Fix check warnings

Author: ysimonson

examples/reranker/src/main.rs

@@ -12,7 +12,6 @@ use std::time::Duration;
 
 use anyhow::{bail, Context, Result};
 use clap::Parser;
-use hf_hub::api::sync::ApiBuilder;
 
 use llama_cpp_2::context::params::{LlamaContextParams, LlamaPoolingType};
 use llama_cpp_2::context::LlamaContext;
@@ -156,7 +155,6 @@ fn main() -> Result<()> {
     // } else {
     //     tokens_lines_list.len()
     // };
-    let mut embeddings_stored = 0;
     let mut max_seq_id_batch = 0;
     let mut output = Vec::with_capacity(tokens_lines_list.len());
 
@@ -169,16 +167,10 @@ fn main() -> Result<()> {
                 &mut ctx,
                 &mut batch,
                 max_seq_id_batch,
-                n_embd,
                 &mut output,
                 normalise,
                 pooling.clone(),
             )?;
-            embeddings_stored += if pooling == "none" {
-                batch.n_tokens()
-            } else {
-                max_seq_id_batch
-            };
             max_seq_id_batch = 0;
             batch.clear();
         }
@@ -191,7 +183,6 @@ fn main() -> Result<()> {
         &mut ctx,
         &mut batch,
         max_seq_id_batch,
-        n_embd,
         &mut output,
         normalise,
         pooling.clone(),
@@ -243,7 +234,6 @@ fn batch_decode(
     ctx: &mut LlamaContext,
     batch: &mut LlamaBatch,
     s_batch: i32,
-    n_embd: i32,
     output: &mut Vec<Vec<f32>>,
     normalise: bool,
     pooling: String,

llama-cpp-2/src/model.rs

@@ -666,11 +666,11 @@ impl LlamaModel {
     /// There is many ways this can fail. See [`LlamaContextLoadError`] for more information.
     // we intentionally do not derive Copy on `LlamaContextParams` to allow llama.cpp to change the type to be non-trivially copyable.
     #[allow(clippy::needless_pass_by_value)]
-    pub fn new_context(
-        &self,
+    pub fn new_context<'a>(
+        &'a self,
         _: &LlamaBackend,
         params: LlamaContextParams,
-    ) -> Result<LlamaContext, LlamaContextLoadError> {
+    ) -> Result<LlamaContext<'a>, LlamaContextLoadError> {
         let context_params = params.context_params;
         let context = unsafe {
             llama_cpp_sys_2::llama_new_context_with_model(self.model.as_ptr(), context_params)

llama-cpp-2/src/model/params.rs

@@ -149,7 +149,7 @@ impl LlamaModelParams {
     /// assert_eq!(count, 0);
     /// ```
     #[must_use]
-    pub fn kv_overrides(&self) -> KvOverrides {
+    pub fn kv_overrides<'a>(&'a self) -> KvOverrides<'a> {
         KvOverrides::new(self)
     }
 

llama-cpp-2/src/model/params/kv_overrides.rs

@@ -78,7 +78,7 @@ pub struct KvOverrides<'a> {
 }
 
 impl KvOverrides<'_> {
-    pub(super) fn new(model_params: &LlamaModelParams) -> KvOverrides {
+    pub(super) fn new<'a>(model_params: &'a LlamaModelParams) -> KvOverrides<'a> {
         KvOverrides { model_params }
     }
 }

llama-cpp-sys-2/src/lib.rs

@@ -3,5 +3,6 @@
 #![allow(non_upper_case_globals)]
 #![allow(non_camel_case_types)]
 #![allow(non_snake_case)]
+#![allow(unpredictable_function_pointer_comparisons)]
 
 include!(concat!(env!("OUT_DIR"), "/bindings.rs"));