Add the DRY dynamic N-gram anti-repetition sampler (#982)

* Add the DRY dynamic N-gram anti-repetition sampler

The DRY (Do not Repeat Yourself) sampler is a dynamic N-gram
repetition penalty that negatively scores tokens that would extend
sequences that already appear in the context.

See this discussion for a motivation and explanation of the sampler:
oobabooga/text-generation-webui#5677

This implementation of DRY mostly aligns with the obabooga version
with a few modifications. It uses a more efficient linear scanning
algorithm to identify repetitions. It also supports multi-token
sequence breakers. As a limitation, this implementation reuses
the rep pen range parameter, rather than introducing a new range
just for the DRY sampler.

There is a separate change to lite.koboldai.net that exposes the DRY
sampler parameters to KoboldAI Lite, so none of the embed files have
been changed as part of this commit.

* Update default DRY parameters to match lite

* Improve DRY token debug logging

* Replace `and` with `&&` to fix MSVC compile error

Little known fact: The C++98 standard defines `and` as an
alternative token for the `&&` operator (along with a bunch
of other digraphs). MSVC does not allow these without using
the /Za option or including the <iso646.h> header. Change to
the more standard operator to make this code more portable.

* Fix MSVC compile error because log is not constexpr

Replace the compile-time computation with a floating-point
approximation of log(std::numeric_limits<float>::max()).

* Remove unused llama sampler variables and clean up sequence breakers.

* Remove KCPP_SAMPLER_DRY as a separate enum entry

The DRY sampler is effectively a repetition penalty and there
are very few reasons to apply it at a different place in sampler
order than the standard single-token penalty. There are also
multiple projects that have dependencies on the existing sampler
IDs, including KoboldAI, KoboldAI Lite, and Silly Tavern. In order
to minimize the impact of the dependencies of adding the DRY sampler
to koboldcpp, it makes the most sense to not add a new ID for now,
and instead to piggyback on KCPP_SAMPLER_REP_PEN. In the future
if we find a use case for splitting the application of rep pen and DRY
we can introduce a new enum entry then.

* Add the dry_penalty_last_n to independently control DRY penalty range

This parameter follows the oobabooga semantics: it's optional, with a
default value of zero. Zero means that DRY should sample the entire
context. Otherwise, it's the number of tokens from the end of the
context that are scanned for repetitions.

* Limit sequence breaker lengths in tokens and characters

The core DRY sampler algorithm is linear in the context length, but
there are several parts of the sampler related to multi-token
sequence breakers that are potentially quadratic. Without any
restrictions, a suitably crafted context and sequence breaker could
result in a denial-of-service attack on a server running koboldcpp.
This change limits the maximum number of characters and the maximum
token length of a sequence breaker in order to limit the maximum
overhead associated with the sampler.

This change also improves some comments, adding more detail and
changing the wording to increase clarity.

Author: pi6am

common/common.h

@@ -113,6 +113,11 @@ struct gpt_params {
     int32_t mirostat          = 0;     // 0 = disabled, 1 = mirostat, 2 = mirostat 2.0
     float   mirostat_tau      = 5.00f; // target entropy
     float   mirostat_eta      = 0.10f; // learning rate
+    float   dry_multiplier    = 0.0f;  // penalty multiplier, 0.0 = disabled
+    float   dry_base          = 1.75f; // exponential base
+    int32_t dry_allowed_length = 2;    // repeated sequences longer than this are penalized
+    int32_t dry_penalty_last_n = 0;    // how many tokens to scan for repetitions (0 = entire context)
+    std::vector<std::string> dry_sequence_breakers; // DRY sequence breakers
 
     // DynaTemp!
     float   dynatemp_range     = 0.0f;  // enables DynaTemp if greater than 0. dynatemp_min = temperature - dt_range, dynatemp_max = temperature + dt_range

expose.h

@@ -5,6 +5,7 @@ const int stop_token_max = 16;
 const int ban_token_max = 16;
 const int tensor_split_max = 16;
 const int logit_bias_max = 16;
+const int dry_seq_break_max = 16;
 const int images_max = 4;
 
 // match kobold's sampler list and order
@@ -83,6 +84,11 @@ struct generation_inputs
     const int mirostat = 0;
     const float mirostat_eta = 0.0f;
     const float mirostat_tau = 0.0f;
+    const float dry_multiplier = 0.0f;
+    const float dry_base = 0.0f;
+    const int dry_allowed_length = 0;
+    const int dry_penalty_last_n = 0;
+    const char * dry_sequence_breakers[dry_seq_break_max] = {};
     const samplers sampler_order[KCPP_SAMPLER_MAX] = {};
     const int sampler_len = 0;
     const bool allow_eos_token = false;