Bug: loading llava models fails

[jakexcosme]

Note: This issue was copied from ggml-org#9455

Original Author: @mudler
Original Issue Number: ggml-org#9455
Created: 2024-09-12T16:59:37Z

---

What happened?

It seems that loading llava models crashes entirely. I can reproduce that 100% hit with moondream models.

this issue has been discussed already in ggml-org#9066 (comment) and in ggml-org#9294 (comment), this ticket is just a tracker to discuss about the issue

Name and Version

Commit still working here: 815b1fb
Commit which is not working: e6b7801 (which includes ggml-org#9082 ), also daa9623 is not working (which is older)

What operating system are you seeing the problem on?

Linux

Relevant log output

10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stderr /home/mudler/_git/LocalAI/backend/cpp/llama-avx2/llama.cpp/ggml/src/ggml.c:13835: GGML_ASSERT(i01 >= 0 &
& i01 < ne01) failed


10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout [Thread debugging using libthread_db enabled]
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout Using host libthread_db library "/lib64/libthread_db.so.1".
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout 0x00007f989b8e94a3 in ?? () from /lib64/libgomp.so.1
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #0  0x00007f989b8e94a3 in ?? () from /lib64/libgomp.so.1
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #1  0x00000000008222e5 in ggml_graph_compute_thread.isra ()
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #2  0x00007f989b8dcd16 in GOMP_parallel () from /lib64/libgomp.so.1
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #3  0x0000000000825a2a in ggml_graph_compute ()
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #4  0x0000000000834010 in ggml_backend_cpu_graph_compute ()
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #5  0x000000000083784c in ggml_backend_graph_compute ()
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #6  0x0000000000652b63 in clip_image_batch_encode.constprop ()
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #7  0x0000000000653553 in clip_image_encode ()
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #8  0x0000000000657ac8 in llava_image_embed_make_with_clip_img ()
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #9  0x00000000004e2c09 in llama_server_context::update_slots() [clone .isra.0] ()
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #10 0x00000000004d7629 in llama_server_queue::start_loop() ()
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout #11 0x000000000048b040 in main ()
10:25PM DBG GRPC(moondream2-text-model-f16.gguf-127.0.0.1:42747): stdout [Inferior 1 (process 13029) detached]

Note

• Flagged as critical as it completely crashes llama.cpp
• llama.cpp is being used as a library ( chore(deps): update llama.cpp mudler/LocalAI#3497 )
• Applying the suggestion described in Bug: MiniCPM-V-2.6 commit d565bb2fd5a2a58b9924a7a34e77a87c78c52137 causing crash in moondream ggml-org/llama.cpp#9066 (comment) seems to workaround the issue for me

diff --git a/examples/llava/clip.cpp b/examples/llava/clip.cpp
index 342042ff..224db9b5 100644
--- a/examples/llava/clip.cpp
+++ b/examples/llava/clip.cpp
@@ -2419,7 +2419,7 @@ bool clip_image_batch_encode(clip_ctx * ctx, const int n_threads, const clip_ima
             struct ggml_tensor * patches = ggml_graph_get_tensor(gf, "patches");
             int* patches_data = (int*)malloc(ggml_nbytes(patches));
             for (int i = 0; i < num_patches; i++) {
-                patches_data[i] = i + 1;
+                patches_data[i] = i;
             }
             ggml_backend_tensor_set(patches, patches_data, 0, ggml_nbytes(patches));
             free(patches_data);