@novastera-oss/llamarn 0.2.7 → 0.2.9

package/cpp/llama.cpp/src/llama-memory-recurrent.h

@@ -11,8 +11,8 @@
 // llama_memory_recurrent
 //
 
-// TODO: extract the cache state used for graph computation into llama_memory_recurrent_state_i
-//       see the implementation of llama_kv_cache_unified_state_i for an example how to do it
+// TODO: extract the cache state used for graph computation into llama_memory_recurrent_context_i
+//       see the implementation of llama_kv_cache_unified_context_i for an example how to do it
 class llama_memory_recurrent : public llama_memory_i {
 public:
 
@@ -34,14 +34,14 @@ public:
     // llama_memory_i
     //
 
-    llama_memory_state_ptr init_batch(
-            const llama_batch & batch,
+    llama_memory_context_ptr init_batch(
+            llama_batch_allocr & balloc,
             uint32_t n_ubatch,
             bool embd_all) override;
 
-    llama_memory_state_ptr init_full() override;
+    llama_memory_context_ptr init_full() override;
 
-    llama_memory_state_ptr init_update(llama_context * lctx, bool optimize) override;
+    llama_memory_context_ptr init_update(llama_context * lctx, bool optimize) override;
 
     void clear(bool data) override;
 
@@ -125,37 +125,34 @@ private:
     bool state_read_data(llama_io_read_i & io, uint32_t cell_count);
 };
 
-class llama_memory_recurrent_state : public llama_memory_state_i {
+class llama_memory_recurrent_context : public llama_memory_context_i {
 public:
     // used for errors
-    llama_memory_recurrent_state(llama_memory_status status);
+    llama_memory_recurrent_context(llama_memory_status status);
 
-    // used to create a full-cache state
-    llama_memory_recurrent_state(
+    // used to create a full-cache or update context
+    llama_memory_recurrent_context(
             llama_memory_recurrent * mem);
 
-    // used to create a state from a batch
-    llama_memory_recurrent_state(
+    // used to create a batch processing context from a batch
+    llama_memory_recurrent_context(
             llama_memory_recurrent * mem,
-            llama_sbatch sbatch,
             std::vector<llama_ubatch> ubatches);
 
-    virtual ~llama_memory_recurrent_state();
+    virtual ~llama_memory_recurrent_context();
 
     //
-    // llama_memory_state_i
+    // llama_memory_context_i
     //
 
     bool next()  override;
     bool apply() override;
 
-    std::vector<int64_t> & out_ids() override;
-
     llama_memory_status  get_status() const override;
     const llama_ubatch & get_ubatch() const override;
 
     //
-    // llama_memory_recurrent_state specific API
+    // llama_memory_recurrent_context specific API
     //
 
     uint32_t get_n_rs() const;
@@ -173,8 +170,6 @@ private:
 
     llama_memory_recurrent * mem;
 
-    llama_sbatch sbatch;
-
     size_t i_next = 0;
 
     std::vector<llama_ubatch> ubatches;

package/cpp/llama.cpp/src/llama-memory.h

@@ -3,10 +3,11 @@
 #include "llama.h"
 
 #include <memory>
-#include <vector>
 
 struct llama_ubatch;
 
+class llama_batch_allocr;
+
 class llama_io_write_i;
 class llama_io_read_i;
 
@@ -26,23 +27,21 @@ enum llama_memory_status {
     LLAMA_MEMORY_STATUS_FAILED_COMPUTE,
 };
 
-// helper function for combining the status of two memory states
+// helper function for combining the status of two memory contexts
 // useful for implementing hybrid memory types (e.g. iSWA)
 llama_memory_status llama_memory_status_combine(llama_memory_status s0, llama_memory_status s1);
 
-// the interface for managing the memory state during batch processing
+// the interface for managing the memory context during batch processing
 // this interface is implemented per memory type. see:
-//   - llama_kv_cache_unified_state
-//   - llama_kv_cache_unified_iswa_state
+//   - llama_kv_cache_unified_context
+//   - llama_kv_cache_unified_iswa_context
 //   ...
 //
-// the only method that can mutate the memory and the memory state is llama_memory_i::apply()
-//
-// TODO: rename to llama_memory_context_i ?
-struct llama_memory_state_i {
-    virtual ~llama_memory_state_i() = default;
+// the only method that should mutate the memory and the memory context is llama_memory_i::apply()
+struct llama_memory_context_i {
+    virtual ~llama_memory_context_i() = default;
 
-    // consume the current ubatch from the state and proceed to the next one
+    // consume the current ubatch from the context and proceed to the next one
     // return false if we are done
     virtual bool next() = 0;
 
@@ -50,17 +49,14 @@ struct llama_memory_state_i {
     // return false on failure
     virtual bool apply() = 0;
 
-    // TODO: this might get reworked in the future when refactoring llama_batch
-    virtual std::vector<int64_t> & out_ids() = 0;
-
     // get the current ubatch
     virtual const llama_ubatch & get_ubatch() const = 0;
 
-    // get the status of the memory state - used for error handling and checking if any updates would be applied
+    // get the status of the memory context - used for error handling and checking if any updates would be applied
     virtual llama_memory_status get_status() const = 0;
 };
 
-using llama_memory_state_ptr = std::unique_ptr<llama_memory_state_i>;
+using llama_memory_context_ptr = std::unique_ptr<llama_memory_context_i>;
 
 // general concept of LLM memory
 // the KV cache is a type of LLM memory, but there can be other types
@@ -68,19 +64,19 @@ struct llama_memory_i {
     virtual ~llama_memory_i() = default;
 
     // split the input batch into a set of ubatches and verify that they can fit into the cache
-    // return a state object containing the ubatches and KV cache state required to process them
-    // check the llama_memory_state_i::get_status() for the result
-    virtual llama_memory_state_ptr init_batch(
-            const llama_batch & batch,
+    // return a context object containing the ubatches and memory state required to process them
+    // check the llama_memory_context_i::get_status() for the result
+    virtual llama_memory_context_ptr init_batch(
+            llama_batch_allocr & balloc,
             uint32_t n_ubatch,
             bool embd_all) = 0;
 
     // simulate full cache, used for allocating worst-case compute buffers
-    virtual llama_memory_state_ptr init_full() = 0;
+    virtual llama_memory_context_ptr init_full() = 0;
 
     // prepare for any pending memory updates, such as shifts, defrags, etc.
     // status == LLAMA_MEMORY_STATUS_NO_UPDATE if there is nothing to update
-    virtual llama_memory_state_ptr init_update(llama_context * lctx, bool optimize) = 0;
+    virtual llama_memory_context_ptr init_update(llama_context * lctx, bool optimize) = 0;
 
     // getters
     virtual bool get_can_shift() const = 0;

package/cpp/llama.cpp/src/llama-model-saver.cpp

@@ -228,6 +228,7 @@ void llama_model_saver::add_kv_from_model() {
     // add_kv(LLM_KV_TOKENIZER_MASK_ID,                 ???);
     add_kv(LLM_KV_TOKENIZER_ADD_BOS,                 vocab.get_add_bos());
     add_kv(LLM_KV_TOKENIZER_ADD_EOS,                 vocab.get_add_eos());
+    add_kv(LLM_KV_TOKENIZER_ADD_SEP,                 vocab.get_add_sep());
     add_kv(LLM_KV_TOKENIZER_ADD_PREFIX,              vocab.get_add_space_prefix());
     add_kv(LLM_KV_TOKENIZER_REMOVE_EXTRA_WS,         vocab.get_remove_extra_whitespaces());
     add_kv(LLM_KV_TOKENIZER_PRECOMPILED_CHARSMAP,    vocab.get_precompiled_charsmap());