cui-llama.rn 1.2.1 → 1.2.3

package/cpp/llama.cpp

@@ -12,9 +12,7 @@
 #  include "ggml-rpc.h"
 #endif
 
-#ifdef LM_GGML_USE_CUDA
-#  include "ggml-cuda.h"
-#elif defined(LM_GGML_USE_VULKAN)
+#if defined(LM_GGML_USE_VULKAN)
 #  include "ggml-vulkan.h"
 #elif defined(LM_GGML_USE_SYCL)
 #  include "ggml-sycl.h"
@@ -24,14 +22,6 @@
 #   include "ggml-cann.h"
 #endif
 
-#ifdef LM_GGML_USE_BLAS
-#  include "ggml-blas.h"
-#endif
-
-#ifdef LM_GGML_USE_METAL
-#  include "ggml-metal.h"
-#endif
-
 // TODO: replace with ggml API call
 #define QK_K 256
 
@@ -227,6 +217,7 @@ enum llm_arch {
     LLM_ARCH_RWKV6,
     LLM_ARCH_GRANITE,
     LLM_ARCH_GRANITE_MOE,
+    LLM_ARCH_CHAMELEON,
     LLM_ARCH_UNKNOWN,
 };
 
@@ -279,6 +270,7 @@ static const std::map<llm_arch, const char *> LLM_ARCH_NAMES = {
     { LLM_ARCH_RWKV6,           "rwkv6"        },
     { LLM_ARCH_GRANITE,         "granite"      },
     { LLM_ARCH_GRANITE_MOE,     "granitemoe"   },
+    { LLM_ARCH_CHAMELEON,       "chameleon"    },
     { LLM_ARCH_UNKNOWN,         "(unknown)"    },
 };
 
@@ -315,6 +307,7 @@ enum llm_kv {
     LLM_KV_DECODER_START_TOKEN_ID,
     LLM_KV_ATTN_LOGIT_SOFTCAPPING,
     LLM_KV_FINAL_LOGIT_SOFTCAPPING,
+    LLM_KV_SWIN_NORM,
     LLM_KV_RESCALE_EVERY_N_LAYERS,
     LLM_KV_TIME_MIX_EXTRA_DIM,
     LLM_KV_TIME_DECAY_EXTRA_DIM,
@@ -422,6 +415,7 @@ static const std::map<llm_kv, const char *> LLM_KV_NAMES = {
     { LLM_KV_DECODER_START_TOKEN_ID,            "%s.decoder_start_token_id"            },
     { LLM_KV_ATTN_LOGIT_SOFTCAPPING,            "%s.attn_logit_softcapping"            },
     { LLM_KV_FINAL_LOGIT_SOFTCAPPING,           "%s.final_logit_softcapping"           },
+    { LLM_KV_SWIN_NORM,                         "%s.swin_norm"                         },
     { LLM_KV_RESCALE_EVERY_N_LAYERS,            "%s.rescale_every_n_layers"            },
     { LLM_KV_TIME_MIX_EXTRA_DIM,                "%s.time_mix_extra_dim"                },
     { LLM_KV_TIME_DECAY_EXTRA_DIM,              "%s.time_decay_extra_dim"              },
@@ -613,9 +607,11 @@ enum llm_tensor {
     LLM_TENSOR_ENC_FFN_DOWN,
     LLM_TENSOR_ENC_FFN_UP,
     LLM_TENSOR_ENC_OUTPUT_NORM,
+    LLM_TENSOR_CLS,
+    LLM_TENSOR_CLS_OUT,
 };
 
-static const std::map<llm_arch, std::map<llm_tensor, std::string>> LLM_TENSOR_NAMES = {
+static const std::map<llm_arch, std::map<llm_tensor, const char *>> LLM_TENSOR_NAMES = {
     {
         LLM_ARCH_LLAMA,
         {
@@ -800,6 +796,8 @@ static const std::map<llm_arch, std::map<llm_tensor, std::string>> LLM_TENSOR_NA
             { LLM_TENSOR_LAYER_OUT_NORM,  "blk.%d.layer_output_norm" },
             { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
             { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
+            { LLM_TENSOR_CLS,             "cls" },
+            { LLM_TENSOR_CLS_OUT,         "cls.output" },
         },
     },
     {
@@ -835,6 +833,7 @@ static const std::map<llm_arch, std::map<llm_tensor, std::string>> LLM_TENSOR_NA
             { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
             { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
             { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
+            { LLM_TENSOR_CLS,             "cls" },
         },
     },
     {
@@ -1510,6 +1509,25 @@ static const std::map<llm_arch, std::map<llm_tensor, std::string>> LLM_TENSOR_NA
             { LLM_TENSOR_FFN_UP_EXPS,     "blk.%d.ffn_up_exps" },
         },
     },
+    {
+        LLM_ARCH_CHAMELEON,
+        {
+            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
+            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
+            { LLM_TENSOR_OUTPUT,          "output" },
+            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
+            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
+            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
+            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
+            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
+            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
+            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
+            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
+            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
+            { LLM_TENSOR_ATTN_Q_NORM,     "blk.%d.attn_q_norm" },
+            { LLM_TENSOR_ATTN_K_NORM,     "blk.%d.attn_k_norm" },
+        },
+    },
     {
         LLM_ARCH_UNKNOWN,
         {
@@ -1549,32 +1567,32 @@ struct LLM_TN {
         return LLM_TENSOR_NAMES.at(arch).at(tensor);
     }
 
-    std::string operator()(llm_tensor tensor, const std::string & suffix) const {
+    std::string operator()(llm_tensor tensor, const char * suffix) const {
         if (LLM_TENSOR_NAMES.at(arch).find(tensor) == LLM_TENSOR_NAMES.at(arch).end()) {
             return "__missing__";
         }
-        return LLM_TENSOR_NAMES.at(arch).at(tensor) + "." + suffix;
+        return std::string(LLM_TENSOR_NAMES.at(arch).at(tensor)) + "." + suffix;
     }
 
     std::string operator()(llm_tensor tensor, int bid) const {
         if (LLM_TENSOR_NAMES.at(arch).find(tensor) == LLM_TENSOR_NAMES.at(arch).end()) {
             return "__missing__";
         }
-        return ::format(LLM_TENSOR_NAMES.at(arch).at(tensor).c_str(), bid);
+        return ::format(LLM_TENSOR_NAMES.at(arch).at(tensor), bid);
     }
 
-    std::string operator()(llm_tensor tensor, const std::string & suffix, int bid) const {
+    std::string operator()(llm_tensor tensor, const char * suffix, int bid) const {
         if (LLM_TENSOR_NAMES.at(arch).find(tensor) == LLM_TENSOR_NAMES.at(arch).end()) {
             return "__missing__";
         }
-        return ::format(LLM_TENSOR_NAMES.at(arch).at(tensor).c_str(), bid) + "." + suffix;
+        return ::format(LLM_TENSOR_NAMES.at(arch).at(tensor), bid) + "." + suffix;
     }
 
-    std::string operator()(llm_tensor tensor, const std::string & suffix, int bid, int xid) const {
+    std::string operator()(llm_tensor tensor, const char * suffix, int bid, int xid) const {
         if (LLM_TENSOR_NAMES.at(arch).find(tensor) == LLM_TENSOR_NAMES.at(arch).end()) {
             return "__missing__";
         }
-        return ::format(LLM_TENSOR_NAMES.at(arch).at(tensor).c_str(), bid, xid) + "." + suffix;
+        return ::format(LLM_TENSOR_NAMES.at(arch).at(tensor), bid, xid) + "." + suffix;
     }
 };
 
@@ -2247,59 +2265,16 @@ static std::string llama_token_to_piece(const struct llama_model * model, llama_
     return piece;
 }
 
-static lm_ggml_backend_buffer_type_t llama_default_buffer_type_cpu(bool host_buffer) {
-    lm_ggml_backend_buffer_type_t buft = nullptr;
-
-#if defined(LM_GGML_USE_CUDA)
-    // host buffers should only be used when data is expected to be copied to/from the GPU
-    if (host_buffer) {
-        buft = lm_ggml_backend_cuda_host_buffer_type();
-    }
-#elif defined(LM_GGML_USE_SYCL)
-    if (host_buffer) {
-        buft = lm_ggml_backend_sycl_host_buffer_type();
-    }
-#elif defined(LM_GGML_USE_CANN)
-    if (host_buffer) {
-        buft = lm_ggml_backend_cann_host_buffer_type();
-    }
-#elif defined(LM_GGML_USE_CPU_HBM)
-    buft = lm_ggml_backend_cpu_hbm_buffer_type();
-#elif defined(LM_GGML_USE_VULKAN)
-    if (host_buffer) {
-        buft = lm_ggml_backend_vk_host_buffer_type();
-    }
-#endif
-
-    if (buft == nullptr) {
-        buft = lm_ggml_backend_cpu_buffer_type();
-    }
-    return buft;
-
-    LM_GGML_UNUSED(host_buffer);
-}
-
 //
 // globals
 //
 
-struct llama_state {
-    llama_state() {
-#ifdef LM_GGML_USE_METAL
-        lm_ggml_backend_metal_log_set_callback(log_callback, log_callback_user_data);
-#elif defined(LM_GGML_USE_CUDA)
-        lm_ggml_backend_cuda_log_set_callback(log_callback, log_callback_user_data);
-#elif defined(LM_GGML_USE_CANN)
-        lm_ggml_backend_cann_log_set_callback(log_callback, log_callback_user_data);
-#endif
-    }
-
-    // We save the log callback globally
+struct llama_logger_state {
     lm_ggml_log_callback log_callback = llama_log_callback_default;
     void * log_callback_user_data = nullptr;
 };
 
-static llama_state g_state;
+static llama_logger_state g_logger_state;
 
 // available llama models
 enum e_model {
@@ -2373,6 +2348,7 @@ struct llama_hparams {
     bool vocab_only;
     bool rope_finetuned;
     bool use_par_res;
+    bool swin_norm;
 
     uint32_t n_vocab;
     uint32_t n_ctx_train; // context size the model was trained on
@@ -2439,7 +2415,7 @@ struct llama_hparams {
 
     // needed by encoder-decoder models (e.g. T5, FLAN-T5)
     // ref: https://github.com/ggerganov/llama.cpp/pull/8141
-    llama_token dec_start_token_id = -1;
+    llama_token dec_start_token_id = LLAMA_TOKEN_NULL;
 
     enum llama_pooling_type      pooling_type            = LLAMA_POOLING_TYPE_NONE;
     enum llama_rope_type         rope_type               = LLAMA_ROPE_TYPE_NONE;
@@ -2881,6 +2857,7 @@ struct llama_model {
     llama_hparams hparams = {};
     llama_vocab   vocab;
 
+    // TODO: should init all tensors to nullptr
     struct lm_ggml_tensor * tok_embd;
     struct lm_ggml_tensor * type_embd;
     struct lm_ggml_tensor * pos_embd;
@@ -2893,16 +2870,25 @@ struct llama_model {
     struct lm_ggml_tensor * output_b;
     struct lm_ggml_tensor * output_norm_enc;
 
+    // classifier
+    struct lm_ggml_tensor * cls;
+    struct lm_ggml_tensor * cls_b;
+    struct lm_ggml_tensor * cls_out   = nullptr;
+    struct lm_ggml_tensor * cls_out_b = nullptr;
+
     std::vector<llama_layer> layers;
 
+    // gguf metadata
+    std::unordered_map<std::string, std::string> lm_gguf_kv;
+
     llama_split_mode split_mode;
     int main_gpu;
     int n_gpu_layers;
 
-    std::vector<std::string> rpc_servers;
+    // list of devices used in this model
+    std::vector<lm_ggml_backend_dev_t> devices;
 
-    // gguf metadata
-    std::unordered_map<std::string, std::string> lm_gguf_kv;
+    std::vector<std::string> rpc_servers;
 
     // layer -> buffer type mapping
     struct layer_buft {
@@ -2945,11 +2931,6 @@ struct llama_model {
             lm_ggml_free(ctx);
         }
         for (lm_ggml_backend_buffer_t buf : bufs) {
-#ifdef LM_GGML_USE_CUDA
-            if (lm_ggml_backend_buffer_get_type(buf) == lm_ggml_backend_cpu_buffer_type()) {
-                lm_ggml_backend_cuda_unregister_host_buffer(lm_ggml_backend_buffer_get_base(buf));
-            }
-#endif
             lm_ggml_backend_buffer_free(buf);
         }
         while (!lora_adapters.empty()) {
@@ -3314,12 +3295,8 @@ struct llama_context {
     std::unordered_map<struct llama_lora_adapter *, float> lora_adapters;
 
     std::vector<lm_ggml_backend_t> backends;
-#ifdef LM_GGML_USE_METAL
-    lm_ggml_backend_t backend_metal = nullptr;
-#endif
-#ifdef LM_GGML_USE_BLAS
-    lm_ggml_backend_t backend_blas = nullptr;
-#endif
+    std::vector<std::pair<lm_ggml_backend_t, lm_ggml_backend_set_n_threads_t>> set_n_threads_fns;
+
     lm_ggml_backend_t backend_cpu = nullptr;
 
     lm_ggml_threadpool_t threadpool       = nullptr;
@@ -3435,72 +3412,112 @@ struct llama_lora_adapter {
     }
 };
 
-static size_t llama_get_device_count(const llama_model & model) {
-    size_t count = 1;
-#if defined(LM_GGML_USE_CUDA)
-    count = lm_ggml_backend_cuda_get_device_count();
-#elif defined(LM_GGML_USE_SYCL)
-    count = lm_ggml_backend_sycl_get_device_count();
+static int llama_get_device_count(const llama_model & model) {
+    int count = (int) model.devices.size();
+
+#if defined(LM_GGML_USE_RPC)
+    count += (int) model.rpc_servers.size();
+#endif
+
+#if defined(LM_GGML_USE_SYCL)
+    count += lm_ggml_backend_sycl_get_device_count();
 #elif defined(LM_GGML_USE_VULKAN)
-    count = lm_ggml_backend_vk_get_device_count();
+    count += lm_ggml_backend_vk_get_device_count();
 #elif defined(LM_GGML_USE_CANN)
-    return lm_ggml_backend_cann_get_device_count();
-#endif
-#if defined(LM_GGML_USE_RPC)
-    count += model.rpc_servers.size();
+    count += lm_ggml_backend_cann_get_device_count();
 #endif
+
     return count;
+
     LM_GGML_UNUSED(model);
 }
 
-static lm_ggml_backend_buffer_type_t llama_default_buffer_type_offload(const llama_model & model, int gpu) {
+static lm_ggml_backend_buffer_type_t llama_default_buffer_type_cpu(const llama_model & model, bool host_buffer) {
     lm_ggml_backend_buffer_type_t buft = nullptr;
 
-#ifdef LM_GGML_USE_RPC
-    int rpc_count = (int)model.rpc_servers.size();
-#else
-    int rpc_count = 0;
+    if (host_buffer) {
+        for (auto * dev : model.devices) {
+            buft = lm_ggml_backend_dev_host_buffer_type(dev);
+            if (buft != nullptr) {
+                break;
+            }
+        }
+    }
+
+#if defined(LM_GGML_USE_SYCL)
+    if (host_buffer) {
+        buft = lm_ggml_backend_sycl_host_buffer_type();
+    }
+#elif defined(LM_GGML_USE_CANN)
+    if (host_buffer) {
+        buft = lm_ggml_backend_cann_host_buffer_type();
+    }
+#elif defined(LM_GGML_USE_CPU_HBM)
+    buft = lm_ggml_backend_cpu_hbm_buffer_type();
+#elif defined(LM_GGML_USE_VULKAN)
+    if (host_buffer) {
+        buft = lm_ggml_backend_vk_host_buffer_type();
+    }
 #endif
-    int local_gpu = gpu - rpc_count;
+
+    if (buft == nullptr) {
+        buft = lm_ggml_backend_cpu_buffer_type();
+    }
+    return buft;
+
+    LM_GGML_UNUSED(host_buffer);
+}
+
+static lm_ggml_backend_buffer_type_t llama_default_buffer_type_offload(const llama_model & model, int device) {
+    lm_ggml_backend_buffer_type_t buft = nullptr;
+
 #if defined(LM_GGML_USE_RPC)
-    if (gpu < rpc_count) {
-        const char * endpoint = model.rpc_servers[gpu].c_str();
+    int rpc_count = (int)model.rpc_servers.size();
+    if (device < rpc_count) {
+        const char * endpoint = model.rpc_servers[device].c_str();
         return lm_ggml_backend_rpc_buffer_type(endpoint);
     }
+    device -= rpc_count;
 #endif
-#if defined(LM_GGML_USE_METAL)
-    buft = lm_ggml_backend_metal_buffer_type();
-#elif defined(LM_GGML_USE_CUDA)
-    buft = lm_ggml_backend_cuda_buffer_type(local_gpu);
-#elif defined(LM_GGML_USE_VULKAN)
-    buft = lm_ggml_backend_vk_buffer_type(local_gpu);
+
+    if (device < (int)model.devices.size()) {
+        return lm_ggml_backend_dev_buffer_type(model.devices[device]);
+    }
+    device -= (int)model.devices.size();
+
+#if defined(LM_GGML_USE_VULKAN)
+    buft = lm_ggml_backend_vk_buffer_type(device);
 #elif defined(LM_GGML_USE_SYCL)
-    buft = lm_ggml_backend_sycl_buffer_type(local_gpu);
+    buft = lm_ggml_backend_sycl_buffer_type(device);
 #elif defined(LM_GGML_USE_KOMPUTE)
-    buft = lm_ggml_backend_kompute_buffer_type(local_gpu);
-    if (buft == nullptr) {
-        LLAMA_LOG_WARN("%s: cannot use GPU %d, check `vulkaninfo --summary`\n", __func__, local_gpu);
-    }
+    buft = lm_ggml_backend_kompute_buffer_type(device);
 #elif defined(LM_GGML_USE_CANN)
-    buft = lm_ggml_backend_cann_buffer_type(local_gpu);
+    buft = lm_ggml_backend_cann_buffer_type(device);
 #endif
 
     if (buft == nullptr) {
-        buft = llama_default_buffer_type_cpu(true);
+        buft = llama_default_buffer_type_cpu(model, true);
     }
     return buft;
+
     LM_GGML_UNUSED(model);
-    LM_GGML_UNUSED(local_gpu);
 }
 
 static lm_ggml_backend_buffer_type_t llama_default_buffer_type_split(const llama_model & model, int fallback_gpu, const float * tensor_split) {
     lm_ggml_backend_buffer_type_t buft = nullptr;
 
-#ifdef LM_GGML_USE_CUDA
-    if (lm_ggml_backend_cuda_get_device_count() > 1) {
-        buft = lm_ggml_backend_cuda_split_buffer_type(tensor_split);
+    // find a backend that supports split buffers
+    for (size_t i = 0; i < lm_ggml_backend_reg_count(); ++i) {
+        lm_ggml_backend_reg_t reg = lm_ggml_backend_reg_get(i);
+
+        auto lm_ggml_backend_split_buffer_type_fn = (lm_ggml_backend_split_buffer_type_t) lm_ggml_backend_reg_get_proc_address(reg, "lm_ggml_backend_split_buffer_type");
+        if (lm_ggml_backend_split_buffer_type_fn) {
+            buft = lm_ggml_backend_split_buffer_type_fn(tensor_split);
+            if (buft != nullptr) {
+                break;
+            }
+        }
     }
-#endif
 
 #ifdef LM_GGML_USE_SYCL
     if (lm_ggml_backend_sycl_get_device_count() > 1) {
@@ -3517,13 +3534,8 @@ static lm_ggml_backend_buffer_type_t llama_default_buffer_type_split(const llama
 }
 
 static size_t llama_get_device_memory(const llama_model & model, int device) {
-#ifdef LM_GGML_USE_RPC
-    int rpc_count = (int)model.rpc_servers.size();
-#else
-    int rpc_count = 0;
-#endif
-    int local_device = device - rpc_count;
 #if defined(LM_GGML_USE_RPC)
+    int rpc_count = (int)model.rpc_servers.size();
     if (device < rpc_count) {
         size_t total;
         size_t free;
@@ -3531,32 +3543,37 @@ static size_t llama_get_device_memory(const llama_model & model, int device) {
         lm_ggml_backend_rpc_get_device_memory(endpoint, &free, &total);
         return free;
     }
+    device = device - rpc_count;
 #endif
-#if defined(LM_GGML_USE_CUDA)
-    size_t total;
-    size_t free;
-    lm_ggml_backend_cuda_get_device_memory(local_device, &free, &total);
-    return free;
-#elif defined(LM_GGML_USE_SYCL)
+
+    if (device < (int)model.devices.size()) {
+        lm_ggml_backend_dev_t dev = model.devices[device];
+        size_t total;
+        size_t free;
+        lm_ggml_backend_dev_memory(dev, &free, &total);
+        return free;
+    }
+
+#if defined(LM_GGML_USE_SYCL)
     size_t total;
     size_t free;
-    lm_ggml_backend_sycl_get_device_memory(local_device, &free, &total);
+    lm_ggml_backend_sycl_get_device_memory(device, &free, &total);
     return free;
 #elif defined(LM_GGML_USE_VULKAN)
     size_t total;
     size_t free;
-    lm_ggml_backend_vk_get_device_memory(local_device, &free, &total);
+    lm_ggml_backend_vk_get_device_memory(device, &free, &total);
     return free;
 #elif defined(LM_GGML_USE_CANN)
     size_t total;
     size_t free;
-    lm_ggml_backend_cann_get_device_memory(local_device, &free, &total);
+    lm_ggml_backend_cann_get_device_memory(device, &free, &total);
     return free;
 #else
     return 1;
 #endif
     LM_GGML_UNUSED(model);
-    LM_GGML_UNUSED(local_device);
+    LM_GGML_UNUSED(device);
 }
 
 //
@@ -3599,7 +3616,7 @@ static bool llama_kv_cache_init(
             buft_layer_count[model.buft_layer[i].buft]++;
         }
     } else {
-        buft_layer_count[llama_default_buffer_type_cpu(true)] = n_layer;
+        buft_layer_count[llama_default_buffer_type_cpu(model, true)] = n_layer;
     }
 
     // create a context for each buffer type
@@ -4891,7 +4908,7 @@ struct llama_model_loader {
     static const int TENSOR_NOT_REQUIRED = 1;
     static const int TENSOR_DUPLICATED   = 2;
 
-    struct lm_ggml_tensor * create_tensor(struct lm_ggml_context * ctx, const std::string & name, const std::vector<int64_t> & ne, int flags = 0) {
+    struct lm_ggml_tensor * create_tensor(struct lm_ggml_context * ctx, const std::string & name, const std::initializer_list<int64_t> & ne, int flags = 0) {
         const struct lm_ggml_tensor * cur = check_tensor_dims(name, ne, !(flags & TENSOR_NOT_REQUIRED));
 
         if (cur == NULL) {
@@ -4901,7 +4918,7 @@ struct llama_model_loader {
         return create_tensor_for(ctx, cur, flags & TENSOR_DUPLICATED);
     }
 
-    struct lm_ggml_tensor * create_tensor_as_view(struct lm_ggml_context * ctx, struct lm_ggml_tensor * base, const std::string & name, const std::vector<int64_t> & ne, size_t offset, bool required = true) {
+    struct lm_ggml_tensor * create_tensor_as_view(struct lm_ggml_context * ctx, struct lm_ggml_tensor * base, const std::string & name, const std::initializer_list<int64_t> & ne, size_t offset, bool required = true) {
         const struct lm_ggml_tensor * cur = check_tensor_dims(name, ne, required);
 
         if (cur == NULL) {
@@ -4914,7 +4931,7 @@ struct llama_model_loader {
 
         std::array<int64_t, LM_GGML_MAX_DIMS> dims;
         for (size_t i = 0; i < LM_GGML_MAX_DIMS; ++i) {
-            dims[i] = i < ne.size() ? ne[i] : 1;
+            dims[i] = i < ne.size() ? ne.begin()[i] : 1;
         }
 
         struct lm_ggml_tensor * tensor = lm_ggml_view_4d(ctx, base,
@@ -5012,7 +5029,7 @@ struct llama_model_loader {
     // Returns false if cancelled by progress_callback
     bool load_all_data(
             struct lm_ggml_context * ctx,
-            llama_buf_map & bufs_mmap,
+            llama_buf_map & bufs,
             llama_mlocks * lmlocks,
             llama_progress_callback progress_callback,
             void * progress_callback_user_data) {
@@ -5021,43 +5038,94 @@ struct llama_model_loader {
         std::vector<no_init<uint8_t>> read_buf;
         std::vector<std::future<std::pair<lm_ggml_tensor *, bool>>> validation_result;
 
-#if defined(LM_GGML_USE_CUDA)
         // 4 staging buffers for async uploads, each sized 1MB seems to be a good default for single NVMe drives.
         // NVMe raid configurations might require more / larger buffers.
         constexpr size_t n_buffers = 4;
         constexpr size_t buffer_size = 1 * 1024 * 1024; // 1MB
 
         std::vector<lm_ggml_backend_buffer_t> host_buffers;
-        std::vector<void*> host_ptrs;
         std::vector<lm_ggml_backend_event_t> events;
+        std::vector<void *> host_ptrs;
         size_t buffer_idx = 0; // buffer to use for async loads
-
-        lm_ggml_backend_t cuda_backend = nullptr;
-        if (!use_mmap && !check_tensors) {
+        lm_ggml_backend_t upload_backend = [&](const char * fn) -> lm_ggml_backend_t {
+            if (use_mmap || check_tensors) {
+                return nullptr;
+            }
             // When not using mmaped io use async uploads from pinned memory to GPU memory.
-            // First determine if the CUDA backend is active, and if so, determine the device ID.
-            lm_ggml_backend_buffer_t buf = bufs_mmap.count(0) ? bufs_mmap.at(0) : nullptr;
-            if (buf) {
-                lm_ggml_backend_buffer_type_t buffer_type = lm_ggml_backend_buffer_get_type(buf);
-                for (int i = 0; i < lm_ggml_backend_cuda_get_device_count(); ++i) {
-                    auto * cuda_buffer_type = lm_ggml_backend_cuda_buffer_type(i);
-                    if (buffer_type == cuda_buffer_type) {
-                        cuda_backend = lm_ggml_backend_cuda_init(i);
-                        break;
-                    }
-                }
+            // First determine if the backend supports the necessary features for async uploads.
+            auto * buf = bufs.count(0) ? bufs.at(0) : nullptr;
+            if (!buf) {
+                LLAMA_LOG_DEBUG("%s: no buffer found for async uploads\n", fn);
+                return nullptr;
+            }
+
+            auto * buft = lm_ggml_backend_buffer_get_type(buf);
+            auto * dev = lm_ggml_backend_buft_get_device(buft);
+            if (!dev) {
+                LLAMA_LOG_DEBUG("%s: no device found for buffer type %s for async uploads\n", fn,
+                    lm_ggml_backend_buft_name(buft));
+                return nullptr;
+            }
+
+            if (buft != lm_ggml_backend_dev_buffer_type(dev)) {
+                LLAMA_LOG_DEBUG("%s: buffer type %s is not the default buffer type for device %s for async uploads\n", fn,
+                    lm_ggml_backend_buft_name(buft), lm_ggml_backend_dev_name(dev));
+                return nullptr;
+            }
+
+            lm_ggml_backend_dev_props props;
+            lm_ggml_backend_dev_get_props(dev, &props);
+            if (!props.caps.async || !props.caps.host_buffer || !props.caps.events) {
+                LLAMA_LOG_DEBUG("%s: device %s does not support async, host buffers or events\n", fn,
+                    lm_ggml_backend_dev_name(dev));
+                return nullptr;
             }
 
-            // If the cuda backend is active create pinned memory buffers and events for synchronisation.
-            if (cuda_backend) {
-                for (size_t idx = 0; idx < n_buffers; ++idx) {
-                    host_buffers.emplace_back(lm_ggml_backend_buft_alloc_buffer(llama_default_buffer_type_cpu(true), buffer_size));
-                    host_ptrs.emplace_back(lm_ggml_backend_buffer_get_base(host_buffers[idx]));
-                    events.emplace_back(lm_ggml_backend_event_new(cuda_backend));
+            auto * host_buft = lm_ggml_backend_dev_host_buffer_type(dev);
+            if (!host_buft) {
+                LLAMA_LOG_DEBUG("%s: no host buffer type found for device %s\n", fn,
+                    lm_ggml_backend_dev_name(dev));
+                return nullptr;
+            }
+
+            // If the backend is supported, create pinned memory buffers and events for synchronisation.
+            for (size_t idx = 0; idx < n_buffers; ++idx) {
+                auto * buf = lm_ggml_backend_buft_alloc_buffer(host_buft, buffer_size);
+                if (!buf) {
+                    LLAMA_LOG_DEBUG("%s: failed to allocate host buffer for async uploads for device %s\n", fn,
+                        lm_ggml_backend_dev_name(dev));
+                    return nullptr;
+                }
+
+                host_buffers.emplace_back(buf);
+                host_ptrs.emplace_back(lm_ggml_backend_buffer_get_base(buf));
+
+                auto * event = lm_ggml_backend_event_new(dev);
+                if (!event) {
+                    LLAMA_LOG_DEBUG("%s: failed to create event for async uploads for device %s\n", fn,
+                        lm_ggml_backend_dev_name(dev));
+                    return nullptr;
                 }
+
+                events.emplace_back(event);
+            }
+
+            lm_ggml_backend_t backend = lm_ggml_backend_dev_init(dev, nullptr);
+            if (!backend) {
+                LLAMA_LOG_DEBUG("%s: failed to initialize backend for device %s for async uploads\n", fn,
+                    lm_ggml_backend_dev_name(dev));
+                return nullptr;
             }
+
+            return backend;
+        }(__func__);
+
+        if (upload_backend) {
+            LLAMA_LOG_DEBUG("%s: using async uploads for device %s, buffer type %s, backend %s\n", __func__,
+                lm_ggml_backend_dev_name(lm_ggml_backend_get_device(upload_backend)),
+                lm_ggml_backend_buft_name(lm_ggml_backend_buffer_get_type(bufs.at(0))),
+                lm_ggml_backend_name(upload_backend));
         }
-#endif
 
         for (struct lm_ggml_tensor * cur = lm_ggml_get_first_tensor(ctx); cur != NULL; cur = lm_ggml_get_next_tensor(ctx, cur)) {
             const auto * weight = get_weight(lm_ggml_get_name(cur));
@@ -5077,8 +5145,8 @@ struct llama_model_loader {
             if (use_mmap) {
                 const auto & mapping = mappings.at(weight->idx);
                 lm_ggml_backend_buffer_t buf_mmap = nullptr;
-                if (bufs_mmap.count(weight->idx)) {
-                    buf_mmap = bufs_mmap.at(weight->idx);
+                if (bufs.count(weight->idx)) {
+                    buf_mmap = bufs.at(weight->idx);
                 }
                 uint8_t * data = (uint8_t *) mapping->addr + weight->offs;
 
@@ -5114,9 +5182,8 @@ struct llama_model_loader {
                         }));
                     }
                 } else {
-#if defined(LM_GGML_USE_CUDA)
-                    // If cuda_backend is valid load the tensor in chunks to pinned memory and upload the buffers asynchronously to the GPU.
-                    if (cuda_backend) {
+                    // If upload_backend is valid load the tensor in chunks to pinned memory and upload the buffers asynchronously to the GPU.
+                    if (upload_backend) {
                         file->seek(weight->offs, SEEK_SET);
 
                         size_t bytes_read = 0;
@@ -5126,17 +5193,14 @@ struct llama_model_loader {
 
                             lm_ggml_backend_event_synchronize(events[buffer_idx]);
                             file->read_raw(host_ptrs[buffer_idx], read_iteration);
-                            lm_ggml_backend_tensor_set_async(cuda_backend, cur, host_ptrs[buffer_idx], bytes_read, read_iteration);
-                            lm_ggml_backend_event_record(events[buffer_idx]);
+                            lm_ggml_backend_tensor_set_async(upload_backend, cur, host_ptrs[buffer_idx], bytes_read, read_iteration);
+                            lm_ggml_backend_event_record(events[buffer_idx], upload_backend);
 
                             bytes_read += read_iteration;
                             ++buffer_idx;
                             buffer_idx %= n_buffers;
                         }
-                    }
-                    else
-#endif
-                    {
+                    } else {
                         read_buf.resize(n_size);
                         file->seek(weight->offs, SEEK_SET);
                         file->read_raw(read_buf.data(), n_size);
@@ -5151,17 +5215,15 @@ struct llama_model_loader {
             size_done += n_size;
         }
 
-#if defined(LM_GGML_USE_CUDA)
-        // free temporary resources used for async cuda uploads
-        if (cuda_backend) {
-            for (size_t idx = 0; idx < n_buffers;++idx) {
-                lm_ggml_backend_event_synchronize(events[idx]);
-                lm_ggml_backend_event_free(events[idx]);
-                lm_ggml_backend_buffer_free(host_buffers[idx]);
-            }
-            lm_ggml_backend_free(cuda_backend);
+        // free temporary resources used for async uploads
+        for (auto * event : events) {
+            lm_ggml_backend_event_synchronize(event);
+            lm_ggml_backend_event_free(event);
         }
-#endif
+        for (auto * buf : host_buffers) {
+            lm_ggml_backend_buffer_free(buf);
+        }
+        lm_ggml_backend_free(upload_backend);
 
         // check validation results
         bool validation_failed = false;
@@ -5477,8 +5539,10 @@ static void llm_load_hparams(
                     }
                 } else {
                     switch (hparams.n_layer) {
+                        case 16: model.type = e_model::MODEL_1B; break; // Llama 3.2 1B
                         case 22: model.type = e_model::MODEL_1B; break;
                         case 26: model.type = e_model::MODEL_3B; break;
+                        case 28: model.type = e_model::MODEL_3B; break; // Llama 3.2 3B
                         // granite uses a vocab with len 49152
                         case 32: model.type = hparams.n_vocab == 49152 ? e_model::MODEL_3B : (hparams.n_vocab < 40000 ? e_model::MODEL_7B : e_model::MODEL_8B); break;
                         case 36: model.type = e_model::MODEL_8B; break; // granite
@@ -5591,11 +5655,11 @@ static void llm_load_hparams(
                 ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS,    hparams.f_norm_eps);
                 ml.get_key(LLM_KV_ATTENTION_CAUSAL,           hparams.causal_attn);
                 ml.get_key(LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT, hparams.n_vocab_type);
-                ml.get_key(LLM_KV_POOLING_TYPE,               hparams.pooling_type);
+                ml.get_key(LLM_KV_POOLING_TYPE,               hparams.pooling_type, false);
                 hparams.f_max_alibi_bias = 8.0f;
 
                 switch (hparams.n_layer) {
-                    case 4: model.type = e_model::MODEL_33M; break; // jina-embeddings-small
+                    case 4:  model.type = e_model::MODEL_33M;  break; // jina-embeddings-small
                     case 12: model.type = e_model::MODEL_137M; break; // jina-embeddings-base
                 }
             } break;
@@ -6095,6 +6159,18 @@ static void llm_load_hparams(
                     default: model.type = e_model::MODEL_UNKNOWN;
                 }
             } break;
+        case LLM_ARCH_CHAMELEON:
+            {
+                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
+                hparams.f_norm_eps = 1e-5;  // eps for qk-norm, torch default
+                ml.get_key(LLM_KV_SWIN_NORM, hparams.swin_norm);
+
+                switch (hparams.n_layer) {
+                    case 32: model.type = e_model::MODEL_7B; break;
+                    case 48: model.type = e_model::MODEL_34B; break;
+                    default: model.type = e_model::MODEL_UNKNOWN;
+               }
+            } break;
         default: (void)0;
     }
 
@@ -6288,6 +6364,7 @@ static void llm_load_vocab(
                     tokenizer_pre == "phi-2"   ||
                     tokenizer_pre == "jina-es" ||
                     tokenizer_pre == "jina-de" ||
+                    tokenizer_pre == "jina-v1-en" ||
                     tokenizer_pre == "jina-v2-es" ||
                     tokenizer_pre == "jina-v2-de" ||
                     tokenizer_pre == "jina-v2-code") {
@@ -6352,6 +6429,11 @@ static void llm_load_vocab(
             } else if (
                 tokenizer_pre == "exaone") {
                 vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_EXAONE;
+            } else if (
+                tokenizer_pre == "chameleon") {
+                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_CHAMELEON;
+                vocab.tokenizer_add_bos = true;
+                vocab.tokenizer_clean_spaces = false;
             } else {
                 throw std::runtime_error(format("unknown pre-tokenizer type: '%s'", tokenizer_pre.c_str()));
             }
@@ -6409,7 +6491,12 @@ static void llm_load_vocab(
 
     for (uint32_t i = 0; i < n_vocab; i++) {
         std::string word = lm_gguf_get_arr_str(ctx, token_idx, i);
-        LM_GGML_ASSERT(unicode_cpts_from_utf8(word).size() > 0);
+
+        //LM_GGML_ASSERT(unicode_cpts_from_utf8(word).size() > 0);
+        if (word.empty()) {
+            LLAMA_LOG_WARN("%s: empty token at index %u\n", __func__, i);
+            word = "[EMPTY_" + std::to_string(i) + "]";
+        }
 
         vocab.token_to_id[word] = i;
         vocab.max_token_len = std::max(vocab.max_token_len, (int) word.size());
@@ -6434,6 +6521,8 @@ static void llm_load_vocab(
     }
     LM_GGML_ASSERT(vocab.id_to_token.size() == vocab.token_to_id.size());
 
+    vocab.init_tokenizer();
+
     // determine the newline token: LLaMA "<0x0A>" == 10 == '\n', Falcon 193 == '\n'
     if (vocab.type == LLAMA_VOCAB_TYPE_SPM) {
         // For Fill-In-the-Middle (FIM)/infill models which where converted
@@ -6488,8 +6577,14 @@ static void llm_load_vocab(
         vocab.linefeed_id = ids[0];
     } else {
         const std::vector<int> ids = llama_tokenize_internal(vocab, "\xC4\x8A", false); // U+010A
-        LM_GGML_ASSERT(!ids.empty() && "model vocab missing newline token");
-        vocab.linefeed_id = ids[0];
+
+        //LM_GGML_ASSERT(!ids.empty() && "model vocab missing newline token");
+        if (ids.empty()) {
+            LLAMA_LOG_WARN("%s: model vocab missing newline token, using special_pad_id instead\n", __func__);
+            vocab.linefeed_id = vocab.special_pad_id;
+        } else {
+            vocab.linefeed_id = ids[0];
+        }
     }
 
     // special tokens
@@ -6864,6 +6959,13 @@ static bool llm_load_tensors(
         void * progress_callback_user_data) {
     auto & hparams = model.hparams;
 
+    // check if the value of main_gpu is valid
+    if (llama_get_device_count(model) > 0 &&
+        split_mode != LLAMA_SPLIT_MODE_LAYER &&
+        (main_gpu < 0 || main_gpu >= llama_get_device_count(model))) {
+        throw std::runtime_error(format("invalid value for main_gpu: %d (available devices: %d)", main_gpu, llama_get_device_count(model)));
+    }
+
     model.split_mode   = split_mode;
     model.main_gpu     = main_gpu;
     model.n_gpu_layers = n_gpu_layers;
@@ -6873,14 +6975,14 @@ static bool llm_load_tensors(
     bool use_mmap_buffer = true;
 
     // there is very little benefit to offloading the input layer, so always keep it on the CPU
-    model.buft_input = llama_default_buffer_type_cpu(true);
+    model.buft_input = llama_default_buffer_type_cpu(model, true);
     //model.buft_input = llama_default_buffer_type_offload(main_gpu);
 
     model.buft_layer.resize(n_layer);
 
     // assign cpu layers
     for (int i = 0; i < i_gpu_start; ++i) {
-        model.buft_layer[i] = llama_default_buffer_type_cpu(true);
+        model.buft_layer[i] = llama_default_buffer_type_cpu(model, true);
     }
 
     if (split_mode == LLAMA_SPLIT_MODE_LAYER) {
@@ -6918,7 +7020,7 @@ static bool llm_load_tensors(
             int layer_gpu = std::upper_bound(splits.begin(), splits.begin() + device_count, float(act_gpu_layers - 1)/act_gpu_layers) - splits.begin();
             model.buft_output = llama_default_buffer_type_offload(model, layer_gpu);
         } else {
-            model.buft_output = llama_default_buffer_type_cpu(true);
+            model.buft_output = llama_default_buffer_type_cpu(model, true);
         }
     } else {
         lm_ggml_backend_buffer_type_t split_buft;
@@ -6942,7 +7044,7 @@ static bool llm_load_tensors(
                 llama_default_buffer_type_offload(model, main_gpu)
             };
         } else {
-            model.buft_output = llama_default_buffer_type_cpu(true);
+            model.buft_output = llama_default_buffer_type_cpu(model, true);
         }
     }
 
@@ -7362,6 +7464,12 @@ static bool llm_load_tensors(
 
                     if (model.arch == LLM_ARCH_BERT) {
                         model.pos_embd = ml.create_tensor(ctx_input, tn(LLM_TENSOR_POS_EMBD,    "weight"), {n_embd, n_ctx_train});
+
+                        model.cls   = ml.create_tensor(ctx_output, tn(LLM_TENSOR_CLS, "weight"), {n_embd, n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
+                        model.cls_b = ml.create_tensor(ctx_output, tn(LLM_TENSOR_CLS, "bias"),   {n_embd},         llama_model_loader::TENSOR_NOT_REQUIRED);
+
+                        model.cls_out   = ml.create_tensor(ctx_output, tn(LLM_TENSOR_CLS_OUT, "weight"), {n_embd, 1}, llama_model_loader::TENSOR_NOT_REQUIRED);
+                        model.cls_out_b = ml.create_tensor(ctx_output, tn(LLM_TENSOR_CLS_OUT, "bias"),   {1},         llama_model_loader::TENSOR_NOT_REQUIRED);
                     }
 
                     model.tok_norm   = ml.create_tensor(ctx_output, tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {n_embd});
@@ -7414,6 +7522,8 @@ static bool llm_load_tensors(
                     model.tok_norm   = ml.create_tensor(ctx_output, tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {n_embd}); // LayerNorm
                     model.tok_norm_b = ml.create_tensor(ctx_output, tn(LLM_TENSOR_TOKEN_EMBD_NORM, "bias"),   {n_embd}); //LayerNorm bias
 
+                    model.cls   = ml.create_tensor(ctx_output, tn(LLM_TENSOR_CLS, "weight"), {n_embd, 1}, llama_model_loader::TENSOR_NOT_REQUIRED);
+                    model.cls_b = ml.create_tensor(ctx_output, tn(LLM_TENSOR_CLS, "bias"),   {1},         llama_model_loader::TENSOR_NOT_REQUIRED);
                     for (int i = 0; i < n_layer; ++i) {
                         lm_ggml_context * ctx_layer = ctx_for_layer(i);
                         lm_ggml_context * ctx_split = ctx_for_layer_split(i);
@@ -8739,6 +8849,45 @@ static bool llm_load_tensors(
                     }
 
                 } break;
+            case LLM_ARCH_CHAMELEON:
+                {
+                 model.tok_embd = ml.create_tensor(ctx_input, tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab});
+
+                 // output
+                    {
+                        model.output_norm = ml.create_tensor(ctx_output,       tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd});
+                        model.output      = ml.create_tensor(ctx_output_split, tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
+
+                        // if output is NULL, init from the input tok embed
+                        if (model.output == NULL) {
+                            model.output = ml.create_tensor(ctx_output, tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
+                        }
+                    }
+
+                    for (int i = 0; i < n_layer; ++i) {
+                        lm_ggml_context * ctx_layer = ctx_for_layer(i);
+                        lm_ggml_context * ctx_split = ctx_for_layer_split(i);
+
+                        auto & layer = model.layers[i];
+
+                        layer.attn_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd});
+                        layer.attn_q_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k, n_head});
+                        layer.attn_k_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k, n_head_kv});
+                        layer.attn_q_norm_b = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_Q_NORM, "bias", i),  {n_embd_head_k, n_head}, llama_model_loader::TENSOR_NOT_REQUIRED);
+                        layer.attn_k_norm_b = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_K_NORM, "bias", i),  {n_embd_head_k, n_head_kv}, llama_model_loader::TENSOR_NOT_REQUIRED);
+
+                        layer.wq = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd});
+                        layer.wk = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa});
+                        layer.wv = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa});
+                        layer.wo = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd});
+
+                        layer.ffn_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd});
+
+                        layer.ffn_gate = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff});
+                        layer.ffn_down = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd});
+                        layer.ffn_up   = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff});
+                    }
+                } break;
             default:
                 throw std::runtime_error("unknown architecture");
         }
@@ -8764,55 +8913,40 @@ static bool llm_load_tensors(
         llama_buf_map bufs;
         bufs.reserve(n_max_backend_buffer);
 
-        // only the mmap region containing the tensors in the model is mapped to the backend buffer
-        // this is important for metal with apple silicon: if the entire model could be mapped to a metal buffer, then we could just use metal for all layers
-        // this allows using partial offloading when the model size exceeds the metal buffer size, but not the RAM size
-        if (ml.use_mmap && use_mmap_buffer && buft == llama_default_buffer_type_cpu(true)) {
+        // check if this backend device supports buffer_from_host_ptr
+        // when using a host buffer as the CPU bakcend buffer, use the CPU device to prioritize using buffer_from_host_ptr over the host buffer
+        lm_ggml_backend_dev_t dev = lm_ggml_backend_buft_get_device(buft == llama_default_buffer_type_cpu(model, true) ? lm_ggml_backend_cpu_buffer_type() : buft);
+        bool buffer_from_host_ptr_supported = false;
+        if (dev) {
+            lm_ggml_backend_dev_props props;
+            lm_ggml_backend_dev_get_props(dev, &props);
+            buffer_from_host_ptr_supported = props.caps.buffer_from_host_ptr;
+        }
+
+        if (ml.use_mmap && use_mmap_buffer && buffer_from_host_ptr_supported) {
             for (uint32_t idx = 0; idx < ml.files.size(); idx++) {
+                // only the mmap region containing the tensors in the model is mapped to the backend buffer
+                // this is important for metal with apple silicon: if the entire model could be mapped to a metal buffer, then we could just use metal for all layers
+                // this allows using partial offloading when the model size exceeds the metal buffer size, but not the RAM size
                 void * addr = nullptr;
-                size_t first, last;
+                size_t first, last; // NOLINT
                 ml.get_mapping_range(&first, &last, &addr, idx, ctx);
                 if (first >= last) {
                     continue;
                 }
-                lm_ggml_backend_buffer_t buf = lm_ggml_backend_cpu_buffer_from_ptr((char *) addr + first, last - first);
+                const size_t max_size = lm_ggml_get_max_tensor_size(ctx);
+                lm_ggml_backend_buffer_t buf = lm_ggml_backend_dev_buffer_from_host_ptr(dev, (char *) addr + first, last - first, max_size);
                 if (buf == nullptr) {
-                    throw std::runtime_error("unable to allocate backend CPU buffer");
+                    throw std::runtime_error(format("unable to allocate %s buffer", lm_ggml_backend_buft_name(buft)));
                 }
                 model.bufs.push_back(buf);
                 bufs.emplace(idx, buf);
-#ifdef LM_GGML_USE_CUDA
-                if (n_layer >= n_gpu_layers) {
-                    lm_ggml_backend_cuda_register_host_buffer(
-                        lm_ggml_backend_buffer_get_base(buf),
-                        lm_ggml_backend_buffer_get_size(buf));
-                }
-#endif
             }
         }
-#ifdef LM_GGML_USE_METAL
-        else if (ml.use_mmap && use_mmap_buffer && buft == lm_ggml_backend_metal_buffer_type()) {
-            for (uint32_t idx = 0; idx < ml.files.size(); idx++) {
-                const size_t max_size = lm_ggml_get_max_tensor_size(ctx);
-                void * addr = nullptr;
-                size_t first, last;
-                ml.get_mapping_range(&first, &last, &addr, idx, ctx);
-                if (first >= last) {
-                    continue;
-                }
-                lm_ggml_backend_buffer_t buf = lm_ggml_backend_metal_buffer_from_ptr((char *) addr + first, last - first, max_size);
-                if (buf == nullptr) {
-                    throw std::runtime_error("unable to allocate backend metal buffer");
-                }
-                model.bufs.push_back(buf);
-                bufs.emplace(idx, buf);
-            }
-        }
-#endif
         else {
             lm_ggml_backend_buffer_t buf = lm_ggml_backend_alloc_ctx_tensors_from_buft(ctx, buft);
             if (buf == nullptr) {
-                throw std::runtime_error("unable to allocate backend buffer");
+                throw std::runtime_error(format("unable to allocate %s buffer", lm_ggml_backend_buft_name(buft)));
             }
             model.bufs.push_back(buf);
             if (use_mlock && lm_ggml_backend_buffer_is_host(buf)) {
@@ -10208,6 +10342,10 @@ struct llm_build_context {
         struct lm_ggml_tensor * cur;
 
         switch (pooling_type) {
+            case LLAMA_POOLING_TYPE_NONE:
+                {
+                    cur = inp;
+                } break;
             case LLAMA_POOLING_TYPE_MEAN:
                 {
                     struct lm_ggml_tensor * inp_mean = build_inp_mean();
@@ -10219,9 +10357,26 @@ struct llm_build_context {
                     struct lm_ggml_tensor * inp_cls = build_inp_cls();
                     cur = lm_ggml_get_rows(ctx0, inp, inp_cls);
                 } break;
-            case LLAMA_POOLING_TYPE_NONE:
+            case LLAMA_POOLING_TYPE_RANK:
                 {
-                    cur = inp;
+                    struct lm_ggml_tensor * inp_cls = build_inp_cls();
+                    inp = lm_ggml_get_rows(ctx0, inp, inp_cls);
+
+                    // classification head
+                    // https://github.com/huggingface/transformers/blob/5af7d41e49bbfc8319f462eb45253dcb3863dfb7/src/transformers/models/roberta/modeling_roberta.py#L1566
+                    LM_GGML_ASSERT(model.cls       != nullptr);
+                    LM_GGML_ASSERT(model.cls_b     != nullptr);
+
+                    cur = lm_ggml_add (ctx0, lm_ggml_mul_mat(ctx0, model.cls, inp), model.cls_b);
+                    cur = lm_ggml_tanh(ctx0, cur);
+
+                    // some models don't have `cls_out`, for example: https://huggingface.co/jinaai/jina-reranker-v1-tiny-en
+                    // https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/blob/cb5347e43979c3084a890e3f99491952603ae1b7/modeling_bert.py#L884-L896
+                    if (model.cls_out) {
+                        LM_GGML_ASSERT(model.cls_out_b != nullptr);
+
+                        cur = lm_ggml_add (ctx0, lm_ggml_mul_mat(ctx0, model.cls_out, cur), model.cls_out_b);
+                    }
                 } break;
             default:
                 {
@@ -11450,8 +11605,8 @@ struct llm_build_context {
             inpL = cur;
         }
 
-        // final output
         cur = inpL;
+
         cb(cur, "result_embd", -1);
 
         lm_ggml_build_forward_expand(gf, cur);
@@ -15883,6 +16038,184 @@ struct llm_build_context {
 
         return gf;
     }
+
+    // ref: https://github.com/facebookresearch/chameleon
+    // based on the original build_llama() function, changes:
+    //   * qk-norm
+    //   * swin-norm
+    //   * removed bias
+    //   * removed MoE
+    struct lm_ggml_cgraph * build_chameleon() {
+        struct lm_ggml_cgraph * gf = lm_ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+
+        // mutable variable, needed during the last layer of the computation to skip unused tokens
+        int32_t n_tokens = this->n_tokens;
+
+        const int64_t n_embd_head = hparams.n_embd_head_v;
+        LM_GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
+        LM_GGML_ASSERT(n_embd_head == hparams.n_rot);
+
+        struct lm_ggml_tensor * cur;
+        struct lm_ggml_tensor * inpL;
+
+        inpL = llm_build_inp_embd(ctx0, lctx, hparams, batch, model.tok_embd, cb);
+
+        // inp_pos - contains the positions
+        struct lm_ggml_tensor * inp_pos = build_inp_pos();
+
+        // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
+        struct lm_ggml_tensor * KQ_mask = build_inp_KQ_mask();
+
+        for (int il = 0; il < n_layer; ++il) {
+            struct lm_ggml_tensor * inpSA = inpL;
+
+            // norm
+            if (hparams.swin_norm) {
+                cur = inpL;
+            } else {
+                cur = llm_build_norm(ctx0, inpL, hparams,
+                    model.layers[il].attn_norm, NULL,
+                    LLM_NORM_RMS, cb, il);
+                cb(cur, "attn_norm", il);
+            }
+
+            // self-attention
+            {
+                // compute Q and K and RoPE them
+                struct lm_ggml_tensor * Qcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, cur);
+                cb(Qcur, "Qcur", il);
+
+                struct lm_ggml_tensor * Kcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wk, cur);
+                cb(Kcur, "Kcur", il);
+
+                struct lm_ggml_tensor * Vcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wv, cur);
+                cb(Vcur, "Vcur", il);
+
+                if (model.layers[il].attn_q_norm) {
+                    Qcur = lm_ggml_view_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens,
+                                lm_ggml_element_size(Qcur) * n_embd_head,
+                                lm_ggml_element_size(Qcur) * n_embd_head * n_head,
+                                0);
+                    cb(Qcur, "Qcur", il);
+
+                    Qcur = llm_build_norm(ctx0, Qcur, hparams,
+                                model.layers[il].attn_q_norm,
+                                model.layers[il].attn_q_norm_b,
+                                LLM_NORM, cb, il);
+                    cb(Qcur, "Qcur", il);
+                }
+
+                if (model.layers[il].attn_k_norm) {
+                    Kcur = lm_ggml_view_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens,
+                                lm_ggml_element_size(Kcur) * n_embd_head,
+                                lm_ggml_element_size(Kcur) * n_embd_head * n_head_kv,
+                                0);
+                    cb(Kcur, "Kcur", il);
+
+                    Kcur = llm_build_norm(ctx0, Kcur, hparams,
+                               model.layers[il].attn_k_norm,
+                               model.layers[il].attn_k_norm_b,
+                               LLM_NORM, cb, il);
+                    cb(Kcur, "Kcur", il);
+                }
+
+                Qcur = lm_ggml_rope_ext(
+                    ctx0, lm_ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens), inp_pos, nullptr,
+                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+                    ext_factor, attn_factor, beta_fast, beta_slow
+                );
+                cb(Qcur, "Qcur", il);
+
+                Kcur = lm_ggml_rope_ext(
+                    ctx0, lm_ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens), inp_pos, nullptr,
+                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+                    ext_factor, attn_factor, beta_fast, beta_slow
+                );
+                cb(Kcur, "Kcur", il);
+
+                cur = llm_build_kv(ctx0, lctx, kv_self, gf,
+                        model.layers[il].wo, nullptr,
+                        Kcur, Vcur, Qcur, KQ_mask, n_tokens, kv_head, n_kv, 1.0f/sqrtf(float(n_embd_head)), cb, il);
+
+                if (hparams.swin_norm) {
+                    cur = llm_build_norm(ctx0, cur, hparams,
+                        model.layers[il].attn_norm, NULL,
+                        LLM_NORM_RMS, cb, il);
+                }
+            }
+
+            if (il == n_layer - 1) {
+                // skip computing output for unused tokens
+                struct lm_ggml_tensor * inp_out_ids = build_inp_out_ids();
+                n_tokens = n_outputs;
+                cur   = lm_ggml_get_rows(ctx0,   cur, inp_out_ids);
+                inpSA = lm_ggml_get_rows(ctx0, inpSA, inp_out_ids);
+            }
+
+            struct lm_ggml_tensor * ffn_inp = lm_ggml_add(ctx0, cur, inpSA);
+            cb(ffn_inp, "ffn_inp", il);
+
+            // feed-forward network
+            if (!hparams.swin_norm) {
+                cur = llm_build_norm(ctx0, ffn_inp, hparams,
+                        model.layers[il].ffn_norm, NULL,
+                        LLM_NORM_RMS, cb, il);
+                cb(cur, "ffn_norm", il);
+            }
+
+            cur = llm_build_ffn(ctx0, lctx, cur,
+                    model.layers[il].ffn_up,   NULL, NULL,
+                    model.layers[il].ffn_gate, NULL, NULL,
+                    model.layers[il].ffn_down, NULL, NULL,
+                    NULL,
+                    LLM_FFN_SILU, LLM_FFN_PAR, cb, il);
+            cb(cur, "ffn_out", il);
+
+            if (hparams.swin_norm) {
+                cur = llm_build_norm(ctx0, cur, hparams,
+                        model.layers[il].ffn_norm, NULL,
+                        LLM_NORM_RMS, cb, il);
+                cb(cur, "ffn_norm", il);
+            }
+
+            cur = lm_ggml_add(ctx0, cur, ffn_inp);
+            cb(cur, "ffn_out", il);
+
+            cur = lctx.cvec.apply_to(ctx0, cur, il);
+            cb(cur, "l_out", il);
+
+            // input for next layer
+            inpL = cur;
+        }
+
+        cur = inpL;
+
+        cur = llm_build_norm(ctx0, cur, hparams,
+                model.output_norm, NULL,
+                LLM_NORM_RMS, cb, -1);
+        cb(cur, "result_norm", -1);
+
+        // lm_head
+        cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
+        cb(cur, "result_output_with_img_logits", -1);
+
+        // TODO: this suppresses the output of image tokens, which is required to enable text-only outputs.
+        // Needs to be removed once image outputs are supported.
+        int img_token_end_idx = 8196;
+        int img_token_start_idx = 4;
+        int num_img_tokens = img_token_end_idx - img_token_start_idx;
+        // creates 1d tensor of size num_img_tokens and values -FLT_MAX,
+        // which ensures that text token values are always at least larger than image token values
+        struct lm_ggml_tensor * img_logits = lm_ggml_new_tensor_1d(ctx0, LM_GGML_TYPE_F32, num_img_tokens);
+        img_logits = lm_ggml_clamp(ctx0, img_logits, -FLT_MAX, -FLT_MAX);
+        cb(img_logits, "img_logits", -1);
+        cur = lm_ggml_set_1d(ctx0, cur, img_logits, lm_ggml_element_size(cur) * img_token_start_idx);
+        cb(cur, "result_output", -1);
+
+        lm_ggml_build_forward_expand(gf, cur);
+
+        return gf;
+    }
 };
 
 static struct lm_ggml_cgraph * llama_build_graph_defrag(llama_context & lctx, const std::vector<uint32_t> & ids) {
@@ -16143,6 +16476,10 @@ static struct lm_ggml_cgraph * llama_build_graph(
             {
                 result = llm.build_rwkv6();
             } break;
+        case LLM_ARCH_CHAMELEON:
+            {
+                result = llm.build_chameleon();
+            } break;
         default:
             LM_GGML_ABORT("fatal error");
     }
@@ -16429,7 +16766,9 @@ static void llama_set_inputs(llama_context & lctx, const llama_ubatch & batch) {
         }
     }
 
-    if (cparams.embeddings && cparams.pooling_type == LLAMA_POOLING_TYPE_CLS) {
+    if (cparams.embeddings && (
+                cparams.pooling_type == LLAMA_POOLING_TYPE_CLS ||
+                cparams.pooling_type == LLAMA_POOLING_TYPE_RANK)) {
         const int64_t n_tokens     = batch.n_tokens;
         const int64_t n_seq_tokens = batch.n_seq_tokens;
         const int64_t n_seqs       = batch.n_seqs;
@@ -16444,7 +16783,7 @@ static void llama_set_inputs(llama_context & lctx, const llama_ubatch & batch) {
             const llama_seq_id seq_id = batch.seq_id[s][0];
 
             // TODO: adapt limits to n_seqs when batch.equal_seqs is true
-            LM_GGML_ASSERT(seq_id < n_tokens && "seq_id cannot be larger than n_tokens with pooling_type == CLS");
+            LM_GGML_ASSERT(seq_id < n_tokens && "seq_id cannot be larger than n_tokens with pooling_type == CLS or RANK");
 
             for (int i = 0; i < n_seq_tokens; ++i) {
                 const llama_pos pos = batch.pos[s*n_seq_tokens + i];
@@ -16646,7 +16985,7 @@ static size_t llama_output_reserve(llama_context & lctx, size_t n_outputs) {
             lctx.embd = nullptr;
         }
 
-        lctx.buf_output = lm_ggml_backend_buft_alloc_buffer(llama_default_buffer_type_cpu(true), new_size);
+        lctx.buf_output = lm_ggml_backend_buft_alloc_buffer(llama_default_buffer_type_cpu(lctx.model, true), new_size);
         if (lctx.buf_output == nullptr) {
             LLAMA_LOG_ERROR("%s: failed to allocate output buffer of size %.2f MiB\n", __func__, new_size / (1024.0 * 1024.0));
             return 0;
@@ -16715,24 +17054,20 @@ static void llama_graph_compute(
             lm_ggml_cgraph * gf,
                     int   n_threads,
         lm_ggml_threadpool * threadpool) {
-#ifdef LM_GGML_USE_METAL
-    if (lm_ggml_backend_is_metal(lctx.backend_metal)) {
-        lm_ggml_backend_metal_set_n_cb(lctx.backend_metal, n_threads);
-    }
-#endif
-
     if (lctx.backend_cpu != nullptr) {
-        lm_ggml_backend_cpu_set_n_threads(lctx.backend_cpu, n_threads);
         lm_ggml_backend_cpu_set_threadpool(lctx.backend_cpu, threadpool);
         lm_ggml_backend_cpu_set_abort_callback(lctx.backend_cpu, lctx.abort_callback, lctx.abort_callback_data);
     }
-#ifdef LM_GGML_USE_BLAS
-    if (lctx.backend_blas != nullptr) {
-        lm_ggml_backend_blas_set_n_threads(lctx.backend_blas, n_threads);
+
+    // set the number of threads for all the backends
+    for (const auto & set_n_threads_fn : lctx.set_n_threads_fns) {
+        set_n_threads_fn.second(set_n_threads_fn.first, n_threads);
     }
-#endif
 
-    lm_ggml_backend_sched_graph_compute_async(lctx.sched, gf);
+    auto err = lm_ggml_backend_sched_graph_compute_async(lctx.sched, gf);
+    if (err != LM_GGML_STATUS_SUCCESS) {
+        LLAMA_LOG_ERROR("%s: lm_ggml_backend_sched_graph_compute_async failed with error %d\n", __func__, err);
+    }
 
     // fprintf(stderr, "splits: %d\n", lm_ggml_backend_sched_get_n_splits(lctx.sched));
 }
@@ -16984,6 +17319,20 @@ static int llama_decode_internal(
                             lm_ggml_backend_tensor_get_async(backend_embd, embd, embd_seq_out[seq_id].data(), (n_embd*seq_id)*sizeof(float), n_embd*sizeof(float));
                         }
                     } break;
+                case LLAMA_POOLING_TYPE_RANK:
+                    {
+                        // extract the rerank score - a single float per sequence
+                        auto & embd_seq_out = lctx.embd_seq;
+
+                        for (uint32_t s = 0; s < ubatch.n_seqs; ++s) {
+                            const llama_seq_id seq_id = ubatch.seq_id[s][0];
+                            if (embd_seq_out.find(seq_id) != embd_seq_out.end()) {
+                                continue;
+                            }
+                            embd_seq_out[seq_id].resize(1);
+                            lm_ggml_backend_tensor_get_async(backend_embd, embd, embd_seq_out[seq_id].data(), (seq_id)*sizeof(float), sizeof(float));
+                        }
+                    } break;
                 case LLAMA_POOLING_TYPE_UNSPECIFIED:
                     {
                         LM_GGML_ABORT("unknown pooling type");
@@ -17190,6 +17539,13 @@ static int llama_encode_internal(
                             lm_ggml_backend_tensor_get_async(backend_embd, embd, embd_seq_out[seq_id].data(), (n_embd*seq_id)*sizeof(float), n_embd*sizeof(float));
                         }
                     } break;
+                case LLAMA_POOLING_TYPE_RANK:
+                    {
+                        // TODO: this likely should be the same logic as in llama_decoder_internal, but better to
+                        //       wait for an encoder model that requires this pooling type in order to test it
+                        //       https://github.com/ggerganov/llama.cpp/pull/9510
+                        LM_GGML_ABORT("RANK pooling not implemented yet");
+                    }
                 case LLAMA_POOLING_TYPE_UNSPECIFIED:
                     {
                         LM_GGML_ABORT("unknown pooling type");
@@ -17527,10 +17883,9 @@ static void llama_tensor_dequantize_internal(
     }
     float * f32_output = (float *) output.data();
 
-    lm_ggml_type_traits_t qtype;
+    const lm_ggml_type_traits * qtype = lm_ggml_get_type_traits(tensor->type);
     if (lm_ggml_is_quantized(tensor->type)) {
-        qtype = lm_ggml_internal_get_type_traits(tensor->type);
-        if (qtype.to_float == NULL) {
+        if (qtype->to_float == NULL) {
             throw std::runtime_error(format("type %s unsupported for integer quantization: no dequantization available", lm_ggml_type_name(tensor->type)));
         }
     } else if (tensor->type != LM_GGML_TYPE_F16 &&
@@ -17544,7 +17899,7 @@ static void llama_tensor_dequantize_internal(
         } else if (tensor->type == LM_GGML_TYPE_BF16) {
             lm_ggml_bf16_to_fp32_row((lm_ggml_bf16_t *)tensor->data, f32_output, nelements);
         } else if (lm_ggml_is_quantized(tensor->type)) {
-            qtype.to_float(tensor->data, f32_output, nelements);
+            qtype->to_float(tensor->data, f32_output, nelements);
         } else {
             LM_GGML_ABORT("fatal error"); // unreachable
         }
@@ -17580,7 +17935,7 @@ static void llama_tensor_dequantize_internal(
             } else if (typ == LM_GGML_TYPE_BF16) {
                 lm_ggml_bf16_to_fp32_row((lm_ggml_bf16_t *)inbuf, outbuf, nels);
             } else {
-                qtype.to_float(inbuf, outbuf, nels);
+                qtype->to_float(inbuf, outbuf, nels);
             }
         };
         workers.emplace_back(compute, tensor->type, (uint8_t *) tensor->data + in_buff_offs, f32_output + out_buff_offs, thr_elems);
@@ -18662,21 +19017,7 @@ struct llama_model_quantize_params llama_model_quantize_default_params() {
 }
 
 size_t llama_max_devices(void) {
-#if defined(LM_GGML_USE_RPC)
-    return LM_GGML_RPC_MAX_SERVERS;
-#elif defined(LM_GGML_USE_METAL)
-    return 1;
-#elif defined(LM_GGML_USE_CUDA)
-    return LM_GGML_CUDA_MAX_DEVICES;
-#elif defined(LM_GGML_USE_SYCL)
-    return LM_GGML_SYCL_MAX_DEVICES;
-#elif defined(LM_GGML_USE_VULKAN)
-    return LM_GGML_VK_MAX_DEVICES;
-#elif defined(LM_GGML_USE_CANN)
-    return LM_GGML_CANN_MAX_DEVICES;
-#else
-    return 1;
-#endif
+    return 16;
 }
 
 bool llama_supports_mmap(void) {
@@ -18688,12 +19029,13 @@ bool llama_supports_mlock(void) {
 }
 
 bool llama_supports_gpu_offload(void) {
-#if defined(LM_GGML_USE_CUDA) || defined(LM_GGML_USE_METAL)   || defined(LM_GGML_USE_VULKAN) || \
+#if defined(LM_GGML_USE_VULKAN) || \
     defined(LM_GGML_USE_SYCL) || defined(LM_GGML_USE_KOMPUTE) || defined(LM_GGML_USE_RPC)
     // Defined when llama.cpp is compiled with support for offloading model layers to GPU.
     return true;
 #else
-    return false;
+    return lm_ggml_backend_dev_by_type(LM_GGML_BACKEND_DEVICE_TYPE_GPU) != nullptr ||
+        lm_ggml_backend_dev_by_type(LM_GGML_BACKEND_DEVICE_TYPE_GPU_FULL) != nullptr;
 #endif
 }
 
@@ -18758,17 +19100,37 @@ struct llama_model * llama_load_model_from_file(
             return true;
         };
     }
+
     if (params.rpc_servers != nullptr && params.rpc_servers[0] != '\0') {
         // split the servers set them into model->rpc_servers
         std::string servers(params.rpc_servers);
         size_t pos = 0;
-        while ((pos = servers.find(",")) != std::string::npos) {
+        while ((pos = servers.find(',')) != std::string::npos) {
             std::string server = servers.substr(0, pos);
             model->rpc_servers.push_back(server);
             servers.erase(0, pos + 1);
         }
         model->rpc_servers.push_back(servers);
     }
+
+    // create list of devices to use with this model
+    // currently, we use all available devices
+    // TODO: rework API to give user more control over device selection
+    for (size_t i = 0; i < lm_ggml_backend_dev_count(); ++i) {
+        lm_ggml_backend_dev_t dev = lm_ggml_backend_dev_get(i);
+        switch (lm_ggml_backend_dev_type(dev)) {
+            case LM_GGML_BACKEND_DEVICE_TYPE_CPU:
+            case LM_GGML_BACKEND_DEVICE_TYPE_CPU_FULL:
+                // skip CPU backends since they are `handled separately
+                break;
+
+            case LM_GGML_BACKEND_DEVICE_TYPE_GPU:
+            case LM_GGML_BACKEND_DEVICE_TYPE_GPU_FULL:
+                model->devices.push_back(dev);
+                break;
+        }
+    }
+
     int status = llama_model_load(path_model, *model, params);
     LM_GGML_ASSERT(status <= 0);
     if (status < 0) {
@@ -18930,60 +19292,61 @@ struct llama_context * llama_new_context_with_model(
 
     if (!hparams.vocab_only) {
         // initialize backends
-#if defined(LM_GGML_USE_RPC)
-        if (model->n_gpu_layers > 0) {
-            for (const auto & endpoint : model->rpc_servers) {
-                lm_ggml_backend_t backend = lm_ggml_backend_rpc_init(endpoint.c_str());
+        int main_gpu = model->main_gpu;
+
+        // with registry
+        if (model->split_mode == LLAMA_SPLIT_MODE_NONE || model->split_mode == LLAMA_SPLIT_MODE_ROW) {
+            if (main_gpu >= 0 && main_gpu < (int)model->devices.size()) {
+                lm_ggml_backend_dev_t main_dev = model->devices[main_gpu];
+                lm_ggml_backend_t backend = lm_ggml_backend_dev_init(main_dev, nullptr);
                 if (backend == nullptr) {
-                    LLAMA_LOG_ERROR("%s: failed to initialize RPC to '%s'\n", __func__, endpoint.c_str());
+                    LLAMA_LOG_ERROR("%s: failed to initialize %s backend\n", __func__, lm_ggml_backend_dev_name(main_dev));
+                    llama_free(ctx);
+                    return nullptr;
+                }
+                ctx->backends.push_back(backend);
+            }
+        } else {
+            // LLAMA_SPLIT_MODE_LAYER requires a backend for each GPU
+            for (auto * dev : model->devices) {
+                lm_ggml_backend_t backend = lm_ggml_backend_dev_init(dev, nullptr);
+                if (backend == nullptr) {
+                    LLAMA_LOG_ERROR("%s: failed to initialize %s backend\n", __func__, lm_ggml_backend_dev_name(dev));
                     llama_free(ctx);
                     return nullptr;
                 }
                 ctx->backends.push_back(backend);
             }
         }
-#endif
+        if (main_gpu >= (int)model->devices.size()) {
+            main_gpu -= (int)model->devices.size();
+        }
 
-#if defined(LM_GGML_USE_METAL)
+#if defined(LM_GGML_USE_RPC)
         if (model->n_gpu_layers > 0) {
-            ctx->backend_metal = lm_ggml_backend_metal_init();
-            if (ctx->backend_metal == nullptr) {
-                LLAMA_LOG_ERROR("%s: failed to initialize Metal backend\n", __func__);
-                llama_free(ctx);
-                return nullptr;
-            }
-            ctx->backends.push_back(ctx->backend_metal);
-        }
-#elif defined(LM_GGML_USE_CUDA)
-        if (model->split_mode == LLAMA_SPLIT_MODE_NONE || model->split_mode == LLAMA_SPLIT_MODE_ROW) {
-            // with split_mode LLAMA_SPLIT_MODE_NONE or LLAMA_SPLIT_MODE_ROW, only the main GPU backend is used
-            lm_ggml_backend_t backend = lm_ggml_backend_cuda_init(model->main_gpu);
-            if (backend == nullptr) {
-                LLAMA_LOG_ERROR("%s: failed to initialize CUDA%d backend\n", __func__, model->main_gpu);
-                llama_free(ctx);
-                return nullptr;
-            }
-            ctx->backends.push_back(backend);
-        } else {
-            // LLAMA_SPLIT_MODE_LAYER requires a backend for each GPU
-            for (int device = 0; device < lm_ggml_backend_cuda_get_device_count(); ++device) {
-                lm_ggml_backend_t backend = lm_ggml_backend_cuda_init(device);
+            for (const auto & endpoint : model->rpc_servers) {
+                lm_ggml_backend_t backend = lm_ggml_backend_rpc_init(endpoint.c_str());
                 if (backend == nullptr) {
-                    LLAMA_LOG_ERROR("%s: failed to initialize CUDA%d backend\n", __func__, device);
+                    LLAMA_LOG_ERROR("%s: failed to initialize RPC to '%s'\n", __func__, endpoint.c_str());
                     llama_free(ctx);
                     return nullptr;
                 }
                 ctx->backends.push_back(backend);
             }
         }
-#elif defined(LM_GGML_USE_VULKAN)
+        if (main_gpu >= (int)model->rpc_servers.size()) {
+            main_gpu -= (int)model->rpc_servers.size();
+        }
+#endif
+
+#if defined(LM_GGML_USE_VULKAN)
         if (model->split_mode == LLAMA_SPLIT_MODE_ROW) {
             LLAMA_LOG_ERROR("%s: Row split not supported. Failed to initialize Vulkan backend\n", __func__);
             llama_free(ctx);
             return nullptr;
         }
         if (model->split_mode == LLAMA_SPLIT_MODE_NONE) {
-            lm_ggml_backend_t backend = lm_ggml_backend_vk_init(model->main_gpu);
+            lm_ggml_backend_t backend = lm_ggml_backend_vk_init(main_gpu);
             if (backend == nullptr) {
                 LLAMA_LOG_ERROR("%s: failed to initialize Vulkan backend\n", __func__);
                 llama_free(ctx);
@@ -19004,9 +19367,9 @@ struct llama_context * llama_new_context_with_model(
 #elif defined(LM_GGML_USE_SYCL)
         // with split_mode LLAMA_SPLIT_MODE_NONE or LLAMA_SPLIT_MODE_ROW, only the main GPU backend is used
         if (model->split_mode == LLAMA_SPLIT_MODE_NONE || model->split_mode == LLAMA_SPLIT_MODE_ROW) {
-            lm_ggml_backend_t backend = lm_ggml_backend_sycl_init(model->main_gpu);
+            lm_ggml_backend_t backend = lm_ggml_backend_sycl_init(main_gpu);
             if (backend == nullptr) {
-                LLAMA_LOG_ERROR("%s: failed to initialize SYCL%d backend\n", __func__, model->main_gpu);
+                LLAMA_LOG_ERROR("%s: failed to initialize SYCL%d backend\n", __func__, main_gpu);
                 llama_free(ctx);
                 return nullptr;
             }
@@ -19025,7 +19388,7 @@ struct llama_context * llama_new_context_with_model(
         }
 #elif defined(LM_GGML_USE_KOMPUTE)
         if (model->n_gpu_layers > 0) {
-            auto * backend = lm_ggml_backend_kompute_init(model->main_gpu);
+            auto * backend = lm_ggml_backend_kompute_init(main_gpu);
             if (backend == nullptr) {
                 LLAMA_LOG_ERROR("%s: failed to initialize Kompute backend\n", __func__);
                 llama_free(ctx);
@@ -19034,39 +19397,44 @@ struct llama_context * llama_new_context_with_model(
             ctx->backends.push_back(backend);
         }
 #elif defined(LM_GGML_USE_CANN)
-    // with split_mode LLAMA_SPLIT_MODE_NONE or LLAMA_SPLIT_MODE_ROW, only the main GPU backend is used
-    // TODO: lm_ggml_backend_cann is not support split tensor now, just leave code here.
-    if (model->split_mode == LLAMA_SPLIT_MODE_NONE || model->split_mode == LLAMA_SPLIT_MODE_ROW) {
-        lm_ggml_backend_t backend = lm_ggml_backend_cann_init(model->main_gpu);
-        if (backend == nullptr) {
-            LLAMA_LOG_ERROR("%s: failed to initialize CANN%d backend\n", __func__, model->main_gpu);
-            llama_free(ctx);
-            return nullptr;
-        }
-        ctx->backends.push_back(backend);
-    } else {
-        // LLAMA_SPLIT_MODE_LAYER requires a backend for each GPU
-        // TODO: currently, CANN can't use multi-gpus, just leave code here for further cann version.
-        for (int32_t device = 0; device < lm_ggml_backend_cann_get_device_count(); ++device) {
-            lm_ggml_backend_t backend = lm_ggml_backend_cann_init(device);
+        // with split_mode LLAMA_SPLIT_MODE_NONE or LLAMA_SPLIT_MODE_ROW, only the main GPU backend is used
+        // TODO: lm_ggml_backend_cann is not support split tensor now, just leave code here.
+        if (model->split_mode == LLAMA_SPLIT_MODE_NONE || model->split_mode == LLAMA_SPLIT_MODE_ROW) {
+            lm_ggml_backend_t backend = lm_ggml_backend_cann_init(main_gpu);
             if (backend == nullptr) {
-                LLAMA_LOG_ERROR("%s: failed to initialize CANN%d backend\n", __func__, device);
+                LLAMA_LOG_ERROR("%s: failed to initialize CANN%d backend\n", __func__, main_gpu);
                 llama_free(ctx);
                 return nullptr;
             }
             ctx->backends.push_back(backend);
+        } else {
+            // LLAMA_SPLIT_MODE_LAYER requires a backend for each GPU
+            // TODO: currently, CANN can't use multi-gpus, just leave code here for further cann version.
+            for (int32_t device = 0; device < lm_ggml_backend_cann_get_device_count(); ++device) {
+                lm_ggml_backend_t backend = lm_ggml_backend_cann_init(device);
+                if (backend == nullptr) {
+                    LLAMA_LOG_ERROR("%s: failed to initialize CANN%d backend\n", __func__, device);
+                    llama_free(ctx);
+                    return nullptr;
+                }
+                ctx->backends.push_back(backend);
+            }
         }
-    }
 #endif
 
-#ifdef LM_GGML_USE_BLAS
-        ctx->backend_blas = lm_ggml_backend_blas_init();
-        if (ctx->backend_blas == nullptr) {
-            LLAMA_LOG_WARN("%s: failed to initialize BLAS backend\n", __func__);
-        } else {
-            ctx->backends.push_back(ctx->backend_blas);
+        // add other backends (such as BLAS)
+        for (size_t i = 0; i < lm_ggml_backend_dev_count(); ++i) {
+            lm_ggml_backend_dev_t dev = lm_ggml_backend_dev_get(i);
+            if (lm_ggml_backend_dev_type(dev) == LM_GGML_BACKEND_DEVICE_TYPE_CPU) {
+                lm_ggml_backend_t backend = lm_ggml_backend_dev_init(dev, nullptr);
+                if (backend == nullptr) {
+                    LLAMA_LOG_ERROR("%s: failed to initialize %s backend\n", __func__, lm_ggml_backend_dev_name(dev));
+                    llama_free(ctx);
+                    return nullptr;
+                }
+                ctx->backends.push_back(backend);
+            }
         }
-#endif
 
         ctx->backend_cpu = lm_ggml_backend_cpu_init();
         if (ctx->backend_cpu == nullptr) {
@@ -19076,6 +19444,18 @@ struct llama_context * llama_new_context_with_model(
         }
         ctx->backends.push_back(ctx->backend_cpu);
 
+        // create a list of the set_n_threads functions in the backends
+        for (auto * backend : ctx->backends) {
+            lm_ggml_backend_dev_t dev = lm_ggml_backend_get_device(backend);
+            lm_ggml_backend_reg_t reg = dev ? lm_ggml_backend_dev_backend_reg(dev) : nullptr;
+            if (reg) {
+                auto lm_ggml_backend_set_n_threads_fn = (lm_ggml_backend_set_n_threads_t) lm_ggml_backend_reg_get_proc_address(reg, "lm_ggml_backend_set_n_threads");
+                if (lm_ggml_backend_set_n_threads_fn) {
+                    ctx->set_n_threads_fns.emplace_back(backend, lm_ggml_backend_set_n_threads_fn);
+                }
+            }
+        }
+
         if (!llama_kv_cache_init(ctx->kv_self, ctx, type_k, type_v, kv_size, cparams.offload_kqv)) {
             LLAMA_LOG_ERROR("%s: llama_kv_cache_init() failed for self-attention cache\n", __func__);
             llama_free(ctx);
@@ -19121,7 +19501,7 @@ struct llama_context * llama_new_context_with_model(
             for (auto * backend : ctx->backends) {
                 if (lm_ggml_backend_is_cpu(backend)) {
                     // use host buffers for the CPU backend compute buffer
-                    backend_buft.push_back(llama_default_buffer_type_cpu(true));
+                    backend_buft.push_back(llama_default_buffer_type_cpu(*model, true));
                 } else {
                     backend_buft.push_back(lm_ggml_backend_get_default_buffer_type(backend));
                 }
@@ -19132,17 +19512,37 @@ struct llama_context * llama_new_context_with_model(
             // buffer used to store the computation graph and the tensor meta data
             ctx->buf_compute_meta.resize(lm_ggml_tensor_overhead()*max_nodes + lm_ggml_graph_overhead_custom(max_nodes, false));
 
+            // TODO: move these checks to lm_ggml_backend_sched
             // enabling pipeline parallelism in the scheduler increases memory usage, so it is only done when necessary
             bool pipeline_parallel =
                 llama_get_device_count(*model) > 1 &&
                 model->n_gpu_layers > (int)model->hparams.n_layer &&
                 model->split_mode == LLAMA_SPLIT_MODE_LAYER &&
                 params.offload_kqv;
-#ifndef LM_GGML_USE_CUDA
-            // pipeline parallelism requires support for async compute and events
-            // currently this is only implemented in the CUDA backend
-            pipeline_parallel = false;
-#endif
+
+            // pipeline parallelism requires support for async compute and events in all devices
+            if (pipeline_parallel) {
+                for (auto * backend : ctx->backends) {
+                    if (lm_ggml_backend_is_cpu(backend)) {
+                        // ignore CPU backend
+                        continue;
+                    }
+                    auto * dev = lm_ggml_backend_get_device(backend);
+                    if (!dev) {
+                        // backend is using old interface, not supported
+                        pipeline_parallel = false;
+                        break;
+                    }
+                    lm_ggml_backend_dev_props props;
+                    lm_ggml_backend_dev_get_props(dev, &props);
+                    if (!props.caps.async || !props.caps.events) {
+                        // device does not support async compute or events
+                        pipeline_parallel = false;
+                        break;
+                    }
+                }
+            }
+
             ctx->sched = lm_ggml_backend_sched_new(ctx->backends.data(), backend_buft.data(), ctx->backends.size(), max_nodes, pipeline_parallel);
 
             if (pipeline_parallel) {
@@ -19268,6 +19668,7 @@ enum llama_rope_type llama_rope_type(const struct llama_model * model) {
         case LLM_ARCH_CHATGLM:
         case LLM_ARCH_GRANITE:
         case LLM_ARCH_GRANITE_MOE:
+        case LLM_ARCH_CHAMELEON:
             return LLAMA_ROPE_TYPE_NORM;
 
         // the pairs of head values are offset by n_rot/2
@@ -21446,15 +21847,9 @@ const std::vector<std::pair<std::string, struct lm_ggml_tensor *>> & llama_inter
 }
 
 void llama_log_set(lm_ggml_log_callback log_callback, void * user_data) {
-    g_state.log_callback = log_callback ? log_callback : llama_log_callback_default;
-    g_state.log_callback_user_data = user_data;
-#ifdef LM_GGML_USE_METAL
-    lm_ggml_backend_metal_log_set_callback(g_state.log_callback, g_state.log_callback_user_data);
-#elif defined(LM_GGML_USE_CUDA)
-    lm_ggml_backend_cuda_log_set_callback(g_state.log_callback, g_state.log_callback_user_data);
-#elif defined(LM_GGML_USE_CANN)
-    lm_ggml_backend_cann_log_set_callback(g_state.log_callback, g_state.log_callback_user_data);
-#endif
+    lm_ggml_log_set(log_callback, user_data);
+    g_logger_state.log_callback = log_callback ? log_callback : llama_log_callback_default;
+    g_logger_state.log_callback_user_data = user_data;
 }
 
 static void llama_log_internal_v(lm_ggml_log_level level, const char * format, va_list args) {
@@ -21463,12 +21858,12 @@ static void llama_log_internal_v(lm_ggml_log_level level, const char * format, v
     char buffer[128];
     int len = vsnprintf(buffer, 128, format, args);
     if (len < 128) {
-        g_state.log_callback(level, buffer, g_state.log_callback_user_data);
+        g_logger_state.log_callback(level, buffer, g_logger_state.log_callback_user_data);
     } else {
         char * buffer2 = new char[len + 1];
         vsnprintf(buffer2, len + 1, format, args_copy);
         buffer2[len] = 0;
-        g_state.log_callback(level, buffer2, g_state.log_callback_user_data);
+        g_logger_state.log_callback(level, buffer2, g_logger_state.log_callback_user_data);
         delete[] buffer2;
     }
     va_end(args_copy);