Open-weight model
FunctionGemma-HermesAgent-TitleGenerator
by Matthew americansquid/FunctionGemma-HermesAgent-TitleGenerator
FunctionGemma-HermesAgent-TitleGenerator is an open-weight model from Matthew. It has 268M parameters and a 32,768-token context. At 16-bit it needs about 0.6 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
A fine-tuned google/functiongemma-270m-it model for naming chat sessions from a user's opening message. It produces a setsessiontitle function call containing a concise title, rather than answering the message.
Runs On
What it takes to serve FunctionGemma-HermesAgent-TitleGenerator (268M 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 | 0.5 GB | 0.6 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.3 GB | 0.3 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.1 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 19, 2026.
Model Card
A fine-tuned google/functiongemma-270m-it model for naming chat sessions from a user's opening message. It produces a setsessiontitle function call containing a concise title, rather than answering the message. The saved model is in functiongemmatitlemodel/, including full model weights, tokenizer, generation configuration, and chat template. - Generate a title of 3–7 words in sentence case. - Describe the user's task or conversation topic. - Preserve technical terms. - Return only a setsessiontitle tool call. Example output from the saved tuned evaluation: This is FunctionGemma's tool-call syntax, not JSON. The application must parse and validate the call before using the title. Install…
Excerpt from the card by Matthew.
Configuration
- Architecture
- Gemma3ForCausalLM
- Context length (tokens)
- 32,768
- Layers
- 18
- Hidden size
- 640
- Feed-forward size
- 2,048
- Attention heads
- 4
- Key/value heads
- 1
- Head dimension
- 256
- Vocabulary size
- 262,144
- Sliding window (tokens)
- 512
- Model type
- gemma3_text
Identity and Version
- Repository
- americansquid/FunctionGemma-HermesAgent-TitleGenerator
- Publisher
- Matthew
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 268M parameters
- Languages
- Not stated by the source
- Revision
- 37e7b415def6db657e5df63f345c03f4ddadf31b
- First published
- 2026-10-07
- Last updated
- 2026-10-07
Files and Weights
46 files, 3.9 GB in total. The weights are 12 files totalling 3.8 GB in bin, pt, pth, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| checkpoint-626/model.safetensors | Weights | 536.2 MB | d8413cb0de92 |
| checkpoint-626/optimizer.pt | Weights | 1.1 GB | f6d08f867247 |
| checkpoint-626/rng_state.pth | Weights | 14.6 KB | a8e2011629d8 |
| checkpoint-626/scheduler.pt | Weights | 1.5 KB | e0017cf91356 |
| checkpoint-626/training_args.bin | Weights | 5.3 KB | e459f7c39469 |
| checkpoint-939/model.safetensors | Weights | 536.2 MB | 904e1def1845 |
| checkpoint-939/optimizer.pt | Weights | 1.1 GB | cc264f636575 |
| checkpoint-939/rng_state.pth | Weights | 14.6 KB | 01f9a0f7843a |
| checkpoint-939/scheduler.pt | Weights | 1.5 KB | 16d01477a374 |
| checkpoint-939/training_args.bin | Weights | 5.3 KB | e459f7c39469 |
| model.safetensors | Weights | 536.2 MB | 904e1def1845 |
| training_args.bin | Weights | 5.3 KB | e459f7c39469 |
| all_results.json | Configuration | 243 B | — |
| baseline_test_metrics.json | Configuration | 123 B | — |
| checkpoint-626/config.json | Configuration | 1.5 KB | — |
| checkpoint-626/generation_config.json | Configuration | 218 B | — |
| checkpoint-626/trainer_state.json | Configuration | 21.6 KB | — |
| checkpoint-939/config.json | Configuration | 1.5 KB | — |
| checkpoint-939/generation_config.json | Configuration | 218 B | — |
| checkpoint-939/trainer_state.json | Configuration | 31.9 KB | — |
| config.json | Configuration | 1.5 KB | — |
| eval_results.json | Configuration | 40 B | — |
| generation_config.json | Configuration | 209 B | — |
| run_manifest.json | Configuration | 1.2 KB | — |
| train_results.json | Configuration | 205 B | — |
| trainer_state.json | Configuration | 32.3 KB | — |
| tuned_test_metrics.json | Configuration | 122 B | — |
| README.md | Documentation | 6.9 KB | — |
| baseline_test_predictions.jsonl | Other | 8.6 KB | — |
| chat_template.jinja | Other | 13.8 KB | — |
| checkpoint-626/chat_template.jinja | Other | 13.8 KB | — |
| checkpoint-939/chat_template.jinja | Other | 13.8 KB | — |
| dataset.jsonl | Other | 37.7 MB | 4be5969ae93e |
| hermes_prompt.txt | Other | 943 B | — |
| model_usage.txt | Other | 282 B | — |
| test.jsonl | Other | 3.8 MB | — |
| train.jsonl | Other | 30.1 MB | c6d7c76d6c4a |
| tuned_test_predictions.jsonl | Other | 7.5 KB | — |
| validation.jsonl | Other | 3.8 MB | — |
| .gitattributes | Repository | 1.7 KB | — |
| checkpoint-626/tokenizer.json | Tokenizer | 33.4 MB | 80d7f800b949 |
| checkpoint-626/tokenizer_config.json | Tokenizer | 822 B | — |
| checkpoint-939/tokenizer.json | Tokenizer | 33.4 MB | 80d7f800b949 |
| checkpoint-939/tokenizer_config.json | Tokenizer | 822 B | — |
| tokenizer.json | Tokenizer | 33.4 MB | 80d7f800b949 |
| tokenizer_config.json | Tokenizer | 822 B | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 3.8 GB
Released by Matthew through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 3.8 GB |
| 16-bit | 0.5 GB |
| 8-bit | 0.3 GB |
| 4-bit | 0.1 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About FunctionGemma-HermesAgent-TitleGenerator
How much GPU memory does FunctionGemma-HermesAgent-TitleGenerator need?
About 0.6 GB at 16-bit and 0.2 GB at 4-bit: the weights (268M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run FunctionGemma-HermesAgent-TitleGenerator 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 FunctionGemma-HermesAgent-TitleGenerator's context length?
32,768 tokens, from the maximum position embeddings in its published configuration.