mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2025-04-14 10:36:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs : bring llama-cli conversation/template docs up-to-date (#12426)

Browse Source
This commit is contained in:
Sigbjørn Skjæret 2025-03-17 21:14:32 +01:00 committed by GitHub
parent b1b132efcb
commit 60c902926c
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 36 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

41
examples/main/README.md
View File

@ -27,12 +27,24 @@ Once downloaded, place your model in the models folder in llama.cpp.
##### Input prompt (One-and-done)
```bash
./llama-cli -m models/gemma-1.1-7b-it.Q4_K_M.gguf --prompt "Once upon a time"
./llama-cli -m models/gemma-1.1-7b-it.Q4_K_M.gguf -no-cnv --prompt "Once upon a time"
```
##### Conversation mode (Allow for continuous interaction with the model)
```bash
./llama-cli -m models/gemma-1.1-7b-it.Q4_K_M.gguf -cnv --chat-template gemma
./llama-cli -m models/gemma-1.1-7b-it.Q4_K_M.gguf --chat-template gemma
```
##### Conversation mode using built-in jinja chat template
```bash
./llama-cli -m models/gemma-1.1-7b-it.Q4_K_M.gguf --jinja
```
##### One-and-done query using jinja with custom system prompt and a starting prompt
```bash
./llama-cli -m models/gemma-1.1-7b-it.Q4_K_M.gguf --jinja --single-turn -sys "You are a helpful assistant" -p "Hello"
```
##### Infinite text from a starting prompt (you can use `Ctrl-C` to stop it):
@ -44,12 +56,24 @@ Once downloaded, place your model in the models folder in llama.cpp.
##### Input prompt (One-and-done)
```powershell
./llama-cli.exe -m models\gemma-1.1-7b-it.Q4_K_M.gguf --prompt "Once upon a time"
./llama-cli.exe -m models\gemma-1.1-7b-it.Q4_K_M.gguf -no-cnv --prompt "Once upon a time"
```
##### Conversation mode (Allow for continuous interaction with the model)
```powershell
./llama-cli.exe -m models\gemma-1.1-7b-it.Q4_K_M.gguf -cnv --chat-template gemma
./llama-cli.exe -m models\gemma-1.1-7b-it.Q4_K_M.gguf --chat-template gemma
```
##### Conversation mode using built-in jinja chat template
```powershell
./llama-cli.exe -m models\gemma-1.1-7b-it.Q4_K_M.gguf --jinja
```
##### One-and-done query using jinja with custom system prompt and a starting prompt
```powershell
./llama-cli.exe -m models\gemma-1.1-7b-it.Q4_K_M.gguf --jinja --single-turn -sys "You are a helpful assistant" -p "Hello"
```
#### Infinite text from a starting prompt (you can use `Ctrl-C` to stop it):
@ -77,6 +101,8 @@ The `llama-cli` program provides several ways to interact with the LLaMA models
- `--prompt PROMPT`: Provide a prompt directly as a command-line option.
- `--file FNAME`: Provide a file containing a prompt or multiple prompts.
- `--system-prompt PROMPT`: Provide a system prompt (will otherwise use the default one in the chat template (if provided)).
- `--system-prompt-file FNAME`: Provide a file containing a system prompt.
- `--interactive-first`: Run the program in interactive mode and wait for input right away. (More on this below.)
## Interaction
@ -89,7 +115,10 @@ In interactive mode, users can participate in text generation by injecting their
- `-i, --interactive`: Run the program in interactive mode, allowing users to engage in real-time conversations or provide specific instructions to the model.
- `--interactive-first`: Run the program in interactive mode and immediately wait for user input before starting the text generation.
- `-cnv, --conversation`: Run the program in conversation mode (does not print special tokens and suffix/prefix, use default chat template) (default: false)
- `-cnv, --conversation`: Run the program in conversation mode (does not print special tokens and suffix/prefix, use default or provided chat template) (default: true if chat template found)
- `-no-cnv`: Disable conversation mode (default: false)
- `-st, --single-turn`: Only process a single conversation turn (user input) and then exit.
- `--jinja`: Enable jinja chat template parser, will use the model's built-in template or a user-provided one (default: false)
- `--color`: Enable colorized output to differentiate visually distinguishing between prompts, user input, and generated text.
By understanding and utilizing these interaction options, you can create engaging and dynamic experiences with the LLaMA models, tailoring the text generation process to your specific needs.
@ -125,6 +154,8 @@ When --in-prefix or --in-suffix options are enabled the chat template ( --chat-t
Example usage: `--chat-template gemma`
`--chat-template-file FNAME`: Load a custom jinja chat template from an external file, useful if the model contains outdated or incompatible template, some examples can be found in models/templates. Up-to-date chat templates can be downloaded from Hugging Face using scripts/get_chat_template.py
## Context Management
During text generation, LLaMA models have a limited context size, which means they can only consider a certain number of tokens from the input and generated text. When the context fills up, the model resets internally, potentially losing some information from the beginning of the conversation or instructions. Context management options help maintain continuity and coherence in these situations.