Llama33 70b awq draft llama32 1b
SPD Overview¶
This SPD defines a dual-model configuration for speculative decoding, where the Target Model ensures high-quality multilingual responses and the Draft Model accelerates inference.
Target Model – llama-3.3-70b-instruct-awq¶
This is the AWQ version of the Llama 3.3 70B Instruct model. The Meta Llama 3.3 multilingual large language model (LLM) is an instruction tuned generative model in 70B (text in/text out). The Llama 3.3 instruction tuned text only model is optimized for multilingual dialogue use cases and outperforms many of the available open source and closed chat models on common industry benchmarks.
- Model Architecture: Llama 3.3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.
- Model Release Date: December 6, 2024.
- Model Source: casperhansen/llama-3.3-70b-instruct-awq
- Model Repository: casper-hansen/AutoAWQ
- License: llama3.3
- Supported languages: English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai.
QPC Configurations¶
| Precision | SoCs / Tensor slicing | NSP-Cores (per SoC) | Full Batch Size | Chunking Prompt Length | Context Length (CL) | Generated URL | Download | Generation Date |
|---|---|---|---|---|---|---|---|---|
| MXFP6 | 2 | 8 | 1 | 128 | 4096 | https://dc00tk1pxen80.cloudfront.net/SDK1.20.4/casperhansen/llama-3.3-70b-instruct-awq/llama-3.3-70b-instruct-awq_qpc_8cores_128pl_4096cl_1fbs_2devices_11spd_mxfp6_mxint8_target.tar.gz | Download | 27-Jan-2026 |
Draft Model – Llama 3.2 1B Instruct¶
The Meta Llama 3.2-1B Instruct is a smaller, instruction-tuned model used here as the Draft Model for speculative decoding. It provides fast draft token generation to reduce latency when paired with the Target 70B model.
- Model Architecture: Lightweight auto-regressive transformer.
- Release Date: September 2024.
- Repository: llama3_2
- Model Source: meta-llama/Llama-3.2-1B-Instruct
- License:Llama 3.2 Community License
- Supported Languages: English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai.
QPC Configurations¶
| Precision | SoCs / Tensor slicing | NSP-Cores (per SoC) | Full Batch Size | Chunking Prompt Length | Context Length (CL) | Generated URL | Download | Generation Date |
|---|---|---|---|---|---|---|---|---|
| MXFP6 | 2 | 8 | 1 | 128 | 4096 | https://dc00tk1pxen80.cloudfront.net/SDK1.20.4/meta-llama/Llama-3.2-1B-Instruct/Llama-3.2-1B-Instruct_qpc_8cores_128pl_4096cl_1fbs_2devices_11spd_mxfp6_mxint8_draft.tar.gz | Download | 27-Jan-2026 |