Llamafile
- 2024/05/19
Overview
Today I read a bit about Llamafile. It lets you distribute and run LLMs with a single file. This project aims to run LLMs on consumer grade hardware (last time I heard about Consumer grade hardware was when learning HDFS, Spark few years ago) .
Following are the advantages of using Llamafile:
- No installation is required
- Support for GPUs from multiple vendors (so far when I read about LLMs online, it has always been NVIDIA / CUDA. there was never a discussion about Intel's or AMD'S GPUs)
- CPU support (via
llama.cpp) - Local Inference
Links
- llamafile: bringing LLMs to the people, and to your own computer
- Llamafile’s progress, four months in
- LLaMA Now Goes Faster on CPUs
- Mozilla-Ocho
- A Quick Guide to Containerizing Llamafile with Docker for AI Applications
Handson
I followed the Quickstart instructions. Downloaded llava-v1.5-7b-q4.llamafile.
# Change file permissions
➜ chmod +x llava-v1.5-7b-q4.llamafile
➜ ls -l llava-v1.5-7b-q4.llamafile
-rwxrwxr-x 1 rk rk 4287904141 May 19 17:53 llava-v1.5-7b-q4.llamafile
# Execute
./llava-v1.5-7b-q4.llamafile
Encountered the following error:
zsh: exec format error: ./llava-v1.5-7b-q4.llamafile
There is already a comment about this on the GitHub page: If you use zsh and have trouble running llamafile, try saying sh -c ./llamafile. This is due to a bug that was fixed in zsh 5.9+. The same is the case for Python subprocess, old versions of Fish, etc.
sh -c ./llava-v1.5-7b-q4.llamafile -ngl 9999
note: if you have an AMD or NVIDIA GPU then you need to pass -ngl 9999 to enable GPU offloading
{"build":1500,"commit":"a30b324","function":"server_cli","level":"INFO","line":2856,"msg":"build info","tid":"11165056","timestamp":1716106492}
{"function":"server_cli","level":"INFO","line":2859,"msg":"system info","n_threads":6,"n_threads_batch":-1,"system_info":"AVX = 1 | AVX_VNNI = 1 | AVX2 = 1 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | FMA = 1 | NEON = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 0 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | ","tid":"11165056","timestamp":1716106492,"total_threads":20}
{"function":"load_model","level":"INFO","line":435,"msg":"Multi Modal Mode Enabled","tid":"11165056","timestamp":1716106492}
clip_model_load: model name: openai/clip-vit-large-patch14-336
clip_model_load: description: image encoder for LLaVA
clip_model_load: GGUF version: 3
clip_model_load: alignment: 32
clip_model_load: n_tensors: 377
clip_model_load: n_kv: 19
clip_model_load: ftype: q4_0
clip_model_load: loaded meta data with 19 key-value pairs and 377 tensors from llava-v1.5-7b-mmproj-Q4_0.gguf
clip_model_load: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
clip_model_load: - kv 0: general.architecture str = clip
clip_model_load: - kv 1: clip.has_text_encoder bool = false
clip_model_load: - kv 2: clip.has_vision_encoder bool = true
clip_model_load: - kv 3: clip.has_llava_projector bool = true
clip_model_load: - kv 4: general.file_type u32 = 2
clip_model_load: - kv 5: general.name str = openai/clip-vit-large-patch14-336
clip_model_load: - kv 6: general.description str = image encoder for LLaVA
clip_model_load: - kv 7: clip.vision.image_size u32 = 336
clip_model_load: - kv 8: clip.vision.patch_size u32 = 14
clip_model_load: - kv 9: clip.vision.embedding_length u32 = 1024
clip_model_load: - kv 10: clip.vision.feed_forward_length u32 = 4096
clip_model_load: - kv 11: clip.vision.projection_dim u32 = 768
clip_model_load: - kv 12: clip.vision.attention.head_count u32 = 16
clip_model_load: - kv 13: clip.vision.attention.layer_norm_epsilon f32 = 0.000010
clip_model_load: - kv 14: clip.vision.block_count u32 = 23
clip_model_load: - kv 15: clip.vision.image_mean arr[f32,3] = [0.481455, 0.457828, 0.408211]
clip_model_load: - kv 16: clip.vision.image_std arr[f32,3] = [0.268630, 0.261303, 0.275777]
clip_model_load: - kv 17: clip.use_gelu bool = false
clip_model_load: - kv 18: general.quantization_version u32 = 2
clip_model_load: - type f32: 235 tensors
clip_model_load: - type f16: 1 tensors
clip_model_load: - type q4_0: 141 tensors
clip_model_load: CLIP using CPU backend
clip_model_load: text_encoder: 0
clip_model_load: vision_encoder: 1
clip_model_load: llava_projector: 1
clip_model_load: model size: 169.18 MB
clip_model_load: metadata size: 0.14 MB
clip_model_load: params backend buffer size = 169.18 MB (377 tensors)
clip_model_load: compute allocated memory: 32.89 MB
llama_model_loader: loaded meta data with 19 key-value pairs and 291 tensors from llava-v1.5-7b-Q4_K.gguf (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv 0: general.architecture str = llama
llama_model_loader: - kv 1: general.name str = LLaMA v2
llama_model_loader: - kv 2: llama.context_length u32 = 4096
llama_model_loader: - kv 3: llama.embedding_length u32 = 4096
llama_model_loader: - kv 4: llama.block_count u32 = 32
llama_model_loader: - kv 5: llama.feed_forward_length u32 = 11008
llama_model_loader: - kv 6: llama.rope.dimension_count u32 = 128
llama_model_loader: - kv 7: llama.attention.head_count u32 = 32
llama_model_loader: - kv 8: llama.attention.head_count_kv u32 = 32
llama_model_loader: - kv 9: llama.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 10: general.file_type u32 = 15
llama_model_loader: - kv 11: tokenizer.ggml.model str = llama
llama_model_loader: - kv 12: tokenizer.ggml.tokens arr[str,32000] = ["<unk>", "<s>", "</s>", "<0x00>", "<...
llama_model_loader: - kv 13: tokenizer.ggml.scores arr[f32,32000] = [0.000000, 0.000000, 0.000000, 0.0000...
llama_model_loader: - kv 14: tokenizer.ggml.token_type arr[i32,32000] = [2, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ...
llama_model_loader: - kv 15: tokenizer.ggml.bos_token_id u32 = 1
llama_model_loader: - kv 16: tokenizer.ggml.eos_token_id u32 = 2
llama_model_loader: - kv 17: tokenizer.ggml.padding_token_id u32 = 0
llama_model_loader: - kv 18: general.quantization_version u32 = 2
llama_model_loader: - type f32: 65 tensors
llama_model_loader: - type q4_K: 193 tensors
llama_model_loader: - type q6_K: 33 tensors
llm_load_vocab: special tokens definition check successful ( 259/32000 ).
llm_load_print_meta: format = GGUF V3 (latest)
llm_load_print_meta: arch = llama
llm_load_print_meta: vocab type = SPM
llm_load_print_meta: n_vocab = 32000
llm_load_print_meta: n_merges = 0
llm_load_print_meta: n_ctx_train = 4096
llm_load_print_meta: n_embd = 4096
llm_load_print_meta: n_head = 32
llm_load_print_meta: n_head_kv = 32
llm_load_print_meta: n_layer = 32
llm_load_print_meta: n_rot = 128
llm_load_print_meta: n_embd_head_k = 128
llm_load_print_meta: n_embd_head_v = 128
llm_load_print_meta: n_gqa = 1
llm_load_print_meta: n_embd_k_gqa = 4096
llm_load_print_meta: n_embd_v_gqa = 4096
llm_load_print_meta: f_norm_eps = 0.0e+00
llm_load_print_meta: f_norm_rms_eps = 1.0e-05
llm_load_print_meta: f_clamp_kqv = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: f_logit_scale = 0.0e+00
llm_load_print_meta: n_ff = 11008
llm_load_print_meta: n_expert = 0
llm_load_print_meta: n_expert_used = 0
llm_load_print_meta: causal attn = 1
llm_load_print_meta: pooling type = 0
llm_load_print_meta: rope type = 0
llm_load_print_meta: rope scaling = linear
llm_load_print_meta: freq_base_train = 10000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_yarn_orig_ctx = 4096
llm_load_print_meta: rope_finetuned = unknown
llm_load_print_meta: ssm_d_conv = 0
llm_load_print_meta: ssm_d_inner = 0
llm_load_print_meta: ssm_d_state = 0
llm_load_print_meta: ssm_dt_rank = 0
llm_load_print_meta: model type = 7B
llm_load_print_meta: model ftype = Q4_K - Medium
llm_load_print_meta: model params = 6.74 B
llm_load_print_meta: model size = 3.80 GiB (4.84 BPW)
llm_load_print_meta: general.name = LLaMA v2
llm_load_print_meta: BOS token = 1 '<s>'
llm_load_print_meta: EOS token = 2 '</s>'
llm_load_print_meta: UNK token = 0 '<unk>'
llm_load_print_meta: PAD token = 0 '<unk>'
llm_load_print_meta: LF token = 13 '<0x0A>'
llm_load_tensors: ggml ctx size = 0.15 MiB
llm_load_tensors: CPU buffer size = 3891.24 MiB
..................................................................................................
llama_new_context_with_model: n_ctx = 2048
llama_new_context_with_model: n_batch = 2048
llama_new_context_with_model: n_ubatch = 512
llama_new_context_with_model: flash_attn = 0
llama_new_context_with_model: freq_base = 10000.0
llama_new_context_with_model: freq_scale = 1
llama_kv_cache_init: CPU KV buffer size = 1024.00 MiB
llama_new_context_with_model: KV self size = 1024.00 MiB, K (f16): 512.00 MiB, V (f16): 512.00 MiB
llama_new_context_with_model: CPU output buffer size = 0.14 MiB
llama_new_context_with_model: CPU compute buffer size = 164.01 MiB
llama_new_context_with_model: graph nodes = 1030
llama_new_context_with_model: graph splits = 1
{"function":"initialize","level":"INFO","line":489,"msg":"initializing slots","n_slots":1,"tid":"11165056","timestamp":1716106493}
{"function":"initialize","level":"INFO","line":498,"msg":"new slot","n_ctx_slot":2048,"slot_id":0,"tid":"11165056","timestamp":1716106493}
{"function":"server_cli","level":"INFO","line":3077,"msg":"model loaded","tid":"11165056","timestamp":1716106493}
llama server listening at http://127.0.0.1:8080
