finetune : add -ngl parameter (#3762)

* Add '-ngl' support to finetune.cpp

* Add fprintf in ggml_cuda_op_add

When I tried CUDA offloading during finetuning following the readme, I got an assert here.
This probably isn't an important case because inference later gives a warning saying you should use f16 or f32 instead when using lora

* Add 'finetune.sh', which currently fails when using GPU

"error: operator (): Finetuning on tensors with type 'f16' is not yet supported"

* tweak finetune.sh

* Suppress some warnings in ggml.c

* Add f16 implementation to ggml_compute_forward_add_f16_f32

* Add an f16 case to ggml_add_cast_impl and llama_build_lora_finetune_graphs

* finetune.sh: Edit comments

* Add "add_f16_f32_f32_cuda"

* Tweak an error message

* finetune.sh: Add an optional LLAMA_MODEL_DIR variable

* finetune.sh: Add an optional LLAMA_TRAINING_DIR variable

* train : minor

* tabs to spaces

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: cebtenzzre <cebtenzzre@gmail.com>

1 files changed, 13 insertions, 1 deletions

diff --git a/examples/finetune/finetune.cpp b/examples/finetune/finetune.cpp
index 35824cd2..60c7faa7 100644
--- a/examples/finetune/finetune.cpp
+++ b/examples/finetune/finetune.cpp
@@ -652,7 +652,7 @@ static struct ggml_tensor * llama_build_lora_finetune_graphs(
     GGML_ASSERT(tokens_input->type == GGML_TYPE_I32);
 
     auto add_to_f32 = [] (struct ggml_context * ctx, struct ggml_tensor * a, struct ggml_tensor * b) {
-        if (ggml_is_quantized(a->type)) {
+        if (ggml_is_quantized(a->type) || a->type == GGML_TYPE_F16) {
             return ggml_add_cast(ctx, a, b, GGML_TYPE_F32);
         } else if (a->type == GGML_TYPE_F32) {
             return ggml_add(ctx, a, b);
@@ -1459,6 +1459,17 @@ static bool train_params_parse(int argc, char ** argv, struct train_params * par
             }
             params->n_rank_w3 = std::stoi(argv[i]);
             params->custom_n_rank_w3 = true;
+        } else if (arg == "--gpu-layers" || arg == "-ngl" || arg == "--n-gpu-layers") {
+            if (++i >= argc) {
+                invalid_param = true;
+                break;
+            }
+#ifdef LLAMA_SUPPORTS_GPU_OFFLOAD
+            params->common.n_gpu_layers = std::stoi(argv[i]);
+#else
+            fprintf(stderr, "warning: not compiled with GPU offload support, --n-gpu-layers option will be ignored\n");
+            fprintf(stderr, "warning: see main README.md for information on enabling GPU BLAS support\n");
+#endif
         } else {
             fprintf(stderr, "error: unknown argument: %s\n", arg.c_str());
             train_print_usage(argc, argv, &default_params);
@@ -1545,6 +1556,7 @@ int main(int argc, char ** argv) {
     srand(params.common.seed);
 
     struct llama_model_params llama_mparams = llama_model_default_params();
+    llama_mparams.n_gpu_layers = params.common.n_gpu_layers;
     llama_mparams.vocab_only = false;
 
     printf("%s: model base = '%s'\n", __func__, params.fn_model_base);