Open-weight model
opt-babylm-350m-20eps-prenorm-seed964
by Zachary Houghton znhoughton/opt-babylm-350m-20eps-prenorm-seed964
opt-babylm-350m-20eps-prenorm-seed964 is an open-weight model from Zachary Houghton. It has 312M parameters and a 1,024-token context. At 16-bit it needs about 0.7 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
Runs On
What it takes to serve opt-babylm-350m-20eps-prenorm-seed964 (312M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 0.6 GB | 0.7 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.3 GB | 0.4 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.2 GB | 0.2 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Oct 18, 2026.
Model Card
The publisher has not written a card for this model.
Configuration
- Architecture
- OPTForCausalLM
- Context length (tokens)
- 1,024
- Layers
- 24
- Hidden size
- 1,024
- Attention heads
- 16
- Vocabulary size
- 8,192
- Stored precision
- float32
- Model type
- opt
Identity and Version
- Repository
- znhoughton/opt-babylm-350m-20eps-prenorm-seed964
- Publisher
- Zachary Houghton
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 312M parameters
- Languages
- opt
- Revision
- c1bbe76e876094fa128f33db16b48af2214e5ab6
- First published
- 2026-09-21
- Last updated
- 2026-09-22
Files and Weights
23 files, 2.5 GB in total. The weights are 4 files totalling 2.5 GB in bin, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| last-checkpoint/model.safetensors | Weights | 1.2 GB | 65e53b2c15c8 |
| last-checkpoint/training_args.bin | Weights | 5.4 KB | 5c00403bc9c8 |
| model.safetensors | Weights | 1.2 GB | 65e53b2c15c8 |
| training_args.bin | Weights | 5.4 KB | 5c00403bc9c8 |
| config.json | Configuration | 766 B | — |
| last-checkpoint/config.json | Configuration | 766 B | — |
| last-checkpoint/generation_config.json | Configuration | 132 B | — |
| last-checkpoint/special_tokens_map.json | Configuration | 552 B | — |
| last-checkpoint/trainer_state.json | Configuration | 61.7 KB | — |
| special_tokens_map.json | Configuration | 552 B | — |
| README.md | Documentation | 20 B | — |
| runs/Sep21_23-58-17_af28f3b40d15/events.out.tfevents.1790035808.af28f3b40d15.2120.0 | Other | 5.3 KB | 386b81b36825 |
| runs/Sep22_00-13-53_af28f3b40d15/events.out.tfevents.1790036045.af28f3b40d15.2833.0 | Other | 5.3 KB | a6ca9437af88 |
| runs/Sep22_00-17-49_af28f3b40d15/events.out.tfevents.1790036280.af28f3b40d15.3169.0 | Other | 85.5 KB | c5ef2a46a74d |
| .gitattributes | Repository | 1.5 KB | — |
| last-checkpoint/merges.txt | Tokenizer | 65.8 KB | — |
| last-checkpoint/tokenizer.json | Tokenizer | 552.9 KB | — |
| last-checkpoint/tokenizer_config.json | Tokenizer | 1.1 KB | — |
| last-checkpoint/vocab.json | Tokenizer | 114.4 KB | — |
| merges.txt | Tokenizer | 65.8 KB | — |
| tokenizer.json | Tokenizer | 552.9 KB | — |
| tokenizer_config.json | Tokenizer | 1.1 KB | — |
| vocab.json | Tokenizer | 114.4 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 2.5 GB
Released by Zachary Houghton through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 2.5 GB |
| 16-bit | 0.6 GB |
| 8-bit | 0.3 GB |
| 4-bit | 0.2 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About opt-babylm-350m-20eps-prenorm-seed964
How much GPU memory does opt-babylm-350m-20eps-prenorm-seed964 need?
About 0.7 GB at 16-bit and 0.2 GB at 4-bit: the weights (312M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run opt-babylm-350m-20eps-prenorm-seed964 on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
What is opt-babylm-350m-20eps-prenorm-seed964's context length?
1,024 tokens, from the maximum position embeddings in its published configuration.