Backend | cuda13 |
| Flash attention | Yes |
| Reading from disk | Yes |
Performance Notes
Deploy
Prerequisites
Ensure your GPU nodes are prepared with the NVIDIA container toolkit:
ansible-playbook prositronic.infra.nvidia_container_toolkitCommand
helmfile --state-values-file <(curl -s https://www.prositronic.eu/values/llama-3-3-70b-instruct/q4_k_m/nvidia-l40s.yaml) applyGenerated values.yaml
/values/llama-3-3-70b-instruct/q4_k_m/nvidia-l40s.yaml
Loading values…Frequently Asked Questions
How much VRAM does Llama 3.3 70B Instruct (Q4_K_M) need?
The Q4_K_M quantization of Llama 3.3 70B Instruct requires 39.6 GB. All 80 layers fit in the 48 GB of VRAM available on Scaleway L40S-1-48G, enabling full GPU acceleration.
Can I run Llama 3.3 70B Instruct on Scaleway L40S-1-48G?
Yes. Scaleway L40S-1-48G provides 48 GB of VRAM, which is enough to run Llama 3.3 70B Instruct (Q4_K_M) with all layers on the GPU for optimal performance.
What is quantization?
Quantization reduces a model's numerical precision from its original floating-point format to a more compact representation. This shrinks the file size and VRAM footprint, making it possible to run large models on consumer hardware. The trade-off is a small reduction in output quality. Q4_K_M compresses Llama 3.3 70B Instruct from its original size down to 39.6 GB.
What quantization should I choose for Llama 3.3 70B Instruct?
Q4_K_M is a medium-quality quantization. Higher-quality quants (Q8, Q6) preserve more model accuracy but need more VRAM. Lower quants (Q4, Q3, Q2) reduce VRAM usage at the cost of some quality. Choose based on your available hardware and quality requirements.
What is flash attention and why is it enabled?
Flash attention is a memory-efficient algorithm that speeds up the attention mechanism in transformer models. It reduces VRAM usage during inference by avoiding the materialisation of the full attention matrix. For Llama 3.3 70B Instruct on Scaleway L40S-1-48G, flash attention is enabled to maximise context length and throughput within the available 48 GB of VRAM.
How do I run Llama 3.3 70B Instruct (Q4_K_M) with Ollama?
Run ollama run llama3.3:70b-instruct-q4_k_m to start Llama 3.3 70B Instruct (Q4_K_M). Ollama handles downloading the model weights automatically on first run.