Naphula commited on
Commit
fc873db
·
verified ·
1 Parent(s): 8c4f85f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +6 -0
README.md CHANGED
@@ -32,6 +32,12 @@ Tools to enhance LLM quantizations and merging
32
  # [metadata_audit.py](https://huggingface.co/spaces/Naphula/model_tools/blob/main/metadata_audit.py)
33
  - Checks multiple models within subdirectories for vocab or rope mismatch (useful for large merges). Calibrated for Mistral Nemo 12B by default.
34
 
 
 
 
 
 
 
35
  # [fp32_to_bf16.py](https://huggingface.co/spaces/Naphula/model_tools/blob/main/fp32_to_bf16.py)
36
  - Converts FP32 to BF16 safetensors
37
 
 
32
  # [metadata_audit.py](https://huggingface.co/spaces/Naphula/model_tools/blob/main/metadata_audit.py)
33
  - Checks multiple models within subdirectories for vocab or rope mismatch (useful for large merges). Calibrated for Mistral Nemo 12B by default.
34
 
35
+ # [eos_scanner.py](https://huggingface.co/spaces/Naphula/model_tools/blob/main/eos_scanner.py)
36
+ - This tool scans the tokenizer jsons to detect any mismatches with EOS tokens, which cause early termination bugs. You can then use the [gen_id_patcher.py](https://huggingface.co/spaces/Naphula/model_tools/blob/main/gen_id_patcher.py) to patch missing `generation_config.json` files for EOS token. See [this post](https://huggingface.co/Naphula/Q0_Bench/discussions/1?not-for-all-audiences=true#6987717c762f0a45f672e250) for more info.
37
+
38
+ # [weight_counter.py](https://huggingface.co/spaces/Naphula/model_tools/blob/main/weight_counter.py)
39
+ - This counts the number of models in a yaml and adds up the total weight values. Useful for large della/ties merges.
40
+
41
  # [fp32_to_bf16.py](https://huggingface.co/spaces/Naphula/model_tools/blob/main/fp32_to_bf16.py)
42
  - Converts FP32 to BF16 safetensors
43