┌──────────────────────────────────────────────────────────────────────┐
│                                 spv                                  │
│                               ───────                                │
│                        insecurity researcher.                        │
│                          embrace the chaos                           │
└──────────────────────────────────────────────────────────────────────┘

spv's files ── LLaMA weights

────────────────────────────────────────────────────────────────────────

explanation
  /ai/LLaMA/llama_{7b,13b,
                   30b,65b}.zip: straight-from-Meta-weights,
                                 unconsolidated_pth. (uploading 30B
                                 right now and will upload 65B soon)
  /ai/LLaMA/llama_7b_ggml_f16.zip: LLaMA 7B, converted to GGML format
  /ai/LLaMA/llama_7b_ggml_q4_0.zip: LLaMA 7B, converted to GGML format
                                    and quantized to 4-bits.
  /ai/LLaMA/llama_13b_ggml_f16.zip: LLaMA 13B, converted to GGML format
  /ai/LLaMA/llama_13b_ggml_q4_0.zip: LLaMA 13B, converted to GGML format
                                     and quantized to 4-bits.

  these files are hosted by the wonderful Internet Archive / archive.org

aside
  despite this site (ftp.w212.org) being hosted by w212 research, this
  is still my (spv.sh) project

────────────────────────────────────────────────────────────────────────
~ spv
  embrace the chaos