Rebrand: openTARS models are now made by MastLLM
Browse files
README.md
CHANGED
|
@@ -15,7 +15,7 @@ language:
|
|
| 15 |
|
| 16 |
# Murph
|
| 17 |
|
| 18 |
-
Murph is the small request classifier built into [openTARS](https://
|
| 19 |
|
| 20 |
It is not a language model. It is a softmax logistic regression over word pieces (2 to 5 letters), words and word pairs, with TF-IDF weights. It runs in plain Python on the CPU in about 0.2 ms per request, uses no GPU, and the whole model is one JSON file of about 4 MB.
|
| 21 |
|
|
@@ -50,7 +50,7 @@ The model also stores a confidence threshold and a coverage minimum, so the app
|
|
| 50 |
|
| 51 |
## Training and results
|
| 52 |
|
| 53 |
-
- Trained on 3,034 request phrases in five languages: examples from
|
| 54 |
- Cross-validation accuracy stored in the file: 88.9%.
|
| 55 |
- openTARS evaluation sets, measured on 18 October 2026: 170 of 175 across languages and 85 of 86 on the check set. The evaluation phrases were seen during development, so treat these as a sanity check, not as a benchmark.
|
| 56 |
|
|
|
|
| 15 |
|
| 16 |
# Murph
|
| 17 |
|
| 18 |
+
Murph is the small request classifier built into [openTARS](https://mastllm.github.io/OpenTARS/). For each request it decides the **kind of task**, so the app can pick the right model and tools before any language model runs.
|
| 19 |
|
| 20 |
It is not a language model. It is a softmax logistic regression over word pieces (2 to 5 letters), words and word pairs, with TF-IDF weights. It runs in plain Python on the CPU in about 0.2 ms per request, uses no GPU, and the whole model is one JSON file of about 4 MB.
|
| 21 |
|
|
|
|
| 50 |
|
| 51 |
## Training and results
|
| 52 |
|
| 53 |
+
- Trained on 3,034 request phrases in five languages: examples from MastLLM plus phrases written by a larger model. Phrases similar to the evaluation phrases were removed from training.
|
| 54 |
- Cross-validation accuracy stored in the file: 88.9%.
|
| 55 |
- openTARS evaluation sets, measured on 18 October 2026: 170 of 175 across languages and 85 of 86 on the check set. The evaluation phrases were seen during development, so treat these as a sanity check, not as a benchmark.
|
| 56 |
|