metadata
library_name: peft
Model 'quality'
As you can see from the example bellow the model is almost correctly respecting the learned format, but the grammatical correction explanations are wrong, it lack sense of correct German gramma.
Input
'### User: Check the gramma of the following sentence and list all error and relevant corrections. Sentence:Mir es geht gut, danke#### AI:Correct version of the sentence:
Output
"Mir geht es gut, danke.":
Repairs:
- Instead of "es geht" it should be "geht" - the verb "gehen" is in the third person singular.
- Instead of "Mir es" it should be "Mir geht" - the verb "gehen" is in the third person singular.#### End of the list of corrections.
Training procedure
The following bitsandbytes quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float16
Framework versions
- PEFT 0.4.0