Upload folder using huggingface_hub

run_train_test.sh

@@ -0,0 +1,33 @@
+
+LOG_ARGS="--log_interval 1 --save_interval 2500 --eval_interval 2500 --eval_iters 10"
+
+TRAIN_ARGS="--train_iters 50000 --lr_decay_style cosine 
+--lr_warmup_iters 50 --lr 1e-5 --min_lr 1e-6  --use_flash_attn 
+--attention_dropout 0.0 --adam_beta1 0.9 --adam_beta2 0.95 --adam_eps 1e-5"
+
+DISTRIBUTED_ARGS="--nproc_per_node 8 --nnodes 1 --node_rank 0 --master_addr localhost --master_port 8000"
+# DISTRIBUTED_ARGS="--nproc_per_node 16 --nnodes 2 --node_rank $RANK --master_addr $MASTER_ADDR --master_port $MASTER_PORT"
+
+LLAMA2_ARGS=" --no_bias_gelu_fusion --no_bias_dropout_fusion 
+--seq_length 4096 --max_position_embeddings 4096 
+--hidden_dropout 0.0 --rope_scaling_factor 1.0"
+
+CUDA_DEVICE_MAX_CONNECTIONS=1 \
+torchrun $DISTRIBUTED_ARGS finetune.py \
+	--tensor_model_parallel_size 8 \
+	--pipeline_model_parallel_size 1 \
+	--load megatron_test/llama2-megatron-t8p1 \
+	--save megatron_test/model_test \
+	--tensorboard_dir megatron_test/model_test/logs/ \
+	--data_path megatron_test/indonesian_mix_data/indonesian_mix_data_text_document  \
+	--model_name llama2 \
+	--tokenizer_type SentencePieceTokenizer \
+	--vocab_file megatron_test/llama2_tokenizer.model \
+	--bf16 \
+	--micro_batch_size 4 \
+	--global_batch_size 16 \
+	--sequence_parallel \
+    --recompute_granularity selective \
+	--use_checkpoint_args \
+    $COMMON_ARGS $LOG_ARGS $TRAIN_ARGS $LLAMA2_ARGS
+