Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

chat_template.jinja

@@ -0,0 +1,4 @@
+<s>{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% for message in messages %}{{'<|im_start|>' + message['role'] + '
+' + message['content'] + '<|im_end|>' + '
+'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant
+' }}{% endif %}

config.json

@@ -0,0 +1,43 @@
+{
+  "architectures": [
+    "ApertusForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "auto_map": {
+    "AutoConfig": "configuration_apertus.ApertusConfig",
+    "AutoModel": "modeling_apertus.ApertusModel",
+    "AutoModelForCausalLM": "modeling_apertus.ApertusForCausalLM"
+  },
+  "bos_token_id": 1,
+  "dtype": "bfloat16",
+  "eos_token_id": 74,
+  "hidden_act": "xielu",
+  "hidden_dropout": 0.0,
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 21504,
+  "max_position_embeddings": 65536,
+  "mlp_bias": false,
+  "model_type": "apertus",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pad_token_id": 3,
+  "post_norm": false,
+  "qk_norm": true,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 8.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3",
+    "type": "llama3"
+  },
+  "rope_theta": 12000000,
+  "tie_word_embeddings": false,
+  "transformers_version": "4.56.1",
+  "use_cache": false,
+  "vocab_size": 131072
+}

generation_config.json

@@ -0,0 +1,10 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": [
+    74
+  ],
+  "pad_token_id": 3,
+  "transformers_version": "4.56.1"
+}

latest

@@ -0,0 +1 @@
+global_step104

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53885a828db7f1548d57e6eecaee8616b5f37ab26d5c26a4212c17c810dc076c
+size 4999776656

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70d446f63ab248de8bd3c223ef9c9b8d3facb9decbbc851854286a82120c733f
+size 4882374192

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ab0d22b58be4a4705a63e3dff958cfe64c714b6e259758d2f2a8b0b7a209551
+size 4974647808

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59845fdef69972f525936551b7842373e76498535810c408f37cc901072b4611
+size 1249928160

model.safetensors.index.json

@@ -0,0 +1,459 @@
+{
+  "metadata": {
+    "total_parameters": 274496,
+    "total_size": 16106676480
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.feedforward_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.act_fn.alpha_n": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.act_fn.alpha_p": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.act_fn.beta": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.act_fn.eps": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.feedforward_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.act_fn.alpha_n": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.act_fn.alpha_p": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.act_fn.beta": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.act_fn.eps": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.2.attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.feedforward_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.act_fn.alpha_n": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.act_fn.alpha_p": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.act_fn.beta": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.act_fn.eps": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.20.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.21.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.3.attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.feedforward_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.act_fn.alpha_n": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.act_fn.alpha_p": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.act_fn.beta": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.act_fn.eps": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.30.attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.feedforward_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.act_fn.alpha_n": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.act_fn.alpha_p": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.act_fn.beta": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.act_fn.eps": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.attention_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.feedforward_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.act_fn.alpha_n": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.act_fn.alpha_p": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.act_fn.beta": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.act_fn.eps": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.4.attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.feedforward_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.act_fn.alpha_n": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.act_fn.alpha_p": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.act_fn.beta": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.act_fn.eps": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.feedforward_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.act_fn.alpha_n": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.act_fn.alpha_p": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.act_fn.beta": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.act_fn.eps": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.feedforward_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.act_fn.alpha_n": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.act_fn.alpha_p": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.act_fn.beta": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.act_fn.eps": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.feedforward_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.act_fn.alpha_n": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.act_fn.alpha_p": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.act_fn.beta": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.act_fn.eps": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.feedforward_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.act_fn.alpha_n": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.act_fn.alpha_p": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.act_fn.beta": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.act_fn.eps": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.feedforward_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.act_fn.alpha_n": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.act_fn.alpha_p": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.act_fn.beta": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.act_fn.eps": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.norm.weight": "model-00004-of-00004.safetensors"
+  }
+}

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f469fb6a869fe76761e1194ed0a7948ca397689bbc8ac0a9ea85a077fd50929
+size 16389

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77e31efd49e7c2510fff79f966c879db58740a4187714c13003ffa53d0d441c5
+size 16389

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:755aba68d004de8b7239e4451f96d8aaad4274ed7f03ec57d204f73d7b768a54
+size 16389

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1e57948798e97ec4bd65e4f2bab0090fd58ab95e9d421be20702021446d2636
+size 16389

rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7930f1c6dd64fe161f166b710675bea007029bf2a54e835287c8517c8d61b7e
+size 16389

rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c098b3d19df4c6a4261105183eb9357e2715d784d681c1426e4bb88c847c317
+size 16389

rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c914eacfcdfa6cf1a18175490235b8bf14f4521cc8ebda28a827fb8611e2958d
+size 16389

rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad02907448ab52a0de6407d8cc85b4523850947654c14a8ca1a3772f6c8c9cf8
+size 16389

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2b9980531e9a03282132d1ab3a420f8e27db73f249236d9aacc098905b0d50b
+size 1465

special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1723037d658a5b5900e4c836ba27c5af3a06ddf6ad73c74bd76e425a91fccd58
+size 17078476

trainer_state.json

@@ -0,0 +1,1178 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.9714285714285715,
+  "eval_steps": 500,
+  "global_step": 104,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01904761904761905,
+      "grad_norm": 1710495.2036469786,
+      "learning_rate": 0.0,
+      "loss": 1.3566,
+      "memory/device_reserved (GiB)": 126.71,
+      "memory/max_active (GiB)": 124.13,
+      "memory/max_allocated (GiB)": 122.77,
+      "step": 1,
+      "tokens_per_second_per_gpu": 3497.88
+    },
+    {
+      "epoch": 0.0380952380952381,
+      "grad_norm": 1558119.299961758,
+      "learning_rate": 8e-07,
+      "loss": 1.3538,
+      "memory/device_reserved (GiB)": 126.73,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 2,
+      "tokens_per_second_per_gpu": 3711.34
+    },
+    {
+      "epoch": 0.05714285714285714,
+      "grad_norm": 4186.518498313145,
+      "learning_rate": 1.6e-06,
+      "loss": 1.3529,
+      "memory/device_reserved (GiB)": 126.73,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 3,
+      "tokens_per_second_per_gpu": 3869.09
+    },
+    {
+      "epoch": 0.0761904761904762,
+      "grad_norm": 1230.1393406412694,
+      "learning_rate": 2.4e-06,
+      "loss": 1.3622,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.96,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 4,
+      "tokens_per_second_per_gpu": 3629.53
+    },
+    {
+      "epoch": 0.09523809523809523,
+      "grad_norm": 1035.6723923215748,
+      "learning_rate": 3.2e-06,
+      "loss": 1.3487,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 5,
+      "tokens_per_second_per_gpu": 3634.6
+    },
+    {
+      "epoch": 0.11428571428571428,
+      "grad_norm": 432.5460621726683,
+      "learning_rate": 4e-06,
+      "loss": 1.3432,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 6,
+      "tokens_per_second_per_gpu": 3776.49
+    },
+    {
+      "epoch": 0.13333333333333333,
+      "grad_norm": 1638.0331848931094,
+      "learning_rate": 4.8e-06,
+      "loss": 1.3677,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 7,
+      "tokens_per_second_per_gpu": 3647.88
+    },
+    {
+      "epoch": 0.1523809523809524,
+      "grad_norm": 2883.046779503214,
+      "learning_rate": 5.6e-06,
+      "loss": 1.3444,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 8,
+      "tokens_per_second_per_gpu": 3677.14
+    },
+    {
+      "epoch": 0.17142857142857143,
+      "grad_norm": 478.1216745871938,
+      "learning_rate": 6.4e-06,
+      "loss": 1.3305,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 9,
+      "tokens_per_second_per_gpu": 3739.63
+    },
+    {
+      "epoch": 0.19047619047619047,
+      "grad_norm": 1025.7505155071237,
+      "learning_rate": 7.2e-06,
+      "loss": 1.3362,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 10,
+      "tokens_per_second_per_gpu": 3730.56
+    },
+    {
+      "epoch": 0.20952380952380953,
+      "grad_norm": 1209.6274892436668,
+      "learning_rate": 8e-06,
+      "loss": 1.3325,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 11,
+      "tokens_per_second_per_gpu": 3661.21
+    },
+    {
+      "epoch": 0.22857142857142856,
+      "grad_norm": 1213.936189837833,
+      "learning_rate": 7.997766254921018e-06,
+      "loss": 1.3575,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 12,
+      "tokens_per_second_per_gpu": 3716.18
+    },
+    {
+      "epoch": 0.24761904761904763,
+      "grad_norm": 942.8786617202861,
+      "learning_rate": 7.991067514492613e-06,
+      "loss": 1.3145,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 13,
+      "tokens_per_second_per_gpu": 3587.32
+    },
+    {
+      "epoch": 0.26666666666666666,
+      "grad_norm": 3603.68277269405,
+      "learning_rate": 7.979911260354016e-06,
+      "loss": 1.3402,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 14,
+      "tokens_per_second_per_gpu": 3712.99
+    },
+    {
+      "epoch": 0.2857142857142857,
+      "grad_norm": 2487.402838754216,
+      "learning_rate": 7.96430995261912e-06,
+      "loss": 1.2956,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 15,
+      "tokens_per_second_per_gpu": 3762.61
+    },
+    {
+      "epoch": 0.3047619047619048,
+      "grad_norm": 667.5903281250161,
+      "learning_rate": 7.944281015960114e-06,
+      "loss": 1.2992,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 16,
+      "tokens_per_second_per_gpu": 3358.5
+    },
+    {
+      "epoch": 0.3238095238095238,
+      "grad_norm": 167.9027323688511,
+      "learning_rate": 7.919846820146347e-06,
+      "loss": 1.3119,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 17,
+      "tokens_per_second_per_gpu": 3675.6
+    },
+    {
+      "epoch": 0.34285714285714286,
+      "grad_norm": 47.46189855084341,
+      "learning_rate": 7.891034655060149e-06,
+      "loss": 1.302,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 18,
+      "tokens_per_second_per_gpu": 3754.58
+    },
+    {
+      "epoch": 0.3619047619047619,
+      "grad_norm": 115.37054783431222,
+      "learning_rate": 7.857876700217507e-06,
+      "loss": 1.3066,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 19,
+      "tokens_per_second_per_gpu": 3763.17
+    },
+    {
+      "epoch": 0.38095238095238093,
+      "grad_norm": 72.03472195336599,
+      "learning_rate": 7.820409988827649e-06,
+      "loss": 1.2876,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 20,
+      "tokens_per_second_per_gpu": 3750.15
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 123.88987560365385,
+      "learning_rate": 7.778676366431674e-06,
+      "loss": 1.2854,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 21,
+      "tokens_per_second_per_gpu": 3556.91
+    },
+    {
+      "epoch": 0.41904761904761906,
+      "grad_norm": 36.51030416393311,
+      "learning_rate": 7.73272244416641e-06,
+      "loss": 1.2799,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 22,
+      "tokens_per_second_per_gpu": 3627.03
+    },
+    {
+      "epoch": 0.4380952380952381,
+      "grad_norm": 37.445205147197846,
+      "learning_rate": 7.682599546705715e-06,
+      "loss": 1.2835,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 23,
+      "tokens_per_second_per_gpu": 3604.91
+    },
+    {
+      "epoch": 0.45714285714285713,
+      "grad_norm": 39.93974794828826,
+      "learning_rate": 7.628363654937363e-06,
+      "loss": 1.2947,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 24,
+      "tokens_per_second_per_gpu": 3782.86
+    },
+    {
+      "epoch": 0.47619047619047616,
+      "grad_norm": 59.41355630536809,
+      "learning_rate": 7.570075343439524e-06,
+      "loss": 1.2702,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 25,
+      "tokens_per_second_per_gpu": 3694.52
+    },
+    {
+      "epoch": 0.49523809523809526,
+      "grad_norm": 34.32373819297229,
+      "learning_rate": 7.507799712826686e-06,
+      "loss": 1.2984,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 26,
+      "tokens_per_second_per_gpu": 3613.01
+    },
+    {
+      "epoch": 0.5142857142857142,
+      "grad_norm": 21.68779916764309,
+      "learning_rate": 7.441606317040558e-06,
+      "loss": 1.2827,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 27,
+      "tokens_per_second_per_gpu": 3616.18
+    },
+    {
+      "epoch": 0.5333333333333333,
+      "grad_norm": 30.472648556953168,
+      "learning_rate": 7.371569085667188e-06,
+      "loss": 1.2801,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 28,
+      "tokens_per_second_per_gpu": 3754.99
+    },
+    {
+      "epoch": 0.5523809523809524,
+      "grad_norm": 19.319274693345776,
+      "learning_rate": 7.297766241367041e-06,
+      "loss": 1.2693,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 29,
+      "tokens_per_second_per_gpu": 3677.68
+    },
+    {
+      "epoch": 0.5714285714285714,
+      "grad_norm": 34.31430237097932,
+      "learning_rate": 7.220280212510252e-06,
+      "loss": 1.2581,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 30,
+      "tokens_per_second_per_gpu": 3730.31
+    },
+    {
+      "epoch": 0.5904761904761905,
+      "grad_norm": 82.8518096206661,
+      "learning_rate": 7.139197541114644e-06,
+      "loss": 1.2687,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 31,
+      "tokens_per_second_per_gpu": 3650.37
+    },
+    {
+      "epoch": 0.6095238095238096,
+      "grad_norm": 36.99675013730897,
+      "learning_rate": 7.0546087861893285e-06,
+      "loss": 1.2809,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 32,
+      "tokens_per_second_per_gpu": 3785.35
+    },
+    {
+      "epoch": 0.6285714285714286,
+      "grad_norm": 10.853195813384238,
+      "learning_rate": 6.96660842259183e-06,
+      "loss": 1.253,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 33,
+      "tokens_per_second_per_gpu": 3666.64
+    },
+    {
+      "epoch": 0.6476190476190476,
+      "grad_norm": 27.05353511161411,
+      "learning_rate": 6.875294735511717e-06,
+      "loss": 1.2601,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 34,
+      "tokens_per_second_per_gpu": 3808.86
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 11.079685605370564,
+      "learning_rate": 6.780769710698569e-06,
+      "loss": 1.2539,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 35,
+      "tokens_per_second_per_gpu": 3708.96
+    },
+    {
+      "epoch": 0.6857142857142857,
+      "grad_norm": 35.34021537624741,
+      "learning_rate": 6.683138920556894e-06,
+      "loss": 1.2362,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 36,
+      "tokens_per_second_per_gpu": 3819.32
+    },
+    {
+      "epoch": 0.7047619047619048,
+      "grad_norm": 47.246402607795154,
+      "learning_rate": 6.582511406235209e-06,
+      "loss": 1.2429,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 37,
+      "tokens_per_second_per_gpu": 3762.22
+    },
+    {
+      "epoch": 0.7238095238095238,
+      "grad_norm": 35.65219209343969,
+      "learning_rate": 6.4789995558409795e-06,
+      "loss": 1.2535,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 38,
+      "tokens_per_second_per_gpu": 3496.79
+    },
+    {
+      "epoch": 0.7428571428571429,
+      "grad_norm": 13.147263166038922,
+      "learning_rate": 6.3727189789174205e-06,
+      "loss": 1.2421,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 39,
+      "tokens_per_second_per_gpu": 3471.55
+    },
+    {
+      "epoch": 0.7619047619047619,
+      "grad_norm": 8.92693366901581,
+      "learning_rate": 6.263788377322381e-06,
+      "loss": 1.2587,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 40,
+      "tokens_per_second_per_gpu": 3700.61
+    },
+    {
+      "epoch": 0.780952380952381,
+      "grad_norm": 25.621463437533773,
+      "learning_rate": 6.152329412653491e-06,
+      "loss": 1.2535,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 41,
+      "tokens_per_second_per_gpu": 3696.17
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 21.356947105637357,
+      "learning_rate": 6.038466570367669e-06,
+      "loss": 1.2437,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 42,
+      "tokens_per_second_per_gpu": 3679.52
+    },
+    {
+      "epoch": 0.819047619047619,
+      "grad_norm": 21.528748134497796,
+      "learning_rate": 5.922327020746735e-06,
+      "loss": 1.2243,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 43,
+      "tokens_per_second_per_gpu": 3654.06
+    },
+    {
+      "epoch": 0.8380952380952381,
+      "grad_norm": 14.734257530424147,
+      "learning_rate": 5.804040476864407e-06,
+      "loss": 1.2326,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 44,
+      "tokens_per_second_per_gpu": 3581.66
+    },
+    {
+      "epoch": 0.8571428571428571,
+      "grad_norm": 13.129280834101875,
+      "learning_rate": 5.68373904971334e-06,
+      "loss": 1.2442,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 45,
+      "tokens_per_second_per_gpu": 3788.2
+    },
+    {
+      "epoch": 0.8761904761904762,
+      "grad_norm": 14.976302382446457,
+      "learning_rate": 5.561557100653979e-06,
+      "loss": 1.2486,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 46,
+      "tokens_per_second_per_gpu": 3636.88
+    },
+    {
+      "epoch": 0.8952380952380953,
+      "grad_norm": 15.967232506668388,
+      "learning_rate": 5.43763109135005e-06,
+      "loss": 1.2338,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 47,
+      "tokens_per_second_per_gpu": 3759.31
+    },
+    {
+      "epoch": 0.9142857142857143,
+      "grad_norm": 16.354797247719976,
+      "learning_rate": 5.312099431358276e-06,
+      "loss": 1.2413,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 48,
+      "tokens_per_second_per_gpu": 3663.89
+    },
+    {
+      "epoch": 0.9333333333333333,
+      "grad_norm": 6.665663198954394,
+      "learning_rate": 5.185102323542536e-06,
+      "loss": 1.2395,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 49,
+      "tokens_per_second_per_gpu": 3727.2
+    },
+    {
+      "epoch": 0.9523809523809523,
+      "grad_norm": 9.1334624753648,
+      "learning_rate": 5.056781607485144e-06,
+      "loss": 1.2268,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 50,
+      "tokens_per_second_per_gpu": 3870.66
+    },
+    {
+      "epoch": 0.9714285714285714,
+      "grad_norm": 17.527340590112377,
+      "learning_rate": 4.927280601070113e-06,
+      "loss": 1.2248,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 51,
+      "tokens_per_second_per_gpu": 3582.22
+    },
+    {
+      "epoch": 0.9904761904761905,
+      "grad_norm": 19.222165420352905,
+      "learning_rate": 4.796743940415344e-06,
+      "loss": 1.2254,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 52,
+      "tokens_per_second_per_gpu": 3727.73
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 16.84364160949164,
+      "learning_rate": 4.66531741833252e-06,
+      "loss": 1.242,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.17,
+      "memory/max_allocated (GiB)": 122.81,
+      "step": 53,
+      "tokens_per_second_per_gpu": 3750.91
+    },
+    {
+      "epoch": 1.019047619047619,
+      "grad_norm": 25.10526965511846,
+      "learning_rate": 4.533147821495116e-06,
+      "loss": 1.2426,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 54,
+      "tokens_per_second_per_gpu": 3667.97
+    },
+    {
+      "epoch": 1.0380952380952382,
+      "grad_norm": 24.822314802816855,
+      "learning_rate": 4.400382766496394e-06,
+      "loss": 1.2394,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 55,
+      "tokens_per_second_per_gpu": 3712.75
+    },
+    {
+      "epoch": 1.0571428571428572,
+      "grad_norm": 19.222938204469422,
+      "learning_rate": 4.267170534980487e-06,
+      "loss": 1.2269,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 56,
+      "tokens_per_second_per_gpu": 3874.53
+    },
+    {
+      "epoch": 1.0761904761904761,
+      "grad_norm": 14.962813195503772,
+      "learning_rate": 4.133659908030698e-06,
+      "loss": 1.233,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 57,
+      "tokens_per_second_per_gpu": 3626.61
+    },
+    {
+      "epoch": 1.0952380952380953,
+      "grad_norm": 23.099619927044888,
+      "learning_rate": 4e-06,
+      "loss": 1.2353,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 58,
+      "tokens_per_second_per_gpu": 3631.82
+    },
+    {
+      "epoch": 1.1142857142857143,
+      "grad_norm": 14.683578827379744,
+      "learning_rate": 3.8663400919693026e-06,
+      "loss": 1.2261,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 59,
+      "tokens_per_second_per_gpu": 3778.88
+    },
+    {
+      "epoch": 1.1333333333333333,
+      "grad_norm": 1363.244724375689,
+      "learning_rate": 3.7328294650195136e-06,
+      "loss": 1.2448,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 60,
+      "tokens_per_second_per_gpu": 3648.86
+    },
+    {
+      "epoch": 1.1523809523809523,
+      "grad_norm": 37.56736283967858,
+      "learning_rate": 3.5996172335036064e-06,
+      "loss": 1.2134,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 61,
+      "tokens_per_second_per_gpu": 3680.45
+    },
+    {
+      "epoch": 1.1714285714285715,
+      "grad_norm": 24.14759116678243,
+      "learning_rate": 3.4668521785048856e-06,
+      "loss": 1.2201,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 62,
+      "tokens_per_second_per_gpu": 3742.93
+    },
+    {
+      "epoch": 1.1904761904761905,
+      "grad_norm": 20.895518933622306,
+      "learning_rate": 3.3346825816674796e-06,
+      "loss": 1.2248,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 63,
+      "tokens_per_second_per_gpu": 3729.87
+    },
+    {
+      "epoch": 1.2095238095238094,
+      "grad_norm": 20.07417789192824,
+      "learning_rate": 3.2032560595846563e-06,
+      "loss": 1.2253,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 64,
+      "tokens_per_second_per_gpu": 3664.34
+    },
+    {
+      "epoch": 1.2285714285714286,
+      "grad_norm": 14.61511907498168,
+      "learning_rate": 3.0727193989298864e-06,
+      "loss": 1.241,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 65,
+      "tokens_per_second_per_gpu": 3721.56
+    },
+    {
+      "epoch": 1.2476190476190476,
+      "grad_norm": 18.1080641996899,
+      "learning_rate": 2.943218392514856e-06,
+      "loss": 1.2027,
+      "memory/device_reserved (GiB)": 127.34,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 66,
+      "tokens_per_second_per_gpu": 3589.14
+    },
+    {
+      "epoch": 1.2666666666666666,
+      "grad_norm": 88.35410261817876,
+      "learning_rate": 2.8148976764574643e-06,
+      "loss": 1.221,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 67,
+      "tokens_per_second_per_gpu": 3718.05
+    },
+    {
+      "epoch": 1.2857142857142856,
+      "grad_norm": 23.72041286077318,
+      "learning_rate": 2.6879005686417232e-06,
+      "loss": 1.2172,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 68,
+      "tokens_per_second_per_gpu": 3764.91
+    },
+    {
+      "epoch": 1.3047619047619048,
+      "grad_norm": 43.54234028579835,
+      "learning_rate": 2.5623689086499492e-06,
+      "loss": 1.2326,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 69,
+      "tokens_per_second_per_gpu": 3359.73
+    },
+    {
+      "epoch": 1.3238095238095238,
+      "grad_norm": 6.104685395227184,
+      "learning_rate": 2.4384428993460207e-06,
+      "loss": 1.2427,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 70,
+      "tokens_per_second_per_gpu": 3681.16
+    },
+    {
+      "epoch": 1.342857142857143,
+      "grad_norm": 9.963394838549585,
+      "learning_rate": 2.3162609502866607e-06,
+      "loss": 1.2322,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 71,
+      "tokens_per_second_per_gpu": 3753.07
+    },
+    {
+      "epoch": 1.361904761904762,
+      "grad_norm": 43.43949979845249,
+      "learning_rate": 2.195959523135592e-06,
+      "loss": 1.2383,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 72,
+      "tokens_per_second_per_gpu": 3764.97
+    },
+    {
+      "epoch": 1.380952380952381,
+      "grad_norm": 14.107017331391786,
+      "learning_rate": 2.077672979253265e-06,
+      "loss": 1.2225,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 73,
+      "tokens_per_second_per_gpu": 3751.34
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 10.549323906590455,
+      "learning_rate": 1.96153342963233e-06,
+      "loss": 1.2214,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 74,
+      "tokens_per_second_per_gpu": 3559.51
+    },
+    {
+      "epoch": 1.4190476190476191,
+      "grad_norm": 18.592940657981064,
+      "learning_rate": 1.8476705873465096e-06,
+      "loss": 1.2171,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 75,
+      "tokens_per_second_per_gpu": 3629.78
+    },
+    {
+      "epoch": 1.438095238095238,
+      "grad_norm": 11.120257290964485,
+      "learning_rate": 1.7362116226776187e-06,
+      "loss": 1.2226,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 76,
+      "tokens_per_second_per_gpu": 3603.12
+    },
+    {
+      "epoch": 1.457142857142857,
+      "grad_norm": 7.078043688121306,
+      "learning_rate": 1.627281021082579e-06,
+      "loss": 1.2345,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 77,
+      "tokens_per_second_per_gpu": 3780.85
+    },
+    {
+      "epoch": 1.4761904761904763,
+      "grad_norm": 5.000285151965608,
+      "learning_rate": 1.521000444159021e-06,
+      "loss": 1.2116,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 78,
+      "tokens_per_second_per_gpu": 3695.41
+    },
+    {
+      "epoch": 1.4952380952380953,
+      "grad_norm": 47.84624251792891,
+      "learning_rate": 1.4174885937647903e-06,
+      "loss": 1.2405,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 79,
+      "tokens_per_second_per_gpu": 3605.95
+    },
+    {
+      "epoch": 1.5142857142857142,
+      "grad_norm": 12.461343395029726,
+      "learning_rate": 1.316861079443107e-06,
+      "loss": 1.2272,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 80,
+      "tokens_per_second_per_gpu": 3613.63
+    },
+    {
+      "epoch": 1.5333333333333332,
+      "grad_norm": 7.656217867750634,
+      "learning_rate": 1.2192302893014308e-06,
+      "loss": 1.2265,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 81,
+      "tokens_per_second_per_gpu": 3752.87
+    },
+    {
+      "epoch": 1.5523809523809524,
+      "grad_norm": 15.082668616044355,
+      "learning_rate": 1.1247052644882832e-06,
+      "loss": 1.2183,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 82,
+      "tokens_per_second_per_gpu": 3677.86
+    },
+    {
+      "epoch": 1.5714285714285714,
+      "grad_norm": 16.44949015042616,
+      "learning_rate": 1.0333915774081697e-06,
+      "loss": 1.2099,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.83,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 83,
+      "tokens_per_second_per_gpu": 3729.01
+    },
+    {
+      "epoch": 1.5904761904761906,
+      "grad_norm": 12.211227945509856,
+      "learning_rate": 9.453912138106721e-07,
+      "loss": 1.2231,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 84,
+      "tokens_per_second_per_gpu": 3649.49
+    },
+    {
+      "epoch": 1.6095238095238096,
+      "grad_norm": 7.074192518964132,
+      "learning_rate": 8.60802458885356e-07,
+      "loss": 1.237,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 85,
+      "tokens_per_second_per_gpu": 3783.87
+    },
+    {
+      "epoch": 1.6285714285714286,
+      "grad_norm": 13.131068251165631,
+      "learning_rate": 7.797197874897485e-07,
+      "loss": 1.2116,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 86,
+      "tokens_per_second_per_gpu": 3671.48
+    },
+    {
+      "epoch": 1.6476190476190475,
+      "grad_norm": 15.417850715738988,
+      "learning_rate": 7.022337586329596e-07,
+      "loss": 1.2209,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 87,
+      "tokens_per_second_per_gpu": 3805.45
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 24.13403904325753,
+      "learning_rate": 6.28430914332812e-07,
+      "loss": 1.217,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 88,
+      "tokens_per_second_per_gpu": 3706.88
+    },
+    {
+      "epoch": 1.6857142857142857,
+      "grad_norm": 13.576166990616798,
+      "learning_rate": 5.583936829594433e-07,
+      "loss": 1.2017,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 89,
+      "tokens_per_second_per_gpu": 3820.7
+    },
+    {
+      "epoch": 1.704761904761905,
+      "grad_norm": 8.573005189398867,
+      "learning_rate": 4.92200287173314e-07,
+      "loss": 1.2096,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 90,
+      "tokens_per_second_per_gpu": 3759.45
+    },
+    {
+      "epoch": 1.723809523809524,
+      "grad_norm": 5.5800010726124025,
+      "learning_rate": 4.299246565604755e-07,
+      "loss": 1.2218,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 91,
+      "tokens_per_second_per_gpu": 3499.8
+    },
+    {
+      "epoch": 1.7428571428571429,
+      "grad_norm": 6.765368030458938,
+      "learning_rate": 3.716363450626372e-07,
+      "loss": 1.2117,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 92,
+      "tokens_per_second_per_gpu": 3468.37
+    },
+    {
+      "epoch": 1.7619047619047619,
+      "grad_norm": 7.504548685452772,
+      "learning_rate": 3.174004532942844e-07,
+      "loss": 1.2299,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 93,
+      "tokens_per_second_per_gpu": 3700.98
+    },
+    {
+      "epoch": 1.7809523809523808,
+      "grad_norm": 8.649122371866438,
+      "learning_rate": 2.672775558335898e-07,
+      "loss": 1.2265,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 94,
+      "tokens_per_second_per_gpu": 3700.55
+    },
+    {
+      "epoch": 1.8,
+      "grad_norm": 11.91832294221251,
+      "learning_rate": 2.2132363356832528e-07,
+      "loss": 1.2185,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 95,
+      "tokens_per_second_per_gpu": 3680.96
+    },
+    {
+      "epoch": 1.819047619047619,
+      "grad_norm": 9.186156818821193,
+      "learning_rate": 1.795900111723503e-07,
+      "loss": 1.2008,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 96,
+      "tokens_per_second_per_gpu": 3658.84
+    },
+    {
+      "epoch": 1.8380952380952382,
+      "grad_norm": 13.72977541399496,
+      "learning_rate": 1.4212329978249415e-07,
+      "loss": 1.2104,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.99,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 97,
+      "tokens_per_second_per_gpu": 3581.27
+    },
+    {
+      "epoch": 1.8571428571428572,
+      "grad_norm": 6.290457692715211,
+      "learning_rate": 1.0896534493985177e-07,
+      "loss": 1.223,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 98,
+      "tokens_per_second_per_gpu": 3791.37
+    },
+    {
+      "epoch": 1.8761904761904762,
+      "grad_norm": 9.702798624165407,
+      "learning_rate": 8.0153179853653e-08,
+      "loss": 1.2285,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 99,
+      "tokens_per_second_per_gpu": 3639.09
+    },
+    {
+      "epoch": 1.8952380952380952,
+      "grad_norm": 11.005975725667684,
+      "learning_rate": 5.571898403988573e-08,
+      "loss": 1.2151,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 100,
+      "tokens_per_second_per_gpu": 3757.98
+    },
+    {
+      "epoch": 1.9142857142857141,
+      "grad_norm": 8.44842365055977,
+      "learning_rate": 3.569004738087988e-08,
+      "loss": 1.2238,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 101,
+      "tokens_per_second_per_gpu": 3661.88
+    },
+    {
+      "epoch": 1.9333333333333333,
+      "grad_norm": 4.816542675360639,
+      "learning_rate": 2.0088739645983455e-08,
+      "loss": 1.2232,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 102,
+      "tokens_per_second_per_gpu": 3730.52
+    },
+    {
+      "epoch": 1.9523809523809523,
+      "grad_norm": 11.749396247795026,
+      "learning_rate": 8.932485507387344e-09,
+      "loss": 1.2118,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 103,
+      "tokens_per_second_per_gpu": 3871.77
+    },
+    {
+      "epoch": 1.9714285714285715,
+      "grad_norm": 7.9532371124526104,
+      "learning_rate": 2.2337450789815526e-09,
+      "loss": 1.2109,
+      "memory/device_reserved (GiB)": 127.42,
+      "memory/max_active (GiB)": 124.18,
+      "memory/max_allocated (GiB)": 122.82,
+      "step": 104,
+      "tokens_per_second_per_gpu": 3582.44
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 104,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 13,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1428859668922368.0,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75728d1f222de640ed232302fa9ccb9af4d5b61b86d229500c0faae54317cbd9
+size 9169

zero_to_fp32.py

@@ -0,0 +1,760 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import gc
+import json
+import numpy as np
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device, weights_only=False)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in tqdm(files, desc='Loading checkpoint shards'):
+        state_dict = torch.load(f, map_location=device, mmap=True, weights_only=False)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+class GatheredTensor:
+    """
+    A pseudo tensor that collects partitioned weights.
+    It is more memory efficient when there are multiple groups.
+    """
+
+    def __init__(self, flat_groups, flat_groups_offset, offset, partitioned_numel, shape):
+        self.flat_groups = flat_groups
+        self.flat_groups_offset = flat_groups_offset
+        self.offset = offset
+        self.partitioned_numel = partitioned_numel
+        self.shape = shape
+        self.dtype = self.flat_groups[0][0].dtype
+
+    def contiguous(self):
+        """
+        Merge partitioned weights from flat_groups into a single tensor.
+        """
+        end_idx = self.offset + self.partitioned_numel
+        world_size = len(self.flat_groups)
+        pad_flat_param_chunks = []
+
+        for rank_i in range(world_size):
+            # for each rank, we need to collect weights from related group/groups
+            flat_groups_at_rank_i = self.flat_groups[rank_i]
+            start_group_id = None
+            end_group_id = None
+            for group_id in range(len(self.flat_groups_offset)):
+                if self.flat_groups_offset[group_id] <= self.offset < self.flat_groups_offset[group_id + 1]:
+                    start_group_id = group_id
+                if self.flat_groups_offset[group_id] < end_idx <= self.flat_groups_offset[group_id + 1]:
+                    end_group_id = group_id
+                    break
+            # collect weights from related group/groups
+            for group_id in range(start_group_id, end_group_id + 1):
+                flat_tensor = flat_groups_at_rank_i[group_id]
+                start_offset = self.offset - self.flat_groups_offset[group_id]
+                end_offset = min(end_idx, self.flat_groups_offset[group_id + 1]) - self.flat_groups_offset[group_id]
+                pad_flat_param_chunks.append(flat_tensor[start_offset:end_offset])
+
+        # collect weights from all ranks
+        pad_flat_param = torch.cat(pad_flat_param_chunks, dim=0)
+        param = pad_flat_param[:self.shape.numel()].view(self.shape).contiguous()
+        return param
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = sum([flat_group.numel() for flat_group in fp32_flat_groups[0]]) * world_size
+
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    flat_groups_offset = [0] + list(np.cumsum([flat_tensor.numel() for flat_tensor in fp32_flat_groups[0]]))
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering sharded weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # memory efficient tensor
+        tensor = GatheredTensor(fp32_flat_groups, flat_groups_offset, offset, partitioned_numel, shape)
+        state_dict[name] = tensor
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def to_torch_tensor(state_dict, return_empty_tensor=False):
+    """
+    Convert state_dict of GatheredTensor to torch tensor
+    """
+    torch_state_dict = {}
+    converted_tensors = {}
+    for name, tensor in state_dict.items():
+        tensor_id = id(tensor)
+        if tensor_id in converted_tensors:  # shared tensors
+            shared_tensor = torch_state_dict[converted_tensors[tensor_id]]
+            torch_state_dict[name] = shared_tensor
+        else:
+            converted_tensors[tensor_id] = name
+            if return_empty_tensor:
+                torch_state_dict[name] = torch.empty(tensor.shape, dtype=tensor.dtype)
+            else:
+                torch_state_dict[name] = tensor.contiguous()
+    return torch_state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                             tag=None,
+                                             exclude_frozen_parameters=False,
+                                             lazy_mode=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+        - ``lazy_mode``: get state_dict in lazy mode. It returns a dict of pesduo tensor instead of torch tensor, which is more memory efficient.
+          Convert the pesduo tensor to torch tensor by ``.contiguous()``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    Note: the above usage may not work if your application doesn't have sufficient free CPU memory.
+    You may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint. Or you can load state_dict in lazy mode ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, lazy_mode=True) # not on cpu
+        for name, lazy_tensor in state_dict.item():
+            tensor = lazy_tensor.contiguous()  # to cpu
+            print(name, tensor)
+            # del tensor to release memory if it no longer in use
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    state_dict = _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+    if lazy_mode:
+        return state_dict
+    else:
+        return to_torch_tensor(state_dict)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                                          tag,
+                                                          exclude_frozen_parameters,
+                                                          lazy_mode=True)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        # an memory-efficient approach for sharding
+        empty_state_dict = to_torch_tensor(state_dict, return_empty_tensor=True)
+        state_dict_split = split_torch_state_dict_into_shards(empty_state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model by shard
+    os.makedirs(output_dir, exist_ok=True)
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard_state_dict = {tensor_name: state_dict[tensor_name] for tensor_name in tensors}
+        shard_state_dict = to_torch_tensor(shard_state_dict)
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard_state_dict, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard_state_dict, output_path)
+        # release the memory of current shard
+        for tensor_name in list(shard_state_dict.keys()):
+            del state_dict[tensor_name]
+            del shard_state_dict[tensor_name]
+        del shard_state_dict
+        gc.collect()
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)