Add files using upload-large-folder tool

config.json

@@ -0,0 +1,35 @@
+{
+  "architectures": [
+    "Qwen2MoeForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "decoder_sparse_step": 1,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen2_moe",
+  "moe_intermediate_size": 2560,
+  "norm_topk_prob": false,
+  "num_attention_heads": 28,
+  "num_experts": 64,
+  "num_experts_per_tok": 8,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "output_router_logits": false,
+  "rms_norm_eps": 1e-06,
+  "rope_theta": 1000000.0,
+  "router_aux_loss_coef": 0.001,
+  "shared_expert_intermediate_size": 20480,
+  "sliding_window": 65536,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.40.1",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}

generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "pad_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "repetition_penalty": 1.05,
+  "temperature": 0.7,
+  "top_p": 0.8,
+  "top_k": 20,
+  "transformers_version": "4.40.0.dev0"
+}

special_tokens_map.json

@@ -0,0 +1,16 @@
+{
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tllm-checkpoint/config.json

@@ -0,0 +1,194 @@
+{
+    "producer": {
+        "name": "modelopt",
+        "version": "0.25.0"
+    },
+    "architecture": "Qwen2MoeForCausalLM",
+    "dtype": "bfloat16",
+    "logits_dtype": "float16",
+    "num_hidden_layers": 28,
+    "num_attention_heads": 28,
+    "num_key_value_heads": 4,
+    "hidden_size": 3584,
+    "norm_epsilon": 1e-06,
+    "vocab_size": 151936,
+    "max_position_embeddings": 32768,
+    "hidden_act": "swiglu",
+    "use_parallel_embedding": true,
+    "embedding_sharding_dim": 0,
+    "head_size": 128,
+    "intermediate_size": 18944,
+    "position_embedding_type": "rope_gpt_neox",
+    "share_embedding_table": false,
+    "residual_mlp": false,
+    "bias": false,
+    "rotary_pct": 1.0,
+    "rank": 1,
+    "decoder": "qwen",
+    "rmsnorm": true,
+    "lm_head_bias": false,
+    "mlp_bias": false,
+    "attn_bias": true,
+    "rotary_base": 1000000.0,
+    "rotary_scaling": null,
+    "disable_weight_only_quant_plugin": false,
+    "num_labels": 1,
+    "use_logn_attn": false,
+    "moe": {
+        "num_experts": 64,
+        "shared_expert_intermediate_size": 0,
+        "top_k": 8,
+        "normalization_mode": 0,
+        "sparse_mixer_epsilon": 0.01,
+        "tp_mode": 0,
+        "device_limited_n_group": 0,
+        "device_limited_topk_group": 0,
+        "device_limited_routed_scaling_factor": 1.0
+    },
+    "runtime_defaults": null,
+    "mapping": {
+        "world_size": 2,
+        "gpus_per_node": 8,
+        "cp_size": 1,
+        "tp_size": 2,
+        "pp_size": 1,
+        "moe_tp_size": 2,
+        "moe_ep_size": 1,
+        "auto_parallel": false
+    },
+    "quantization": {
+        "quant_algo": "FP8",
+        "kv_cache_quant_algo": null,
+        "group_size": 128,
+        "smoothquant_val": 0.5,
+        "clamp_val": null,
+        "use_meta_recipe": false,
+        "has_zero_point": false,
+        "pre_quant_scale": false,
+        "exclude_modules": [
+            "transformer.layers.13.mlp.router",
+            "transformer.layers.1.mlp.shared_expert_gate",
+            "transformer.layers.24.mlp.router",
+            "transformer.layers.6.mlp.router",
+            "transformer.layers.25.post_layernorm",
+            "transformer.layers.1.input_layernorm",
+            "transformer.layers.18.post_layernorm",
+            "transformer.layers.6.mlp.shared_expert_gate",
+            "transformer.layers.18.input_layernorm",
+            "transformer.layers.16.input_layernorm",
+            "transformer.layers.21.mlp.router",
+            "transformer.layers.2.mlp.router",
+            "transformer.layers.11.post_layernorm",
+            "transformer.layers.21.input_layernorm",
+            "transformer.layers.6.input_layernorm",
+            "transformer.layers.25.mlp.router",
+            "transformer.layers.26.post_layernorm",
+            "transformer.layers.23.mlp.shared_expert_gate",
+            "transformer.layers.26.mlp.shared_expert_gate",
+            "transformer.layers.23.mlp.router",
+            "transformer.layers.3.mlp.shared_expert_gate",
+            "transformer.vocab_embedding",
+            "transformer.layers.15.mlp.router",
+            "transformer.ln_f",
+            "transformer.layers.25.mlp.shared_expert_gate",
+            "transformer.layers.19.input_layernorm",
+            "transformer.layers.0.post_layernorm",
+            "transformer.layers.22.input_layernorm",
+            "transformer.layers.15.mlp.shared_expert_gate",
+            "transformer.layers.9.mlp.shared_expert_gate",
+            "transformer.layers.26.mlp.router",
+            "transformer.layers.27.mlp.router",
+            "transformer.layers.2.input_layernorm",
+            "transformer.layers.9.input_layernorm",
+            "transformer.layers.16.mlp.shared_expert_gate",
+            "transformer.layers.20.mlp.shared_expert_gate",
+            "transformer.layers.9.post_layernorm",
+            "transformer.layers.17.input_layernorm",
+            "transformer.layers.13.post_layernorm",
+            "transformer.layers.22.mlp.router",
+            "transformer.layers.2.mlp.shared_expert_gate",
+            "transformer.layers.19.mlp.router",
+            "lm_head",
+            "transformer.layers.20.post_layernorm",
+            "transformer.layers.7.post_layernorm",
+            "transformer.layers.7.mlp.router",
+            "transformer.layers.18.mlp.router",
+            "transformer.layers.17.mlp.router",
+            "transformer.layers.24.mlp.shared_expert_gate",
+            "transformer.layers.23.input_layernorm",
+            "transformer.layers.12.mlp.shared_expert_gate",
+            "transformer.layers.4.input_layernorm",
+            "transformer.layers.11.mlp.shared_expert_gate",
+            "transformer.layers.20.mlp.router",
+            "transformer.layers.23.post_layernorm",
+            "transformer.layers.10.mlp.shared_expert_gate",
+            "transformer.layers.13.input_layernorm",
+            "transformer.layers.17.mlp.shared_expert_gate",
+            "transformer.layers.14.mlp.shared_expert_gate",
+            "transformer.layers.4.mlp.shared_expert_gate",
+            "transformer.layers.14.input_layernorm",
+            "transformer.layers.5.input_layernorm",
+            "transformer.layers.6.post_layernorm",
+            "transformer.layers.12.input_layernorm",
+            "transformer.layers.15.post_layernorm",
+            "transformer.layers.17.post_layernorm",
+            "transformer.layers.4.post_layernorm",
+            "transformer.layers.1.post_layernorm",
+            "transformer.layers.18.mlp.shared_expert_gate",
+            "transformer.layers.0.input_layernorm",
+            "transformer.layers.10.mlp.router",
+            "transformer.layers.16.post_layernorm",
+            "transformer.layers.12.mlp.router",
+            "transformer.layers.5.post_layernorm",
+            "transformer.layers.8.post_layernorm",
+            "transformer.layers.16.mlp.router",
+            "transformer.layers.24.input_layernorm",
+            "transformer.layers.22.post_layernorm",
+            "transformer.layers.3.input_layernorm",
+            "transformer.layers.5.mlp.router",
+            "transformer.layers.14.post_layernorm",
+            "transformer.layers.8.mlp.shared_expert_gate",
+            "transformer.layers.19.post_layernorm",
+            "transformer.layers.12.post_layernorm",
+            "transformer.layers.3.post_layernorm",
+            "transformer.layers.24.post_layernorm",
+            "transformer.layers.4.mlp.router",
+            "transformer.layers.11.input_layernorm",
+            "transformer.layers.0.mlp.shared_expert_gate",
+            "transformer.layers.10.post_layernorm",
+            "transformer.layers.2.post_layernorm",
+            "transformer.layers.19.mlp.shared_expert_gate",
+            "transformer.layers.22.mlp.shared_expert_gate",
+            "transformer.layers.26.input_layernorm",
+            "transformer.layers.27.input_layernorm",
+            "transformer.layers.21.post_layernorm",
+            "transformer.layers.9.mlp.router",
+            "transformer.layers.5.mlp.shared_expert_gate",
+            "transformer.layers.27.mlp.shared_expert_gate",
+            "transformer.layers.7.input_layernorm",
+            "transformer.layers.3.mlp.router",
+            "transformer.layers.7.mlp.shared_expert_gate",
+            "transformer.layers.8.mlp.router",
+            "transformer.layers.8.input_layernorm",
+            "transformer.layers.21.mlp.shared_expert_gate",
+            "transformer.layers.27.post_layernorm",
+            "transformer.layers.20.input_layernorm",
+            "transformer.layers.0.mlp.router",
+            "transformer.layers.25.input_layernorm",
+            "transformer.layers.14.mlp.router",
+            "transformer.layers.1.mlp.router",
+            "transformer.layers.13.mlp.shared_expert_gate",
+            "transformer.layers.11.mlp.router",
+            "transformer.layers.10.input_layernorm",
+            "transformer.layers.15.input_layernorm"
+        ]
+    },
+    "qk_layernorm": false,
+    "rotary_embedding_dim": 128,
+    "seq_length": 8192,
+    "qwen_type": "qwen2_moe",
+    "moe_intermediate_size": 2560,
+    "moe_shared_expert_intermediate_size": 20480,
+    "tie_word_embeddings": false,
+    "model_type": "qwen"
+}

tllm-checkpoint/rank0.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:144703e2b8bf6ce02fea53206a4e8c73e861079aa132032fdab31077024c9b83
+size 38520404352

tllm-checkpoint/rank1.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7dd89ce139dd77badd5932c8f495e8a309f9f7f74c6081100d449c746414b8fd
+size 38520404352

tokenizer_config.json

@@ -0,0 +1,41 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": ["<|im_start|>", "<|im_end|>"],
+  "bos_token": null,
+  "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "model_max_length": 65536,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}
+