diff --git a/data/datasets/__init__.py b/data/datasets/__init__.py
index 7b2c077ea4..721c29bdab 100644
--- a/data/datasets/__init__.py
+++ b/data/datasets/__init__.py
@@ -34,6 +34,7 @@
     "reasoning_bg_oa": "0x22almostEvil/reasoning_bg_oa",
     "reasoning_gsm_qna_oa": "0x22almostEvil/reasoning-gsm-qna-oa",
     "semantics_ws_qna_oa": "0x22almostEvil/semantics-ws-qna-oa",
+    "alpaca-fa-instruction": "pourmand1376/alpaca-fa-instruction",
 }
 
 SAFETY_DATASETS = {
diff --git a/data/datasets/alpaca-fa-instruction/README.md b/data/datasets/alpaca-fa-instruction/README.md
new file mode 100644
index 0000000000..ba76e71d10
--- /dev/null
+++ b/data/datasets/alpaca-fa-instruction/README.md
@@ -0,0 +1,2 @@
+This is a persian instruction dataset. The dataset is uploaded
+[here](https://huggingface.co/datasets/pourmand1376/alpaca-fa-instruction).
diff --git a/data/datasets/alpaca-fa-multi/README.md b/data/datasets/alpaca-fa-multi/README.md
new file mode 100644
index 0000000000..8128647b41
--- /dev/null
+++ b/data/datasets/alpaca-fa-multi/README.md
@@ -0,0 +1,4 @@
+This is an multi-turn persian dataset which is in
+[orca-chat](https://huggingface.co/datasets/shahules786/orca-chat) format. It is
+published in
+[huggingface](https://huggingface.co/datasets/pourmand1376/alpaca-fa-multi).
diff --git a/model/model_training/custom_datasets/instruction.py b/model/model_training/custom_datasets/instruction.py
index 7b6ad39787..a746940dc0 100644
--- a/model/model_training/custom_datasets/instruction.py
+++ b/model/model_training/custom_datasets/instruction.py
@@ -32,6 +32,7 @@
     "evol_instruct_code": "nickrosh/Evol-Instruct-Code-80k-v1",
     "evol-codealpaca-v1": "theblackcat102/evol-codealpaca-v1",
     "cot_submix_original": "conceptofmind/cot_submix_original",
+    "alpaca-fa-instruction": "pourmand1376/alpaca-fa-instruction",
 }