Open-weight model
mt5-thai-post-correction-15epoch
by Jiratchaya natttt07/mt5-thai-post-correction-15epoch
mt5-thai-post-correction-15epoch is an open-weight model from Jiratchaya. It has 582M parameters. At 16-bit it needs about 1.4 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 mt5-thai-post-correction-15epoch (582M 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 | 1.2 GB | 1.4 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.6 GB | 0.7 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.3 GB | 0.3 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.
mt5-thai-post-correction-15epoch on every accelerator the SAVRN Index prices, at every precision
Model Card
The publisher has not written a card for this model.
Configuration
- Architecture
- MT5ForConditionalGeneration
- Vocabulary size
- 250,112
- Model type
- mt5
Identity and Version
- Repository
- natttt07/mt5-thai-post-correction-15epoch
- Publisher
- Jiratchaya
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 582M parameters
- Languages
- Not stated by the source
- Revision
- 002110426fc97b7794be2cb64a3c0fd54b968b2d
- First published
- 2026-10-10
- Last updated
- 2026-10-10
Files and Weights
27 files, 16.4 GB in total. The weights are 12 files totalling 16.3 GB in bin, pt, pth, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| checkpoint-52679/model.safetensors | Weights | 2.3 GB | 0108d61d8075 |
| checkpoint-52679/optimizer.pt | Weights | 4.7 GB | 753c6de8b166 |
| checkpoint-52679/rng_state.pth | Weights | 14.6 KB | 400bc8dd1bd4 |
| checkpoint-52679/scheduler.pt | Weights | 1.5 KB | 3f8fef5fd15f |
| checkpoint-52679/training_args.bin | Weights | 5.4 KB | add438334bd4 |
| checkpoint-71835/model.safetensors | Weights | 2.3 GB | 864beb4cb38e |
| checkpoint-71835/optimizer.pt | Weights | 4.7 GB | 5172e2cff8b6 |
| checkpoint-71835/rng_state.pth | Weights | 14.6 KB | a9e5c86479ac |
| checkpoint-71835/scheduler.pt | Weights | 1.5 KB | a0fe7ca432a5 |
| checkpoint-71835/training_args.bin | Weights | 5.4 KB | add438334bd4 |
| model.safetensors | Weights | 2.3 GB | 0108d61d8075 |
| training_args.bin | Weights | 5.4 KB | add438334bd4 |
| checkpoint-52679/config.json | Configuration | 830 B | — |
| checkpoint-52679/generation_config.json | Configuration | 152 B | — |
| checkpoint-52679/trainer_state.json | Configuration | 101.1 KB | — |
| checkpoint-71835/config.json | Configuration | 830 B | — |
| checkpoint-71835/generation_config.json | Configuration | 152 B | — |
| checkpoint-71835/trainer_state.json | Configuration | 137.9 KB | — |
| config.json | Configuration | 830 B | — |
| generation_config.json | Configuration | 152 B | — |
| .gitattributes | Repository | 1.7 KB | — |
| checkpoint-52679/tokenizer.json | Tokenizer | 16.3 MB | 8cf2a14a94ce |
| checkpoint-52679/tokenizer_config.json | Tokenizer | 305 B | — |
| checkpoint-71835/tokenizer.json | Tokenizer | 16.3 MB | 8cf2a14a94ce |
| checkpoint-71835/tokenizer_config.json | Tokenizer | 305 B | — |
| tokenizer.json | Tokenizer | 16.3 MB | 8cf2a14a94ce |
| tokenizer_config.json | Tokenizer | 305 B | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 16.3 GB
Released by Jiratchaya through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 16.3 GB |
| 16-bit | 1.2 GB |
| 8-bit | 0.6 GB |
| 4-bit | 0.3 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About mt5-thai-post-correction-15epoch
How much GPU memory does mt5-thai-post-correction-15epoch need?
About 1.4 GB at 16-bit and 0.3 GB at 4-bit: the weights (582M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run mt5-thai-post-correction-15epoch 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.