...
llama_context: n_ctx_seq (2048) < n_ctx_train (262144) -- the full capacity of the model will not be utilized
llama_context: CUDA_Host output buffer size = 2.32 MiB
llama_kv_cache: CUDA0 KV buffer size = 192.00 MiB
llama_kv_cache: size = 192.00 MiB ( 2048 cells, 48 ​​layers, 4/1 seqs), K (f16): 96.00 MiB, V (f16): 96.00 MiB
sched_reserve: reserving ...
sched_reserve: Flash Attention was auto, set to enabled
sched_reserve: CUDA0 compute buffer size = 300.75 MiB
sched_reserve: CUDA_Host compute buffer size = 12.01 MiB
sched_reserve: graph nodes = 3031
sched_reserve: graph splits = 2
sched_reserve: reserve took 14.34 ms, sched copies = 1
common_init_from_params: warming up the model with an empty run - please wait ... (--no-warmup to disable)
ggml_cuda_compute_forward: MUL_MAT failed
CUDA error: device kernel image is invalid 
current device: 0, in function ggml_cuda_compute_forward at D:\a\llama.cpp\llama.cpp\ggml\src\ggml-cuda\ggml-cuda.cu:2758
err
D:\a\llama.cpp\llama.cpp\ggml\src\ggml-cuda\ggml-cuda.cu:97: CUDA error

e:\AI>c:\llama\b7951\llama-bench.exe   -m e:\AI\nomic-embed-text-v1.Q4_0.gguf   -t 4   -embd 1   -mmp 0
ggml_cuda_init: found 1 CUDA devices:
  Device 0: NVIDIA GeForce GTX 970M, compute capability 5.2, VMM: yes
load_backend: loaded CUDA backend from c:\llama\b7951\ggml-cuda.dll
load_backend: loaded RPC backend from c:\llama\b7951\ggml-rpc.dll
load_backend: loaded CPU backend from c:\llama\b7951\ggml-cpu-haswell.dll
| model                          |       size |     params | backend    | ngl |       embd |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | ---------: | --------------: | -------------------: |
| nomic-bert 137M Q4_0           |  73.48 MiB |   136.73 M | CUDA       |  99 |          1 |           pp512 |      4721.43 + 17.09 |
| nomic-bert 137M Q4_0           |  73.48 MiB |   136.73 M | CUDA       |  99 |          1 |           tg128 |        227.81 + 3.72 |

build: 22cae8321 (7951)

e:\AI>c:\llama\b7952\llama-bench.exe   -m e:\AI\nomic-embed-text-v1.Q4_0.gguf   -t 4   -embd 1   -mmp 0
ggml_cuda_init: found 1 CUDA devices:
  Device 0: NVIDIA GeForce GTX 970M, compute capability 5.2, VMM: yes
load_backend: loaded CUDA backend from c:\llama\b7952\ggml-cuda.dll
load_backend: loaded RPC backend from c:\llama\b7952\ggml-rpc.dll
load_backend: loaded CPU backend from c:\llama\b7952\ggml-cpu-haswell.dll
| model                          |       size |     params | backend    | ngl |       embd |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | ---------: | --------------: | -------------------: |
D:\a\llama.cpp\llama.cpp\ggml\src\ggml-cuda\ggml-cuda.cu:97: CUDA error

e:\AI>c:\llama\b7964\llama-bench.exe   -m e:\AI\nomic-embed-text-v1.Q4_0.gguf   -t 4   -embd 1   -mmp 0
ggml_cuda_init: found 1 CUDA devices:
  Device 0: NVIDIA GeForce GTX 970M, compute capability 5.2, VMM: yes
load_backend: loaded CUDA backend from c:\llama\b7964\ggml-cuda.dll
load_backend: loaded RPC backend from c:\llama\b7964\ggml-rpc.dll
load_backend: loaded CPU backend from c:\llama\b7964\ggml-cpu-haswell.dll
| model                          |       size |     params | backend    | ngl |       embd |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | ---------: | --------------: | -------------------: |
| nomic-bert 137M Q4_0           |  73.48 MiB |   136.73 M | CUDA       |  99 |          1 |           pp512 |      4735.51 + 15.17 |
| nomic-bert 137M Q4_0           |  73.48 MiB |   136.73 M | CUDA       |  99 |          1 |           tg128 |        224.12 + 7.50 |

build: b83111815 (7964)