diff options
| author | Kawrakow <iwankawrakow@gmail.com> | 2025-06-03 17:35:09 +0300 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-06-03 17:35:09 +0300 |
| commit | f6d5fbdc5780b6dca770c896b8463de3239c7f8b (patch) | |
| tree | 5174cb76b596d23383a4434ab2179d8b5213512f /examples | |
| parent | ccb265c01676aad9ae5860ba50e74e61dfcd1cf8 (diff) | |
Adding top-n-sigma sampler (#489)
* Adding top-n-sigma sampler
* Fix typos in XTC PR
* Update README.md for main and server
* More README
* More README
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Diffstat (limited to 'examples')
| -rw-r--r-- | examples/main/README.md | 16 | ||||
| -rw-r--r-- | examples/server/README.md | 3 |
2 files changed, 19 insertions, 0 deletions
diff --git a/examples/main/README.md b/examples/main/README.md index 9396a34f..dd627456 100644 --- a/examples/main/README.md +++ b/examples/main/README.md @@ -239,6 +239,22 @@ The `--mirostat-ent` option sets the Mirostat target entropy (tau), which repres Example usage: `--mirostat 2 --mirostat-lr 0.05 --mirostat-ent 3.0` +### XTC Sampling (Exclude Top Choices) + +The function of this sampler is conrolled by `--xtc-probability` and `--xtc-threshold`. `--xtc-probability` takes values between +0 and 1 (<=0 turns this sampler off) and defines the probability for randomly invoking the sampler. `--xtc-threshold` +defines the token probability threshold. Tokens with probability greater than this threshold will be excluded from the sampling. +The sampler is turned off for `threshold > 0.5`. + +- --xtc-probability p: xtc probability (default: 0.0 => disabled) +- --xtc-threshold t : xtc threshold (default: 1.0 => disabled) + +### Top-n-sigma Sampling + +Sets all logits $L_i$ to $-\infty$ where $L_i < L_{\rm max} - n \sigma$. Here $L_{\rm max}$ is the maximum logit, $\sigma$ is the logit standard deviation, and $n$ (a floating point number) is the top-n-sigma parameter. Increasing $n$ increases the fraction of tokens considered for sampling. In the limit of $n$ close to zero, one effectively gets greedy sampling (only top probability token considered). + +- --top-n-sigma t top-n-sigma parmeter (default: 0.0 => disabled) + ### Logit Bias - `-l TOKEN_ID(+/-)BIAS, --logit-bias TOKEN_ID(+/-)BIAS`: Modify the likelihood of a token appearing in the generated text completion. diff --git a/examples/server/README.md b/examples/server/README.md index cb1eb7c9..c5a4b3e7 100644 --- a/examples/server/README.md +++ b/examples/server/README.md @@ -98,6 +98,9 @@ sampling: (default: 0, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0) --mirostat-lr N Mirostat learning rate, parameter eta (default: 0.1) --mirostat-ent N Mirostat target entropy, parameter tau (default: 5.0) + --xtc-probability p xtc probability (default: 0.0 => disabled) + --xtc-threshold t xtc threshold (default: 1.0 => disabled) + --top-n-sigma t top-n-sigma parmeter (default: 0.0 => disabled) -l TOKEN_ID(+/-)BIAS modifies the likelihood of token appearing in the completion, i.e. `--logit-bias 15043+1` to increase likelihood of token ' Hello', or `--logit-bias 15043-1` to decrease likelihood of token ' Hello' |