Cleanup

Signed-off-by: Eric Curtin <ecurtin@redhat.com>
containers · Jul 26, 2024 · e1502a8 · e1502a8
1 parent 8df8552
commit e1502a8
Show file tree

Hide file tree

Showing 17 changed files with 242 additions and 759 deletions.
diff --git a/.clang-format b/.clang-format
diff --git a/README.md b/README.md
@@ -1,3 +1,81 @@
-# Project archived
+# Project undergoing rewrite
 
-New version at https://github.com/containers/ramalama
+The project is undergoing a complete rewrite in python3, the dependancy on containers will be removed. It will run on macOS and Linux native at a minimum and will support pulling, running and serving models from the following types of repos: huggingface, oci, ollama.
+
+You can still play away with the POC though :)
+
+# ramalama
+
+The goal of ramalama is to make AI even more boring.
+
+## Install
+
+Install ramalama by running this one-liner:
+
+```
+curl -fsSL https://raw.githubusercontent.com/containers/ramalama/main/install.sh | sudo bash
+```
+
+## Usage
+
+### Pulling Models
+
+You can pull a model using the `pull` command. By default, it pulls from the ollama registry.
+
+```
+$ ramalama pull granite-code
+```
+
+## Diagram
+
+```
++----------------+
+|                |
+| ramalama run   |
+|                |
++-------+--------+
+        |
+        v
++----------------+    +-----------------------+    +------------------+
+|                |    | Pull runtime layer    |    | Pull model layer |
+| Auto-detect    +--->| for llama.cpp         +--->| i.e. granite     |
+| hardware type  |    | (CPU, Vulkan, AMD,    |    |                  |
+|                |    |  Nvidia, Intel,       |    +------------------+
++----------------+    |  Apple Silicon, etc.) |    | Repo options:    |
+                      +-----------------------+    +-+-------+------+-+
+                                                     |       |      |
+                                                     v       v      v
+                                             +---------+ +------+ +----------+
+                                             | Hugging | | quay | | Ollama   |
+                                             | Face    | |      | | Registry |
+                                             +-------+-+ +---+--+ +-+--------+
+                                                     |       |      |
+                                                     v       v      v
+                                                   +------------------+
+                                                   | Start container  |
+                                                   | with llama.cpp   |
+                                                   | and granite      |
+                                                   | model            |
+                                                   +------------------+
+```
+
+## In development
+
+Regard this alpha, everything is under development, so expect breaking changes, luckily it's easy to reset everything and re-install:
+
+```
+rm -rf /var/lib/ramalama # only required if running as root user
+rm -rf $HOME/.local/share/ramalama
+```
+
+and install again.
+
+## Credit where credit is due
+
+For the vast majority of AI/LLM software we use, under the covers the heavy lifting is being done by:
+
+https://github.com/ggerganov/llama.cpp
+
+so if you like this tool, give llama.cpp repo a :star:, and hey, give us a :star: too while you are at it.
+
+![image](https://github.com/user-attachments/assets/d7a91662-5903-4117-ad41-2b193a852ea1)
diff --git a/ci.sh b/ci.sh
@@ -7,13 +7,12 @@ main() {
   curl -fsSL https://raw.githubusercontent.com/containers/ramalama/main/install.sh | sudo bash
 
   set +o pipefail
-  ramalama -h | grep Usage:
+  ./ramalama.py -h | grep Usage:
   set -o pipefail
 
-  ramalama pull granite
-  ramalama list | grep granite
-  ramalama rm granite
-  shellcheck "$(command -v ramalama)"
+  ramalama.py pull granite-code
+#  ramalama list | grep granite-code
+#  ramalama rm granite-code
 }
 
 main
diff --git a/container-images/ramalama/39-nvidia/Containerfile b/container-images/ramalama/39-nvidia/Containerfile
diff --git a/container-images/ramalama/39/Containerfile b/container-images/ramalama/39/Containerfile
diff --git a/container-images/ramalama/41-amd/Containerfile b/container-images/ramalama/41-amd/Containerfile
diff --git a/container-images/ramalama/latest-amd/Containerfile b/container-images/ramalama/latest-amd/Containerfile
diff --git a/container-images/ramalama/latest-nvidia/Containerfile b/container-images/ramalama/latest-nvidia/Containerfile
diff --git a/hf-db/granite b/hf-db/granite
diff --git a/hf-db/granite:3b b/hf-db/granite:3b
diff --git a/hf-db/merlinite b/hf-db/merlinite
diff --git a/hf-db/merlinite:7b b/hf-db/merlinite:7b
diff --git a/hf-db/mistral b/hf-db/mistral
diff --git a/hf-db/mistral:7b b/hf-db/mistral:7b
diff --git a/install.sh b/install.sh
@@ -62,7 +62,7 @@ main() {
   TMP="$(mktemp -d)"
   trap cleanup EXIT
   local from="ramalama"
-  local url="raw.githubusercontent.com/containers/ramalama/s/$from"
+  local url="raw.githubusercontent.com/containers/ramalama/main/$from"
   local from="$TMP/$from"
   download
   install -D -m755 "$from" "$bindir/"