Hugging Face's logo

merabro
/
Apertus-ParlTopics-8B

Text Classification
PEFT
Safetensors
German
French
Italian
lora
decision-model
swiss-parliament
Model card Files
xet
 Community
merabro commited on
Commit
0f6e7c1
·
verified ·
1 Parent(s): 8a7a8bc

Model card: naming

Browse files
Files changed (1) hide show
  1. README.md +2 -4
README.md CHANGED
@@ -19,8 +19,6 @@ Apertus-ParlTopics-8B is a LoRA adapter and a yes/no head for the text part of A
19
 
20
  Hugging Face: https://huggingface.co/merabro/Apertus-ParlTopics-8B
21
 
22
- Internal label: training run G. Folder and file names in the repository keep that label (`docs/model/run_G/`, `models/runG/`, `calibration_*_runG.json`, run id `runG_corpus`).
23
-
24
  The Hugging Face repository holds the adapter, the head, the adapter and training configurations and the training metrics. It does not include the Apertus tokenizer or weights: accept the terms of [swiss-ai/Apertus-v1.5-8B](https://huggingface.co/swiss-ai/Apertus-v1.5-8B), convert its text part with `python -m parltopics.convert_text`, and copy its tokenizer files into the folder before serving.
25
 
26
  ## Files and checksums
@@ -30,7 +28,7 @@ The Hugging Face repository holds the adapter, the head, the adapter and trainin
30
  | `adapter_model.safetensors` | 159 MB | `61df7d434bc3662f6dadea5eb63ea213a617b343e1d06fcaad04b39894d975a0` |
31
  | `head.pt` | 8.4 MB | `cb93496bcf2ada57f00e133944b4b4eac5e8daa3b6d9908d616d3ac8eda0747f` |
32
 
33
- The checksums were computed on 19 October 2026 on the local copy (`models/runG/`) and on the copy that the running server loads (`kev_runs/G_desc/` on the GPU server); both match. The adapter configuration, the training configuration and the training metrics are in this folder (`adapter_config.json`, `training_config.json`, `training_metrics.json`).
34
 
35
  ## Model
36
 
@@ -124,7 +122,7 @@ The frontier-model judge is Claude Opus 5.5 (Claude Sonnet 5.5 for the digitalis
124
  Kev's server loads the base model, the adapter and the head and answers `POST /v1/systemone`. On our GPU server:
125
 
126
  ```bash
127
- scripts/hydra/serve_runG.sh 1 8350
128
  ```
129
 
130
  which runs, inside a Kev checkout at `1f04cfa` with `apertus.patch` applied,
 
19
 
20
  Hugging Face: https://huggingface.co/merabro/Apertus-ParlTopics-8B
21
 
 
 
22
  The Hugging Face repository holds the adapter, the head, the adapter and training configurations and the training metrics. It does not include the Apertus tokenizer or weights: accept the terms of [swiss-ai/Apertus-v1.5-8B](https://huggingface.co/swiss-ai/Apertus-v1.5-8B), convert its text part with `python -m parltopics.convert_text`, and copy its tokenizer files into the folder before serving.
23
 
24
  ## Files and checksums
 
28
  | `adapter_model.safetensors` | 159 MB | `61df7d434bc3662f6dadea5eb63ea213a617b343e1d06fcaad04b39894d975a0` |
29
  | `head.pt` | 8.4 MB | `cb93496bcf2ada57f00e133944b4b4eac5e8daa3b6d9908d616d3ac8eda0747f` |
30
 
31
+ The checksums were computed on 19 October 2026 on the local copy and on the copy that the running server loads (the GPU server's run folder on the GPU server); both match. The adapter configuration, the training configuration and the training metrics are in this folder (`adapter_config.json`, `training_config.json`, `training_metrics.json`).
32
 
33
  ## Model
34
 
 
122
  Kev's server loads the base model, the adapter and the head and answers `POST /v1/systemone`. On our GPU server:
123
 
124
  ```bash
125
+ scripts/hydra/serve_model.sh 1 8350
126
  ```
127
 
128
  which runs, inside a Kev checkout at `1f04cfa` with `apertus.patch` applied,