README example shows incorrect model path and command line options after recent changes

[7shi]

The "Basic usage" section in the README contains several inconsistencies that were introduced when PR #137 was merged. Additionally, a related issue #138 has been identified regarding the command line argument handling.

The example in the README currently shows:

# Run inference with the quantized model
python run_inference.py -m models/Falcon3-7B-Instruct-1.58bit/ggml-model-i2_s.gguf -cnv "You are a helpful assistant"

# Output:
# Daniel went back to the the the garden. Mary travelled to the kitchen. Sandra journeyed to the kitchen. Sandra went to the hallway. John went to the bedroom. Mary went back to the garden. Where is Mary?
# Answer: Mary is in the garden.

This example has several problems:

1. The model path is incorrect. This appears to be caused by not distinguishing Instruct variants when the model was added to setup_env.py.

   • Current: models/Falcon3-7B-Instruct-1.58bit/ggml-model-i2_s.gguf
   • Actual: models/Falcon3-7B-1.58bit/ggml-model-i2_s.gguf

2. The command line option -cnv is incorrect and should be -p. However, this is pending the fix for issue New "-cnv" option cannot work correctly!! #138.

3. The example prompt and output do not match:

   • Prompt: You are a helpful assistant
   • Output: Mary is in the garden.

Proposed Changes:

Update the example to correctly demonstrate the command that generates the shown output:

python run_inference.py -m models/Falcon3-7B-1.58bit/ggml-model-i2_s.gguf -p "Daniel went back to the the the garden. Mary travelled to the kitchen. Sandra journeyed to the kitchen. Sandra went to the hallway. John went to the bedroom. Mary went back to the garden. Where is Mary?\nAnswer:" -n 6 -temp 0

This update:

1. Uses the correct model path that matches what setup_env.py actually generates
2. Uses the proper command line option (-p)

Note that this command will not work correctly until issue #138 is resolved, as there are currently problems with the command line argument handling. Additionally, instead of modifying the model path in the README, it might be better to update setup_env.py to properly distinguish between Base and Instruct model variants.

[potassiummmm]

Thanks for pointing this out! Should be fixed in #141.