// Yi (01.AI) · Text LLM
Yi-1.5-34B-Chat-16K: VRAM & local deployment
34.4B parameters. About 25.4 GB at 4-bit and 78.2 GB at BF16 with one 8K-token request.
released 2024-05-15licence: Commercial OKmemory figures are estimates
01Specs
| Publisher | 01.AI · 01-ai |
|---|---|
| Family / series | Yi (01.AI) · Yi-1.5-16K |
| Type | Text LLM (text-generation) |
| Parameters | 34.4B Source: counted from official safetensors weights |
| Context length | 16K tokens (16,384)max_position_embeddings in config.json |
| Architecture | 60 layers · GQA |
| Released weights | BF16 · 68.8 GBdownload |
| Licence | Apache-2.0 · Commercial OKPermissive licence (Apache-2.0 / MIT / CC-BY / OpenMDW): commercial use allowed. |
| Released | 2024-05-15 (Hugging Face repo created) |
02How much VRAM it needs (estimate)
Total = weights + 10% + 0.5GB overhead + KV cache for one request (FP16 KV). More concurrent users multiply the KV column.
| Precision | Weights | Total @ 8K | Smallest GPU | Mac |
|---|---|---|---|---|
| BF16 / FP16 (original) | 68.8 GB | 78.2 GB | RTX PRO 6000 96GB | Mac 128GB |
| FP8 / INT8 / Q8_0 | 36.5 GB | 42.7 GB | 2 × RTX 5090 (64GB) | Mac 64GB |
| 4-bit (Q4_K_M / AWQ / GPTQ-Int4) | 20.9 GB | 25.4 GB | RTX 5090 32GB | Mac 64GB |
KV cache per request: 8K 2.0 GB
Bits per weight: BF16 16, FP8 8.5, 4-bit 4.85. “Fits” = total ≤ 90% of memory; Mac GPU can use ~2/3 of unified memory (≤36GB) or ~3/4 (larger). For MoE, all experts must be in memory, so size by total parameters. Method →
03How to run it
Only runtimes the publisher or the runtime’s official library names for this model.
- Transformers — library declared on Hugging Face
04Official repositories & quantised builds
1 official repo(s) from 01-ai.
- Yi-1.5-34B-Chat-16K Instruct
05Other Yi (01.AI) models
All 9 →| Model | Params | 4-bit total | Released |
|---|---|---|---|
| Yi-1.5-9B-Chat-16K · same series | 8.83B | 7.2 GB | 2024-05-15 |
| Yi-Coder-1.5B-Chat | 1.48B | 3.1 GB | 2024-08-21 |
| Yi-Coder-9B-Chat | 8.83B | 7.2 GB | 2024-08-21 |
| Yi-1.5-9B-32K | 8.83B | 7.2 GB | 2024-05-15 |
| Yi-1.5-34B-32K | 34.4B | 25.4 GB | 2024-05-15 |
| Yi-1.5-6B | 6.06B | 5.1 GB | 2024-05-11 |
| Yi-1.5-9B-Chat | 8.83B | 7.2 GB | 2024-05-10 |
| Yi-1.5-34B-Chat | 34.4B | 25.4 GB | 2024-05-10 |
06Sources
Data checked 2024-06-26. Parameters, context and licence are copied from the official repository; memory figures are our estimates.