Add new model

4 files changed, 40 insertions, 1 deletions

diff --git a/README.md b/README.md
index 9ee0241..fa8c3a9 100644
--- a/README.md
+++ b/README.md
@@ -56,3 +56,7 @@ This builds a Docker image and drops you into a shell with the prompt binary and
 ```bash
 make clean
 ```
+
+## Reading material
+
+- https://www.tinyllm.org/
diff --git a/context.txt b/context.txt
new file mode 100644
index 0000000..12d3073
--- /dev/null
+++ b/context.txt
@@ -0,0 +1,9 @@
+Gandalf: wizard, Lord of the Rings, grey beard, staff, Istari, Grey Pilgrim, Mithrandir, fought Sauron, helped destroy One Ring.
+
+Frodo: hobbit, Lord of the Rings, Bilbo's nephew, Shire, carried One Ring to Mount Doom, Fellowship of the Ring.
+
+Example: Who is Gandalf? Gandalf is a wizard from The Lord of the Rings.
+Example: Who is Frodo? Frodo is a hobbit from The Lord of the Rings.
+Example: Who is Harry Potter? I don't have that information.
+
+Answer this question. Use only the facts from above. If unknown, say "I don't have that information." Just give the answer, no prefix:
diff --git a/models.h b/models.h
index 2e1edb7..ac242fc 100644
--- a/models.h
+++ b/models.h
@@ -31,6 +31,30 @@ model_config models[] = {
         .min_p = 0.05f,
         .seed = LLAMA_DEFAULT_SEED,
     },
+    {
+        .name = "phi-4-mini-instruct",
+        .filepath = "models/Phi-4-mini-instruct.Q2_K.gguf",
+        .n_gpu_layers = 0,
+        .use_mmap = false,
+        .n_ctx = 131072,
+        .n_batch = 4096,
+        .embeddings = false,
+        .temperature = 0.8f,
+        .min_p = 0.05f,
+        .seed = LLAMA_DEFAULT_SEED,
+    },
+    {
+        .name = "tinyllama-1",
+        .filepath = "models/TinyLlama-1.1B-intermediate-step-1431k-3T-Q2_K.gguf",
+        .n_gpu_layers = 0,
+        .use_mmap = false,
+        .n_ctx = 2048,
+        .n_batch = 4096,
+        .embeddings = false,
+        .temperature = 0.8f,
+        .min_p = 0.05f,
+        .seed = LLAMA_DEFAULT_SEED,
+    },
 };
 
 const model_config *get_model_by_name(const char *name) {
diff --git a/models.txt b/models.txt
index 7cbb7c6..3b7a3bd 100644
--- a/models.txt
+++ b/models.txt
@@ -1 +1,3 @@
-https://huggingface.co/Felladrin/gguf-flan-t5-small/resolve/main/flan-t5-small.F16.gguf?download=true
+https://huggingface.co/Felladrin/gguf-flan-t5-small/resolve/main/flan-t5-small.F16.gguf
+https://huggingface.co/MaziyarPanahi/Phi-4-mini-instruct-GGUF/resolve/main/Phi-4-mini-instruct.Q2_K.gguf
+https://huggingface.co/andrijdavid/TinyLlama-1.1B-intermediate-step-1431k-3T-GGUF/resolve/main/TinyLlama-1.1B-intermediate-step-1431k-3T-Q2_K.gguf